United States Office of July 1983
Environmental Protection Drinking \ A ater (WH 550) EPA 570-9-83-001
Agency Washingtoh DC 20460
EPA Assessment of
Microbiology and
Turbidity Standards
for Drinking Water
Proceedings of a
Workshop
December 2-4, 1981
-------
[ )ISCLAI MER
This report has been reviewed by the Office of Drinking Wafer, U.S. Fnvironnentcil
Protection Agency, and approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the U.S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or recommendation for
use.
-------
United States July 1983
Environmental Protection EPA 570-9-83-001
Agency
Assessment of
Microbiology and
Turbidity Standards
for Drinking Water
Paul S. Berger, Ph.D.
Workshop Chairman and Editor
Criteria and Standards Division
Office of Drinking Water
Yerachmiel Argaman, Ph.D., P.E.
Co-Editor
AWARE Corporation
Nashville. Tennessee
This document is available through the National Technical
Information Service, Springfield, VA 22151.
-------
ACKNOWLEDGMENTS
Special recognition must go to the panel of experts who, as members of the workshop steering
committee, provided invaluable advice on the planning and conduct of the workshop. They are:
Dr. Elmer W. Akin Dr. Vincent P. Olivieri
Health Effects Research Lab (EPA) Johns Hopkins University
Dr. Richard S. Englebrecht Dr. Betty H. Olson
University of Illinois University of California
Mr. Edwin E. Geldreich Dr. Wesley 0. Pipes, Jr.
Municipal Environmental Research Drexel University
Laboratory (EPA)
Mr. A. E. Greenberg Dr. Ramon J. Seidler
California Dept. of Health Services Oregon State University
Dr. William J. Hausler Dr. Mark D. Sobsey
University of Iowa University of North Carolina
Dr. Riley D. Housewright
Executive Director, American
Society for Microbiology
I am indebted to the panel chairmen who volunteered to prepare issues papers on short notice
and did a superb job In this and in guiding their panel discussion. They also served on the steering
committee. They are Dr. Engelbrecht, Mr. Greenberg, Dr. Olivieri, Dr. Olson, Dr. Pipes,
Dr. Seidler, and Dr. Sobsey.
Each of the six workshop panels included an EPA resource person who assisted the panel
chairman by providing papers and other materials to the chairman during issues paper development,
by helping the chairman keep panel discussions on track during the workshop, and by reviewing
drafts of revised issues papers after the workshop. These individuals are Mr. Robert H. Bordner
(EMSL-Cl), Dr. Alfred P. Dufour (HERL), Mr. Edwin E. Geldreich (MERL), Mr. Walter Jakubowski
(HERL), Mr. Donald Maddox (Region V), and Mr. Raymond H. Taylor (MERL).
The workshop was sponsored by the Office of Drinking Water, in conjunction with the
American Society for Microbiology. The assistance provided by the ASM in planning and
participating in the workshop is thankfully acknowledged.
Acknowledgment and thanks are due to the AWARE Corporation personnel, our contractor for
the workshop. Ms. Becky Boone took care of the mechanics of the workshop and stayed up past
midnight typing panel summaries during the meeting. Dr. Yerachmiel Argaman was particularly
helpful as co-editor of these proceedings. Dr. Ann Clarke provided substantial support on the final
draft manuscript. Special appreciation goes to Mr. Douglas Williams, EPA/Center for Environ-
mental Research Information, who was project officer for the AWARE Corporation contract, and
who kept contractual arrangements on course in sometimes rough seas.
Finally, I am especially indebted to Dr. Joseph A. Cotruvo (Director, Criteria and Standards
Division, Office of Drinking Water) and Dr. William L. Lappenbusch (Chief, Health Effects Branch,
Criteria and Standards Division, Office of Drinking Water) who provided invaluable guidance and
encouragement during the workshop planning process.
Paul S. Berger, Ph.D.
Workshop Chairman
-------
LIST OF CONTENTS
INTRODUCTION
OPENING SESSION
Joseph A. Cotruvo
Riley D. Housewright
Gunther F. Craun and
Leland J. McCabe
Edwin E. Geidreich
EXECUTIVE SUMMARIES OF ISSUES PAPERS
ISSUES PAPERS
Mark D. Sobsey and Betty H. Olson
Vincent P. Olivieri
Wesley 0. Pipes
Ramon J. Seidler and
Thomas M. Evans
Richard S. Engelbrecht
Arnold E. Greenberg
Introductory Remarks
Keynote Address
Waterborne Disease in the United States
and the Coliform Standard
Emerging Microbiological Issues
Microbial Agents of Waterborne Disease
Measurement of Microbial Quality
Monitoring of Microbial Water Quality
Analytical Methods for Microbial Water
Quality
Source, Treatment and Distribution
Compliance and Policy Issues
-------
INTRODUCTION
-------
INTRODUCTION
On December 2-4, 1981, the U.S. EPA Office of Drinking Water (ODW), in conjunction with
the American Society for Microbiology, sponsored a microbiology standards workshop at the Airlie
House in Warrenton, Virginia. The purpose WQ5 to e omine options on revising the current drinking
water microbiology arid turbidity standards, arid make recommendations to ODW. The workshop
was scientifically-oriented and addressed such issues as indicators of water quality, monitoring,
analytical procedures, water treatment and distribution, and the scientific supportability for and
the advisability of standards for specific waterborne pothogens. Non-technical policy issues and
regulatory options were also discussed.
Participants in the workshop included professional microbiologists, public health experts,
engineers, Federal, State and local public officials, pdblic water supply personnel, and U.S. EPA
personnel. The participants were assigned to s x panels, each charged with specific topics for
discussion and recommendations. The chairman of each panel prepared an issues paper which
served as the starting point for consideration and debate of the issues assigned to that panel.
The six panels and their chairmen were as follows:
Agents of Waterborne Disease - Dr. Mark D. Sobsey and Dr. Betty H. Olson
Measurement of Microbial Water Quality - Dr. Vincent P. Olivieri
Monitoring - Dr. Wesley 0. Pipes
Methods - Dr. Ramon J. Se Idler
Source, Treatment, and Distribution - Dr. Richard S. Engelbrecht
Compliance and Policy Issues - Mr. Arnold E. Greenberg
Subsequent to the workshop, panel chairmen revised their issues papers to reflect the views,
conclusions, and recommendations of the panel. The final ssues papers, including recornmenda-
tions on policy and research, have been assembled In this publication. The publication also contains
formal presentations delivered at the beginning of the workshop. Panel members are identified at
the beginning of each issues paper.
The recommendations contained herein do not necessarily reflect the view of the U.S. EPA,
but will be influentiaL in the development of revised microbiology and turbidity regulations.
Paul S. Berger
Workshop Chairman
-------
OPENING SESSION
-------
INTRODUCTORY REMARKS
Dr. Joseph A. Cotruvo*
Good morning - I want to welcome all of you to U.S. EPA’s Workshop on the Microbiology of
Drinking Water. This is one of a number of workshops that we have sponsored on the scientific
basis for development of Revised National Primary Drinking Water Regulations. It is a great
opportunity for us in EPA to get together again with our old and new friends who are experts in the
various technical aspects of drinking water regulation development and it is a pleasure for us to
have you here and working with us. I know it will be a very productive session too, because it is
from your technical insights that the basic decision processes have to spring so that we can
properly develop drinking water standards to protect drinking water quality in the country.
I would like to acknowledge several people who are here. We know Dr. Riley Housewright
from the Safe Drinking Water Committee of the National Academy of Sciences, and the
participation of the American Society for Microbiology through him is much appreciated. Also, a
special note goes to a large contingent from EPA’s Office of Research and Development arid we
appreciate their direct participation in this project. A great number of microbiologists from many
walks of life are here including leading university scientists and representatives from states and
some from the water industry. I think we have gathered together the best possible group to provide
the advice necessary to take the next step in implementing the Safe Drinking Water Act.
The subject of the conference is the biological quality of drinking water and implementation
of the Safe Drinking Water Act. The basic question that we are asking is, how can the most
appropriate regulatory mechanisms be designed to assure the biological safety of drinking water in
the United States? From you we ask, what are the basic scientific principles that should be applied
to derive those regulatory mechanisms? And ultimately, what are the practical considerations that
must be included in translating from the scientific ideal to the actual application of these controls
in the field? I want to make it clear that we are not only looking for regulations and controls, we
are looking for concepts and principles, because there are many ways that all of us can have an
effect on the community health protection and it is not only specifying rules, but also by providing
leadership guidance and information. When good information is condensed and disseminated, then
*Dir Ij ia and Standards Division, Office of Drinking Water, U.S. EPA, Washington, D.C.
-------
2
states, and the public water systems wilt have a body of knowledge upon which to rely when
needed.
Drinking water sanitary microbiology is probably the oldest public health discipline that
exists. One could ask the question, why is it necessary to talk about drinking water sanitation in
the United States in 1981? Some would think that we know enough and have advanced far enough
in public health protection such that basic questions like this need not be bothered with. There are
treatment processes, like disinfection, and we know how to remove pathogens from water, but the
fact of the matter is, waterborne disease still occurs in the United States and in developed
countries, as well as in developing countries. Obviously, something is not being done properly, or
perhaps we do not really know enough about all the factors that relate to the protection of water
quality and the treatment and transmission of water to consumers.
There is a great record of success, of course. Most communities are essentially free from
risks from biological disease in drinking water, but nevertheless, those cases still do occur at least
in situations where there has been a breakdown In the protective layers in the treatment processes
or other parts of the delivery system. We will hear about a number of cases later, but clearly the
numbers are increasing. This does not necessarily mean that the actual cases are Increasing, but
we are now able to identify more cases of waterborne disease. So the more we look, the more we
find.
I must conclude that drinkiflg water sanitation Is still the most significant drinking water
quality problem in the U.S. In the last few years more and more has been said about drinking water
quality and toxic chemicals and Cancer risks and the like, but the fact remains that waterborne
disease outbreaks still occur in large numbers - 20,000 reported in 1980. In the case of waterborne
infectious disease vs. trace chemical contamination, we are talking about the difference between
the observed and the projected, or In matters of fact where actual cases are detected versus
matters of theory. Trace contaminants are being found In drinking water and drinking water
quality must be protected, but there is still an element of conjecture as to actual human risks that
result in most cases. There is no Conjecture as far as human risks that relate to the presence of
pathogens in drinking water. So since the problem is so obvious, it must be resolved.
The Safe Drinking Water Act published in I 97k was the first comprehensive national attempt
to deal with assuring the quality of drinking water. The charge Issued by Congress in that law is
-------
3
that we should protect health to the extent feasible, taking costs and other factors into considera-
tion. So there is certainly a strong protective element, but there is also the element of
practicality as to what is achievable under present conditions. The object is to identify goals or
optimum levels that we should strive for and to march inexorably toward those goals as quickly and
reasonably as we can, given feasibility and economics. Fortunately, in the case of microbiology,
economics should not be a controlling factor and technologies are available that are often of very
minimal cost; therefore, in most cases it is a matter of applying what everyone knows should be
done, rather than debates over costs and choices of technologies.
There are two mechanisms that are available in the Safe Drinking Water Act to provide the
basic level of protection at the national level: Maximum Contaminant Levels (MCLs) and specific
treatment requirements. The specific charge in the law is - if monitoring is technically and
economically feasible then the appropriate way to regulate is to establish an MCL. Maximum
contaminant levels identify individual substances, pathogens, or indirect measurements of poten-
tially harmful agents; limits for those substances can be established which should not be exceeded.
Maximum contaminant levels are appropriate in most cases. However, the law also provides for
circumstances where there may be a harmful component in the water but where monitoring is not
practical, too expensive, or perhaps the methodology is not even existent. Under those
circumstances, there is a provision that allows for the designation of particular treatment
technologies that would be appropriate for the control of those substances and protection of public
health. In the committee reports which were published with the Safe Drinking Water Act, one
particular example that was discussed where monitoring may be infeasible and where treatment
technologies may be appropriate is the case of viruses. There monitoring is particularly difficult,
expensive and inexact and not available for routine use, and therefore, a treatment approach could
be more appropriate than an MCL.
The two essential microbiology-oriented standards in the current interim primary regulations
are the coliform and turbidity standards. Monitoring for chlorine residual is also permissable in
certain situations where that can be substituted for up to 75 percent of scheduled coliform
samples. But basically that is it; there is no distinction of individual pathogens, only coliforms as a
surrogate. The Safe Drinking Water Act required EPA to produce revised national standards and
establish minimum national requirements that would be applied and perhaps extended by the states.
-------
4
So our object today is to determine what the minimum revised national requirements for
microbiology of drinking water should be to assure the quality of drinking Water. In addition, what
information and advice can be provided such that states will be able to translate and extend those
requirements as needed for their particular circumstances.
Figure I contains a few statistics on the national picture to date since the passage of the
Safe Drinking Water Act. Public water systems are those that are covered by the Safe Drinking
Water Act and these are defined as those that provide piped water to communities of 15 or more
service connection or 25 or more people for at least two months out of the year. The universe of
situations that would fit that definition are divided into community supplies which have resident
populations, and non-community supplies. Non-community public water systems have transient
populations and those could be campgrounds, interstate highway stops, or other places which may
not have 25 permanent residents but certainly do have an average of 25 people passing through on a
regular basis. The totals are certainly inexact; but more than 200,000 total public water systems
are estimated of which about 64,000 are community supplies and perhaps 150,000 are non-
community supplies. EPA ’s greatest emphasis has been on those community supplies since the vast
majority of people in the United States draw water from those supplies - from 180,000,000 to
200,000,000 people, and probably virtually everybody does at one time or another. People often
work in communities although they may live in rural areas. The vast majority of the individual
supplies are groundwaters in origin by a ratio of about five to one, roughly 50,000 to about 11,000.
Population wise, on the other hand, the vast majority of people consume drinking water that Is
derived from surface sources.
The majority of community supplies are groundwaters, as one would expect. Figure 2
contains a breakdown of supplies by type. The population groupings in this designation have been
categorized as very small, small, medium, large, and very large.
• Very large supplies are those that we have identified as larger than 100,000 population.
• The large supplies are in less than 00,000 population down to about 10,000.
• The medium supplies are those from 10,000 down to 3,300 population.
• Small supplies are 3,300 to 500.
• Very small supplies are 500 and below.
-------
FIGURE 1
PUBLIC WATER SYSTEMS
WATER SOURCE
di
POPULATION SERVED
FY80
-------
FIGURE 2
SIZE DISTRIBUTION AND POPULATION SERVED
BY THE COMMUNITY PUBLIC WATER SYSTEMS
FYI)
61.3% 150
50%
C’)
100 —
>. 0
Lii 20 — (0
Z
0
21.8% 50
10
2 91%
::::: 6.1% :::::
I __ __ 11 Fl 2 .4 [ — -
VERY SMALL MEDIUM LARGE’ VERY VERY SMALL MEDIUM LARGE VERY
SMALL LARGE SMALL LARGE
SYSTEM SIZE CATEGORY SYSTEM SIZE CATEGORY
-------
7
The numerical majority, by f or, are in the smaller categories. Among those that are 10,000
and above, the large plus the very large, are only about 5 percent of the total number of
community supplies. However, that group provides water for about 80 percent of the permanent
resident population of the country. Therefore, a very large number of supplies provide water to a
relatively small number of people and that obviously leads to some significant logistics and delivery
problems not only for the Federal Government but primarily for the states and communities in
making sure that quality water iS being maintained in each one of those more than 0,000
community public water systems.
The pie chart (Figure 3) contains data compiled by ODW’s State Programs Division from
regional reports. It gives an indication of compliance status for coliforms and turbidity. This is
the latest data of 1980 which is not totally comprehensive in its coverage, but I think reasonabiy
indicative of what is happening. Sixty-five percent of the systems according to these dat a’ were in
compliance. But 35 percent have had some type of violation of one of those two standards over the
last year or two. These are divided into persistent violations and intermittent violations, to some
degree a function of the way the standards are written and the way monitoring is required. Some
of the violations exceed the maximum contaminant level but most of the violations were either
insufficient or lack of monitoring. Some communities hove both monitored inadequately and also
indicated that they exceed the standards. Sixteen percent have inadequate monitoring and
reporting, so obviously we may not even know whether or not they are in compliance with the MCL.
So it could be that the actual counts of communities that are out of compliance with maximum
contaminant levels are higher than the roughly 3 percent that would be indicated from these data.
Figure 4 ’ is basically a replot of that same Informofion but note this is for maximum
contaminant vlol tion only for the coliforrn standard broken down by system size category and
also by intermittent and persistent violation. The national average aggregate of this group is that
about 10 percent failed the coliform standard at least occasionally.
Overall, the aggregate is about 35 percent violation with about a 5 percent overlap. Keep in
mind that to some degree either regulations or advisores in American Water Works Association
guidelines and goals have been in existence for a great number of years and these types of probIe ns
are correctable very simply in most cases.
-------
FIGURE 3
NATIONAL COMPLIANCE (% SYSTEM)
MICROBIOLOGICAL (63,9Th CWS)
COMPL lANCE
55% OF THE SYSTEMS
VIOLATION
35% OF TIlE SYSTEMS
PERSISTENT: 19%
ONLYMCL2%
ROTH MCL&M/R 1%
ONLY M 15%
INTERMITTENT: 16%
ONLY MCL 3%
BOTHMCL&MIR4%
ONLYMIR*%
COMPL lANCE
FY80
TURBIDITY (11,433 CWS)
FOR MCLOR M/R
PERSiSTENT VIOLATION: !4 MONTHS
INTERMITTENT VIOLATION: 4 MONTHS
*3% OF THE SYSTEMS
VIOLATION
17% OF THE SYSTEMS
TERMITTENT: 6%
PERSISTENT: 11% BOTHNCL&Mffi 1%
ONLY MCL 2%
ONLY MCL 3%
ONLY MIll 3%
ONLYMIRS%
-------
FIGURE 4
MICROBIOLOGICAL VIOLATION BY SYSTEM SIZE CATEGORY
FY80
MCI VIOLATIONS
1N1ERMITTANT VIOLATION (<4 MONTHS)
PERSISTENT VIOLATION( 4 MONTHS)
>.
0
0
U i
I-
>- 40
0
0
L ii
I-
l
U
U i
N
30-’
I
U
<
L i i
0
U.
z
20-
0
5
!
U) 10.8
10
>-
U)
I-
Z 5
Ui
U
Ui
0.
SYSTEM SIZE VERY SMALL MEDIUM
CATEGORY: SMALL
VERY VERY
LARGE SMALL
SMALL MEDIUM LARGE VERY
LARGE
LARGE
10.1 NATIONAL AVERAGE: 10.2%
• IN EACH 43056 13,969 3,881 2767
SiZE CATEGORY:
POPULATION (IN 1000) SERVED BY VIOLATING SYSTEMS:
302
43,056 13,969 3,881 2,767
INTERMITTENT:
457
1,101
1,200
2,721
1,090
881
2,626
2,592
4,758
2,727
PERSISTENT:
243
671
596
1,088
3,315
1,085
2,123
3,396
4,345
3,927
302
-------
l0
Figure 5 is a similar situation viewing turbidity. Since turbidity standards apply only to
surface supplies, only about 11,000 supplies are covered. A greater percentage of them do meet
the turbidity standards, but about 17 percent overall have had compliance problems. Again a split
occurs between actual MCL violations and monitoring and reporting violations and combinations.
Of course, there are some very large communities in the United States today that are using surface
waters and do not filter those waters, just to name New York City and Los Angeles for starters.
They do not necessarily exceed the turbidity standard regularly; however, they might intermit-
tently. The latest AWWA advice on turbidity in a recent issue of the Journal of AWWA strongly
recommends universal filtration for surface water supplies, which is not a new idea either. One of
the reasons for filtration is to control turbidity, and an essential reason for controlling turbidity
and applying filtration is to provide a very important barrier against the transmission of a greater
number of pathogens that are resistant to disinfection or for which disinfection alone should not be
relied upon as the total protection. On the plus side, the FY81 data shows some marked
improvements as can be seen in the comparison in Figures 6 and 7.
There are many unanswered questions, and obviously there are some very important reasons
for all of you to be here today, to be dwelling on those issues on compliance, on goals, on treatment
problems, on how to get to where the U.S. water industry ought to be in the future, to name a few.
Obviously source water quality is important. How does source water quality relate to finished
water quality, relative to appropriate technologies and treatments that will be applied in between,
versus particular kinds of pathogens, or other particular kinds of risks that may be present, because
of source water contamination problems. Other questions relate to the elements of an essential
treatment process relative to certain source quality conditions; protection of water during
transfer, the distribution system, the need for disinfectant residuals if there is one, if so what
level, what types, contact time, concentrations, etc.?
So the record is spotty, although the regulations have only been In effect since essentially
1978, so there is a translational problem here. In addition, we do know that there are communities
trying to get into compliance that ore having problems perhaps arranging financing or are in
construction phases, etc. But compliance should be in the high 90’s rather than in the 60’s, 70’s, and
80 percent range of compliance for something that is so fundamental as biological quality of
drinking water.
-------
FIGURE 5
TURBIDITY VIOLATION BY SYSTEM SIZE CATEGORY
FY80
20
M/R VIOLATIONS
NATIONAL AVERAGE: 11.6%
10.1
1•••
VERY SMALL MEDIUM LARGE VERY
LARGE
4,252 3,760 1,639 1,550
232
4,252 3,760 1,639 1,550
POPULATION (IN 1000) SERVED BY VIOLATING SYSTEMS:
25 199 288 767 2,349
519
22 252 377 1,740 3,040
86 450 580 898 610
>.
0
LU
I-
C.)
LU
N
U)
I
U
LU
0
U.
2
0
I-
-J
0
>
2
U)
LU
I-
U)
>.
U)
I-
2
LU
a
LU
0.
SYSTEM SIZE
CATEG DRY:
13.6
‘ 1
>.
0
0
LU
I-
4
a
LU
N
U)
I
a
4
LU
0
U.
2
0
I-
4
-J
0
>
z
U)
LU
>.
U)
I-
2
LU
a
LU
0.
15 —
10
5
SMALL
VERY SMALL MEDIUM LARGE VERY
CWS* IN EACH
SIZE CATEGORY:
LARGE
-J
INTERMITTENT:
SMALL
232
PERSISTENT: 22 227 236 1,114
-------
FIGURE 6
OMPL lANCE
NATIONAL MICROBIOLOGICAL COMPLIANCE
FY 80 MICRO-MCI COMPLIANCE
7% (INTERMITTENT)
3% (PERSISTENT)
FY 81 MICRO-MCI COMPLIANCE
COMPLIANCE
7.6% (INTERMITTENT)
0.9% (PERSISTENT)
-------
FIGURE 7
NATIONAL TURBIDITY COMPLIANCE
6% (INTERMITTENT)
11% (PERSISTENT)
COMPLIANCE
7.2% (INTERMITTENT)
(PERSISTENT)
COMPLIANCE
FY80
FY81
-------
14
We are learning more and more about individual pathogens, which ones are the important ones
to be concerned about at the national level, that can be monitored for or for which surrogates can
be identif led. And what about the coliform standard which has been an extremely effective tool
that has been applied for many years -- is it adequate? What are its failings? How can it be
improved? There must be a re-examination of the concepts of monitoring frequencies, monitoring
locations, factors that influence the amount and type of monitoring that should be done in large
systems and medium systems, and small systems, particularly where finances can become a
problem and where available expertise becomes less accessible. Then there is turbidity and its
relationship to risks associated with drinking water quality, whether it be particular organisms or
aggregates of organisms. At what turbidity level can a relationship be demonstrated? To what
point can we reasonably extrapolate to assure protection?
Ultimately as all of the figures indicate, the vast majority of problems are in the small
communities. Are there some unique characteristics in those small communities that can be dealt
with scientifically and in a regulatory context? Are there some unique solutions that can be
proposed through both of those contexts?
We have learned a lot in the few years that the Safe Drinking Water Act has been in
existence and in the fewer years that the standards have been in existence, and the compliance
problem is only part of it. A number of other significant issues have arisen; public notification and
circumstances when it is appropriate and when it is not appropriate. What amount of flexibility
ought to be available for states, communities, health departments, in applying regulations? My
feeling is that we must achieve a marriage between both these minimum national requirements and
the capacity of local officials to exercise expert judgment to assure general public health
protection. We must build in as much flexibility as possible into the application of these
requirements so that the states and communities can be guided to particular circumstances where
additional monitoring is necessary to perhaps where less monitoring is acceptable. I am not
referring to standards that would be different in different communities but rather judicially
applying standards that take into consideration the local practical and economic factors that must
always be taken into consideration when one is trying to deal with such a great number of diverse
circumstances.
-------
15
EPA has been considering several ways to restructure standards to maximize operational
flexibility not only for microbiology standards but in all revised standards in the Safe Drinking
Water Act. One method would be kind of a multi—tiered approach that would establish several
categories of standards or provide a series of rules, guidelines, and advice that could be applied
judicially by the states rather than just a single standard that applies to everybody. As an example,
in a three tiered system we could have in the first tier those controls which are absolutely
unequivocally of national concern. Virtually every public water system must take into considera-
tion coliforms or some other measurements that relate to biological quality -- perhaps turbidity,
perhaps many others.
In the second tier could be placed contaminants of more localized significance - or those
which occur only under specific identifiable circumstances. If those conditions do not exist, there
would be no particular obligation placed upon the community. In other words, the community
would make this judgment based upon the facts.
The third tier could be more in the sense of guidance that would go beyond what would be the
essentials of national standards or these localized applications of standards. This guidance could
deal with the things to do when dealing with contaminants that are not regulated, or optimal
approaches to identify the presence of contamination, or to identify acceptable daily intake levels
or health advisories for short exposure to chemicals.
I think this three tier idea can be applied in the microbiology area; I am sure that it can be
applied in the chemicals area, arid this is one of the things we would like you to think about. What
are the absolute essentials that should apply universally? What are other important protective
measures that should apply under certain conditions that may not be universal conditions? What
additional information, guidance, or advice should be provided that will be useful to those states
and communities that are in a position to do more or have a need to do more or which are
encountering problems which were not dealt with explicitly before?
I hope what we are asking of you does not sound too oppressive, but I am sure there is no
better group that could hove been gathered to discuss these issues and advise us. The product of
this workshop, which is the First step in the implementation of the Revised National Drinking Water
Standards, is absolutely esses’tial to our activities in EPA’s Drinking Water Office. It will lead very
directly to an Advance Notice of Proposed Rulemaking, which will discuss these principles for a
-------
16
wider audience to consider. It would then lead eventually to a proposal in the Federal Register and
the explicit regulatory administrative process to ultimately decide upon that mix of national
regulations, guidance, advice, and local options. The object is to assure the quality of drinking
water, wherever it is in the United States. It is not to impose unnecessary requirements, it is not
to require monitoring for the sake of monitoring or MCL’s for the sake of MCL’s, it is to streamline
the national regulctory requirements so that they will be useful tools that the states and
communities will use to assure the safety of water for their citizens at the right price, i.e.,
reasonable cost and practicality.
So again, welcome; I think what comes out of this conference will not only accomplish what I
have just mentioned for the U.S., but will probably be an essential element in the way the rest of
the world deals with biological problems. I am sure the follow-on from what happens here and in
those national regulations will have a substantial impact of what happens in Europe and the rest of
the world, so I think the incentives are great. It is very unlikely that anything like this will happen
again in the next ten years or more so it is a matter of giving it all our best shot right now with the
expectation that the next opportunity will be in the distant future. The significance of what we
are doing cannot be overstated.
-------
KEYNOTE ADDRESS
Dr. Riley D. Housewright*
I am delighted to be here this morning. I must compliment the EPA on their solicitation of
advice from scientific and professional organizations. It is comforting to know that within the
federal structure there are organizations that consult professionals in the areas they are regu-
lating. This shows an enlightened attitude that is not always found in Washington. This is a great
opportunity for you to participate and make a real contribution to water microbiology regulations.
All too often, scientific and professional considerations are neglected in the name of expediency or
some other less valid reason. The American Society for Microbiology, with its 32,000 members, is
an organization with the expertise required for the solution to many of the problems that the
federal government is used to solve. A number of you are here today as a result of suggestions we
have made regarding qualified participants.
It will be my purpose this morning to direct your attention to some ignored, unsolved,
neglected, and possibly some overworked problems associated with the detection and identification
of human pothogens in drinking water. Our task really is simply defined, but not easily
accomplished, It is to detect and identify and quantify human pathogens in drinking water and to
take whatever action is appropriate. Detection and identification are the parts of the problem that
concern us this morning. The list of waterborne pathogens is a long one. It contains
representatives from practically every major group of microorganisms — bacteria, viruses, and
protozoa. Little information is available on mycoplasma, the pathogenic yeast and pathogenic
fungi, that are occasionally found in drinking water. In order to put our problem into perspective,
what we need to do is establish first of all, how many of these organisms are required to cause
illnesses in humans. This is a fundamental matter because the number varies a great deal from one
pathogen to another. These numbers determine how sensitive our methods must be to detect the
organisms causing disease in humans. The best quantitat ion of the numbers required to infect man
comes from controlled human exposures during the evaluation of vaccines for enteric pathogens.
Extensive information is available on two of the organisms that concern us: the Salmonellae and
Executive Director, American Society for Microbiology, Washington, D.C.
-------
2
Shigellae. The number of Salmonella required to infect 50 percent of the people exposed to it is
about 10 million organisms. On the other hand, certain strains of Shigella give the same results
with an infective dose of 10 to 100 organisms. Results similar to those obtained with Shigella are
claimed with certain viruses, but not everybody agrees with this conclusion. So whether the
number is one or it is 10 to 100, we know that the detection method must be sensitive. The direct
quantitotion of pathogens is not a simple matter in such certain circumstances. These pathogens
have been isolated from reservoirs, rivers, groundwaters, and such sources. However, the detection
of these pathogens in process and disinfected water is more difficult. The scientific literature has
presented a vast array of media for the direct detection of these organisms in finished water.
Proposed are modifications of media, time, temperatures, selective inhibitors, and so on. There is
no single procedure that is useful for the isolation or identification and quantitation of all these
pathogens. Only for the Salmonella are the available procedures sufficiently accurate. Methods
for most of the other major pathogens such as Shigella, Vibrio , and Leptospira are Inadequate. In
summary we may say that the procedure for direct detection of these pathogens quantitatively Qnd
the Isolation of them In small numbers is not satisfactory. You should be aware of opinions that
have been expressed by a variety of persons having a wide variety of talents and expertise in these
areas. I plan to mention several of these possibilities. You will then have an opportunity to either
agree or disagree with them. I hope in your sessions you will take the opportunity to either be on
advocate for a particular position or to discard once and for all, some of those half truths that have
been advocated. What I hope to do is to raise enough questions in your mind that you can formulate
an opinion. You should give very careful consideration to such matters as whether or not one
should attempt to detect the pathogen itself or whether we should be concerned only with indicator
organisms.
An indicator organism, as used in water microbiology, means a microorganism whose presence
is evidence that pollution has occurred. This usually is in association with fecal contamination
from man or other warm blooded animals. An indicator organism may be accompanied by
pathogens but it does not necessarily cause disease. The pathogens usually appear in smaller
numbers than indicator organisms and are therefore less likely to be isolated. Indicator organisms
should have the following characteristics: I) they should be applicable to all types of water;
2) they sho.uJd be present in sewage qndpolh4ed wqter when pathogen4 we present; 3) the numt r
-------
3
should be correlated with the amount of pollution found in the water; 4) they should be present in
greater numbers than the pathogens, therefore, more easily detected; 5) there should be no after
growth in the water; 6) they should hove a survival time that is greater than the pathogen; 7) they
should be absent from unpolluted waters; 8) they should be easily detected by simple laboratory
tests in the shortest time consistent with accurate results; 9) they should have constant char-
acteristics; and 10) ideally they should be harmless both to man and animals. Now, no group of
organisms meets all of these criteria. There is a long list of those that have been proposed as
indicator organisms: just to name a few, E. coli, total coliforms, coliphoge, Clostridium
per fringens, Pseudornonas aerug inosa, V ibr io, Staphylococcus (coagulase positive), Bifidobacterium,
Cundidci albicons and other yeasts, fecal streptococci, Yersiriia , and acid-fast organisms. I would
like to list for you some recommendations and comments made regarding some of these proposed
indicators just to give you some idea of the variety and to assure that we not limit our thinking to
E. coil or coliform organisms. About Clostridium perfringens , it was said it appears to be the
fecal indicator of choice for measuring remote or intermittent pollution and in situations where
resistance to disinfectants is at a premium. It is one of the indicators of choice according to this
source. On Pseudomonas aeruginosa , demonstration of this species in surface water suggests the
influence of man and its number reflects the degree of pollution. This source, however, did admit
that there was little relationship between its population and other pathogens or feca indicators.
Regarding Vibrio species, it has been suggested as indicators of water quality. Regarding
Staphylococcus , the major obstacle to its use as on indicator organism is the lack of a selective
medium. Candida albicans is described as a most reasonable candidate. Enterococci and fecol
streptococci may not prove to be ideal but may be advised under certain circumstances. The
presence of KIebsiella in water indicates degraded quality and it is probably as significant a finding
as E. coil and a good indicator of pollution from certain organic wastes.
This gives you some idea of the diversity of opinion regarding the use of indicators when a tot
of people thought this problem was solved a long time ago by the use in Europe of E. coil Itself or
collform organisms in this country. Each of these organisms obviously has qualities that have
recommended it to those people who advocate them, but none meet afl of the criteria. It has been
emphasized by McCabe and a number of us at the Notional Academy of Sciences that the presence
of these indicators should relate directly to the risk of the disease (that Is, to the presence of
human pathogeris).
-------
4
There have been one or two studies that have made a serious attempt to correlate the
presence of indicator organisms with the incidences of disease. Unfortunately, most of them are
rather old. We really have to go back to the studies of Kehr and Butterfield who correlated
coliform count directly with disease incidence in 1943. They concluded that the excretion rate of
coliforms would be the same in a healthy population as in a sick population, but that the latter
would also excrete typhoid bacteria. They then compared the stability of E. coli and Salmonella
typhosa and found that this ratio was quite consistent, as they both died off at approximately the
same rate.
This brings us to E. coli or the total coliform group. I will not try to list their virtues. You
know what they are, but they have been criticized in several respects. One of them has to do with
the die-off of these organisms. Some report that Salmonella die off at a slower rate than E. coli,
thus resulting in false negative test results. It has been known for many years that atypical lactose
reactions also lead to false negative results. One of the problems here that is foremost, and has
been given less attention by the general public as well as the regulatory groups, is the interference
in this test caused by other waterborne organisms. There has been a very clear demonstration by
Ed Geldreich and his group, as well as a number of others, that the presence of other organisms in
large numbers very definitely interferes with the recovery of the coliforms. There is no time to go
into the problem of enteric viruses, but again the presence of coliforms do not, in my opinion,
correlate well with the presence of enteric bacteria. I was interested in another statement Joe
Cotruvo made earlier in this session, regarding the lack of filtration of drinking water in certain
parts of the country, particularly some high population areas. Giardia would not be removed in
these instances and obviously may not be called to our attention by the coliform indicator tests.
A real time and continuous monitoring technique is needed. Some methods have been
proposed for this, including radioactivity and enzymatic tests. But, it is premature for us to
consider any of those unless there are developments, of which I am unaware. The use of chlorine
determinations as a substitute for the coliforms test has been proposed, and it has been said that a
substantial reason for abandoning the coliform test, and its use in the past, has been that the test
really was not done by plant operators. It was found in a national survey that 85 percent of the
water systems were not collecting samples at the prescribed rate. I wonder if it really makes any
difference whether the samples are not collected for doing coliform tests, or they are not collected
-------
5
for doing chlorine tests. This finding says a lot more about the training of the employees than it
does about the validity of these tests. I am particularly concerned about the substitution in this
case, because of so many studies that have found microorganisms (col iforms and a number of
others) in the distribution system where there were extremely high levels of chlorine. I am
Impressed with the idea of getting an online, real time method, but it seems to me that because a
bacteriological test is a little more difficult to run, is not a valid excuse for substituting a less
decisive test. I would like to close now with a remark or two that comes from people in one of the
largest, if not the largest water plant, in this country. They have said, “It seems logical to retain
the total coliforms as the indicators, restrictions on maximum turbidity units, and maximum free
chlorine residuals.”
I am really disappointed to see the results that Joe Cotruvo reported this morning on the
35 percent noncompliance since 1978. First samples must be collected and analyzed. Too often
statistics are collected that are not meanin ful because the test simply is not done or it is done
Incorrectly. It’s a little bit like one of the old time baseball players used to say about his pitching.
I -Ic said, “Ain’t nothing going to happen until I throw the ball.”: That’s about the way it is when it
comes to taking samples; until you take the sample and do something with it, there’s not much
going to happen regardless of chlorine determination or any other kind of determination. There
still ore a lot of questions that need answering about the indicator systems. I hope that the
questions raised this morning will be sufficiently provocative to stimulate ilvely discussions and
creative thought processes In their solution.
-------
WATERBORNE DISEASE OUTBREAKS IN THE UNITED STATES
AND THE COLIFORM STANDARD
Gunther F. Craun* and Leland J. McCabe**
There are many reasons why we have the Safe Drinking Water Act and the resulting
Federal/State program. It’s hard to identify the reason(s) it was finally passed, but every
Congressman who was in favor of the legislation quoted the statistics summarized by Gunther
Craun and me showing that 130 outbreaks of waterborne disease and 46,374 cases of illness
occurred in the United States between 1961 and 1970. It’s pretty ominous to think you are
responsible for those numbers. People seem to feel that we should be able to provide a drinking
water that does not make people sick. Concerns about the effects of low-level chronic exposure to
chemical contaminants in drinking water, and Joe Cutruvo alluded to this, seem to have developed
since the passage of the Safe Drinking Water Act. Granted, the situation where organic chemicals
were discovered in the drinking water source for New Orleans had something to do with getting
President Ford to sign the bill, but the Congressional concerns, or at least the quoted concerns,
were based on the outbreaks that had occurred In that ten-year period.
It Is significant that we are meeting at the Airlie Conference Center because here occurred
one of the outbreaks that are Included in this period. If you stayed at the Groves Cottage or the
Carriage House, you slept in a very historic site. The Washington Post on May 2, 1967, carried the
story: “AIrlIe hunts for the bug that made 60 III.” The attack rate was 83 percent in the first
group that was investigated, and there were two following meeting groups that had excess illness.
It was later found that four previous meeting groups also hod excess illness, which had never been
brought to anyone’s attention. Periodic water sampling by the County Health Department had
always shown the coliform counts at Airlie to be within the limits, but the samples were always
collected from this particular building in which we are meeting. Several wells supply water to the
complex. The water supply for this building was obtained from a good well, but during the
epidemiological investigation, the two wells serving the Groves Cottage and Carriage House were
found to be contaminated. One was near a septic tank tile field and the other drew water from the
creek. It might be interesting to have the manager show us where these two wells were located at
that time and what corrections were made. Stools samples from the individuals and water samples
Sanit T er, Toxicology and Microbiology Division, Health Effects Research Laboratory,
U.S. EPA, Cincinnati, Ohio.
**Science Advisor, Toxicology and Microbiology Division, Health Effects Research Laboratory,
U.S. EPA, Cincinnati, Ohio.
-------
2
from the contaminated wells yielded enteropathogenic E. calf (Oil l:B4). That was the first time
that organism had been responsible for an adult disease in a woterborne outbreak in this country.
The article in Loncet outlining the outbreak investigation suggested that the best way to prevent
disease outbreaks is to investigate them. Routine coliform monitoring of the drinking water in this
particular situation certainly didn’t help in preventing the outbreak.
Our first review of wuterborne disease outbreaks was from I 946 to 1960. We continued the
same summary as Abe Wolman who reviewed outbreaks earlier in the century. The historical
epidemiologiccil data that was used to support the Safe Drinking Water Act came from an article in
the Journal of the American Water Works Association in January 1973, in which Gunther Craun and
I paid homage to John Snow. We had a cartoon from the London Times on Snow’s study of the
Southwork Water Supply when he investigated an outbreak of cholera which was shown to be
waterborne. His work was facilitated because people on the same street were served by alternate
water systems. Since then water utility managers haven’t given epidemiologists the privilege of
studying water systems where alternate houses are served by different water systems. Usually the
distribution system is so complex that you really can’t tell who drinks what water, especially if
different sources ore used. Back then, in the late 60’s and early 70’s, we reported the occurrence
of about two waterborne outbreaks a month. We’ve continued this activity through the 1970’s and
in 1980 we can report the occurrence of about four outbreaks per month, about twice as many
outbreaks as before the Safe Drinking Water Act. In 1980, we can also report about 20,000 cases of
wciferborrie illness. It is hard to tell whether more outbreaks are occurring or the increase in
outbreaks is due to better reporting.
We have tabulated and investigated waterborne disease outbreaks in cooperation with the
Centers for Disease Control since the early I 970’s. Ed Lippy has been in charge of this particular
activity for us the past several years and has begun some cooperative studies with Washington
State, Colorado, and Vermont to determine how many waterborne outbreaks occur when you
dramatically increase outbreak surveillance at the local/State level. We are finding two to three
times the number of outbreaks than we previously reported in these States; this is about what we
predicted in our review of the data for 196 1-70.
During 1971-79, 265 outbreaks of waterborne disease affecting 57,913 persons were reported
by 45 states and Puerto Rico (Table I). Most of the outbreaks (5 14 percent) occurred in
-------
3
Table I. Waterborne Disease Outbreaks in the United States, 1971-1979
Year
Outbreaks
Cases of Illness
1971
19
5,182
1972
29
1,638
1973
26
1,774
1974
25
8,356
1975
24
10,879
1976
35
5,068
1977
34
3,860
1978
32
11,435
1979
41
9,720
265
57,913
-------
4
noncommunity water systems, which serve a large transitory population (Table 2). A large number
(78 percent) of the outbreaks in noncommunity water systems affected travelers, campers,
restaurant patrons, and visitors to recreational areas; of the outbreaks involving this transitory
population, 74 percent occurred from May to August. There are approximately (50,000 non-
community water systems in the United States most of which (96 percent) use groundwater sources.
Most of the illness (69 percent) resulted from outbreaks in community water systems, which
number about 64,000 and serve some 200 million people. Twelve percent of the reported
waterborne outbreaks and I percent of the reported illness occurred in individual water systems,
which depend primarily on untreated groundwater; however, outbreaks in individual water systems
are unlikely to be recognized and reported.
Treatment deficiencies, such as inadequate or ineffective pretreatment and filtration and
interrupted disinfection, were responsible for 36 percent of the outbreaks and 50 percent of the
illness in community water systems. Contamination of the distribution system, primarily through
cross connections and backsiphonoge, also resulted In a large number of outbreaks (34 percent) and
Illness (24 percent) in community systems. in noncommunlty water systems, the use of untreated,
contaminated groundwater accounted for many of the outbreaks (44 percent) and much of the
illness (41 percent); only I I percent of the outbreaks and 2 percent of the illness in community
water systems were caused by the use of untreated, contaminated groundwater (Tables 3 and 4).
More outbreaks occurred in groundwater than surface water systems. The overflow or
seepage of sewage, primarily from seplic tanks or cesspools, was responsible for 43 percent of the
outbreaks and 63 percent of the illness caused by the use of untreated, contaminated groundwater.
Chemical contamination of groundwater (arsenic, ethyl acrylate, leaded gasoline, nitrate, phenol,
polychlorinated blphenyl, selenium, and waste oil) and contamination by surface runoff or flooding
caused 19 percent of the outbreaks in systems using untreated groundwater.
Groundwater systems usually depend on a source water of relatively good bacteriologic
qualify where disinfection is used to protect against possible contamination of the distribution
system. In these situations, unexpected cr ntamInatIan of the source could overwhelm the
disinfection. For groundwater systems that use a source known to be frequently or intermittently
contaminated with bacteria, continuous disinfeclion Is necessary to ensure potability until the
sources of contamination are located and removed, Improvements are made In source protection, or
-------
5
Table 2. Waterborne Disease Outbreaks in the United States, 1971-1979
Type of
Water System
Outbreaks
(Percent)
Cases of Illness
(Percent)
Municipal
34
69
Semipublic
54
30
Individual
12
i
100
100
-------
6
Table 3. Deficiencies Responsible for Waterborne Disease Outbreaks in Community
Water Systems in the United States, 1971- 1979
Outbreaks illness
(Percent) (Percent)
Contaminated Surface Water, No Treatment* 13 23
Contaminated Groundwater, No Treatment I I 2
Deficiencies in Treatment Facilities 36 50
Deficiencies in Distribution System 34 24
Miscellaneous
100 100
*Includes outbreaks of giardiosis in which surface water was chlorinated, but not filtered.
-------
7
Table L• Deficiencies Responsible for Waterborne Disease Outbreaks in
Noncommunity Water Systems in the United States, 197 1-1979
Outbreaks Illness
(Percent) (Percent)
Contaminated Surface Water, No Treat ment* 8 2
Contaminated Groundwater, No Treatment 44 41
Deficiencies in Treatment FacilitIes 36 39
Deficiencies in Distribution System 17
Miscellaneous
too 100
*lnaludes outbreaks of glardlasls In which surface water was chlorinated, but not filtered.
-------
8
alternative wafer sources are developed. Outbreaks were caused by both inadequate disinfection
and interruption of the disinfection process (equipment malfunction or insufficient disinfectant).
An etiologic agent was identified in 145 percent of the outbreaks (Table 5). The remaining
outbreaks were classified as acute gastrointestinal illness, characterized by symptoms such as
abdominal cramps, nausea, vomiting, and diarrhea occurring 214 to 148 hours after consumption of
water. Five of the six outbreaks of acute gastroenteritis where a viral etiology was identified,
occurred in groundwater systems where either no disinfection was provided or disinfection was
interrupted or inadequate. The sixth outbreak involved a groundwater system but occurred due to a
cross connection. The single outbreak of Campylobacter gastroenteritis and the majority of
waterborne outbreaks of giardiasis occurred in surface water systems.
Establishing a maximum contaminant level (MCL) for biological agents in drinking water
requires knowledge of the pathogens that are important causes of woterborne disease. For over
half of the outbreaks that occur, we do not know the etiologic agent. Certainly the panel on
etiologic agents will have plenty to discuss. What do you do in this situation? An MCL should
protect against waterborne disease, but if we don’t know all of the agents that re responsible for
waterborne disease, how can we be certain the coliform test is appropriate as an MCL. Simple
disinfection as the only treatment for surface wafer sources has also been shown to be ineffective
in preventing the waterborne transmission of giardiasis, arid all surface waters should receive
pretreatment and filtration in addition to disinfection. Outbreaks have occurred in filtered water
systems but have been related to lack of pretreatment of filtration in addition to disinfection, poor
design, and poor operation. Giardia cysts can be reduced dramatically by properly functioning
conventional sand filters but the water must be effectively pretreated prior to filtration. Gravity
and pressure sand filters have also been ineffective in removing Glardia cysts under conditions of
poor operation but have achieved removal of turbidity. It cannot be assumed that simply meeting
the turbidity limit is sufficient to prevent outbreaks of glardiasis. The usefulness of routine
coliform surveillance is also questioned because of these giardiasis outbreaks. Waterborne
outbreaks of gicirdiasis and other waterborne disease have occurred In water systems where
coliforms have either not been detected or have not been found to exceed the MCL.
-------
9
Table 5. Etiology of Waterborne Disease in the United States, 197 1-1979
Outbreaks
(Percent)
Acute Gastrointestinal Illness 55
Chemical Poisoning I I
Giardiasis II
Shige llosis 8
Hepatitis A 6
Salmonellosis 3
Viral Gastroenterltis 2
TyphoId 2
Toxigenic E. coil Gastroenterilis
Campylobocter Gastroenteritis I
100
-------
l0
Because of the number of outbreaks caused by treatment deficiencies, especially interrupted
disinfection, it would seem appropriate to use a monitoring system other than one that detects the
presence of bacteria to let us know that the treatment is not optimal. It would seem that the most
logical test would be to determine chlorine residuals in the system. Immediately, the operator
could adjust the chlorine feed or make other corrections, once he determines the absence of, or
lowering of, the normal chlorine residual. Many times, the most obvious solution to a problem is
overlooked when our thinking is only one dimensional. Contamination of the distribution system
through cross connections and backsiphoncige can be minimized by control programs but never
entirely eliminated. It does not make sense to think that routine coliform monitoring will prevent
disease caused by contamination that occurs in the distribution system or that potential cross-
connections can be monitored by coliform surveillance. Again, maintenance of a proper residual
will help neutralize bacterial contamination from cross connections and also alert the operators of a
potential problem long before results of a coliform test are received.
Outbreak data indicate that increased surveillance of small water supplies is needed to
reduce waterborne disease. Increased emphasis must be placed on protection of groundwater
sources from both chemical and biological contamination for small water systems and systems used
on a seasonal basis. Where disinfection of groundwater supplies is indicated by sanitary survey,
proper operation is necessary to ensure continuous, effective disinfection is maintained. For these
small water supp’ies, collection of one coliform sample per quarter or per month is not monitoring,
and we don’t see how small systems can afford to collect a sufficient number of coliform samples
to really say we are monitoring. At the minimum, daily sampling must be conducted. To address
the problem in a cost-effective manner will require a sanitary survey to define sources of con-
tamination, sampling of the untreated water over a period of time in order to recommend proper
treatment, reliable operation of the recommended treatment, and periodic surveys to detect
changes in sources of contamination and deficiencies in treatment. This is the paradox - the larger
systems which generally have satisfactory treatment and operation must collect more coliform
samples than the smaller systems that generally have unreliable treatment and operation. It w9uld
seem that the opposite should be the case.
The American Water Works Association has a committee on waterborne diseases and the
committee just released their report, which should be of Interest. The committee feels that
-------
II
surveillance systems should be structured to help prevent woterborne disease. Their conclusions on
the usefulness of the coliform test in preventing waterborne disease are of interest:
“The committee feels that it is more important to know the quality of the source water
and potential sources of contamination so that source protection and treatment can be
provided. A water supply surveillance program should emphasize frequent engineering
evaluation and sanitary surveys to identify and correct potential deficiencies. Micro-
biological resources are better applied to assessing raw water quality, identifying sources of
contamination, and evaluating the efficiency of treatment than the routine surveillance of
water quality in the distribution system. In some instances, alternative microbiological or
chemical indicators, such as chlorine residual analysis, may be more useful in preventing
outbreaks than coliform surveillance of the distribution system.”
“Routine coliform surveillance of treated water is mandated by the Safe Drinking Water
Act; however, the committee feels that its importance is to provide a historical record of
operation rather than to prevent waterborne outbreaks.”
For the previous symposium on the evaluation of the microbiology standards for drinking
water chaired in 978 by Chuck Hendricks, Gunther Craun reviewed the bacteriological records of
water supplies that experienced outbreaks. Although this was more of a case history study, it did
reveal situations where outbreaks occurred in water systems that met the coliform MCL. Routine
bacteriologic surveillance was shown to be of little use in preventing outbreaks caused by use of
untreated groundwater or surface water, distribution deficiencies, and treatment deficiencies.
Important factors were: (I) length of time between sample collection/analysis and recognition that
the positive result represented a real problem; (2) false sense of security provided by the few
distribution system samples collected from small water systems using untreated ground or surface
waters; (3) little or no knowledge of bacterial water quality of water sources and thus little
incentive to maintain continuous disinfection; and (4) too few samples collected to detect sporadic
occurrences of contamination (e.g., water systems contaminated during a lengthy time interval
between collection of samples). A more thorough study comparing routine coliform results from
communities experiencing an outbreak with communities not experiencing an outbreak
was completed in 1979 by Odette Batik and Gunther Craun; Wes Pipes conducted a similar study of
non-community water systems in Pennsylvania We found it very hard to get the coliform records
-------
from the communities. Some States just didn’t keep their records aroung long enough or the States
wouldn’t give the records to us. Many of the outbreaks occurred before passage of the Safe
Drinking Water Act and the study probably ought to be done again now that we have increased
monitoring and reporting activities. in general, we found no differences between water systems
experiencing an outbreak and those not experiencing an outbreak in terms of whether they met the
coliform MCL. These results will be published shortly (Batik, et aL, J. of Environ. Health,
March-April, t983).
in closing, both Gunther Craun and I stick by our recommendations of the previous symposium
chaired by Chuck Hendricks. We subscribe to the philosophy of placing the burden of preventing an
outbreak on the operator with frequent sanitary surveys by State engineer sanitarians. This
removes the mystique of collecting a sample, mailing it to a Stale Laboratory, and hearing weeks
later whether the water quality was satisfactory or unsatisfactory. The typical response of the
operator is frustration, as he is chided for having done something wrong way back then but is In no
position to remedy the situation. If he performed daily chlorine residuals, he would know
immediately when there was a problem and possibly take corrective action to prevent on outbreak.
Routine bacteriological surveillance of wafer distribution systems for coliforms, such as
proposed by the Standards, is satisfactory for historical purposes but is of little value in preventing
the occurrence of waterborne outbreaks. Results of such sampling cannot be received in time for
corrective action to be taken. The length of time required for analysis and the time delay between
collection and receiving results is much too long to enable the coiiforrn test to be used for routine
quality control purposes. The idea of cheek samples also adds to the time delay, since the first
positive sample is generally assumed to be a sampling error rather than real contaminatIon. It also
Is not logical to permit a lower frequency of sampling for small systems. The ideal quality control
program should offer continuous or frequent monitoring that provides an Indication of quality
within a relatively short time after sampling. In addition to these points, there are instances where
negative bacteriological results are misleading and meaningless.
We propose, instead, an alternative monitoring program that would allow our limited
microbiological resources to concentrate on research activities, such as identification of etlologic
agents responsible for outbreaks of gastrointestinal illness of unknown etiology. The emphasis of
this alternative program would be on engineering evaluations or sanitary surveys. A thorough
-------
‘3
engineering evaluation would include bacteriological monitoring of source water rather than water
in the distribution system. It is necessary to know the quality of source water and potential
sources of contamination so that appropriate treatment can be recommended. The emphasis then
must be on providing and maintaining this treatment, especially continuous disinfection at adequate
concentrations. For product quality control, a simple test such as chlorine residual can be used,
Of course, chlorine residual must be defined by an engineer for each system so that all factors of
effective disinfection are included, such as time, pH, temperature, species of chlorine, turbidity,
and representative sampling points and frequency of collection must be determined. Bacteri-
ological surveillance of the distribution system should be used to check the validity of chlorine
residual monitoring, but this can be done at a lower frequency. This would also provide a historical
record which many seem to feel is important. If a disinfectant other than chlorine is used, then
disinfection markers/residuals should be monitored for these chemicals. If supplies are not
disinfected, chlorine residual measurements would, of course, not work; however, it is anticipated
that a minimum number of such supplies would exist.
The real importance of measuring chlorine residual is that the primary responsibility for
taking corrective action is placed where it should be - with the operator. Time is of the essence in
preventing an outbreak from occurring and the operator must be the first to know if something is
amiss, not the last to know. He can provide the quickest and simplest response simply by adjusting
the chlorinator.
-------
EMERGING MICROBIOLOGICAL ISSUES
Edwin E. Geldreich*
INTRODUCTION
The potential consequences of microbial contamination in drinking water have always been of
paramount importance to both the supplier and consumer. For a variety of reasons, new
microbiological concerns in wafer supply are emerging that can be traced to source water quality,
treatment practices, distribution systems and monitoring data reliability. These concerns do not
reflect on the adequacy of conventional water treatment when skillfully applied by trained
operators in water plants designed to cope with both the changing qualities of source water and the
peak water supply needs of the population served. Unfortunately, the continued application of
marginal treatment to source waters receiving increasing pollutional loadings is an alarming trend
particularly in light of decreasing capacity of streams to assimilate increasing waste discharges
from expanding populations on the watershed, and in view of relaced State requirements for
disinfection of wastewater effluents. Sediment accumulations in distribution systems often
compound treatment problems by becoming sites for microbial colonization that ultimately cause
water quality deterioration at the consumers tap. Finally, monitoring problems including sampling
logistics and methodology limitations are definite factors affecting the ability to detect stressed
coliforms or pathogens such as Yersinia, Campylobacter , and Giardia .
SOURCE WATER QUALITY
A number of factors enter into the choice of a best available raw source water for drinking
water, including adequate quantity during seasonal variations in flow, water quality that Is
amenable to treatment and some measure of watershed protection from domestic, industrial, and
agricultural pollution (Office of Drinking Water, 1978). A series of microbial barriers are of
particular importance In watershed water quality management and Involve application of poInt
source wostewater treatment technology, natural self-purification processes In surface waters and
water treatment engineering to produce a safe drinking water supply (Swoyne, et al., 1980).
Unfortunately, controlling activities on the watershed through land ownership, State regulations or
local ordinances is becoming more difficult because of pressures from various groups.
*Chief, Micr R logical Treatment Branch, Water Supply Research Division, Municipal Environ-
mental Research Laboratory, U.S. EPA, Cincinnati, Ohio.
-------
2
Recently several States have proposed suspension of sewage effluent disinfection for waste
treatment plans more than 20 to 40 miles upstream from a water supply intake (Huff, 1981). The
argument is that chlorination is not cost effective for controlling fecal coliform densities in
sewage effluents. Furthermore, based on bacteriological quality studies, it is argued that quality
at downstream water supply intakes has not improved substantially as a direct result of sewage
effluent disinfection practices. The crux of the problem is that chlorination of sewage effluents
would be more effective if disinfection demand were reduced through effective control of
Biological Oxygen Demand (BOD), total suspended solids and ammonia levels in sewage treatment.
Since many plants have difficulty in consistently achieving limits below 20 mg/I BOD and 30 mg/I
suspended solids, operators have tended to over-chlorinate.
Receiving waters between the point discharge of sewage effluents and water intakes not only
serve as a buffer to accidental spills and treatment bypasses but can contribute to water quality
improvements. Stream self-purification is a delicate blend of complex and poorly defined
processes that involve bacterial adsorption with sedimentation, predation, dilution, water tempera-
ture and solar radiation (Frost and Streeter, 1924; Kittrell and Furfori, 1963; Shilo and Bruff, 1965;
Wuhrmcn, 1972; and Hendricks, 1974). While natural self-purification may be effective in a given
water course during periods of warm temperature and limited precipitation, results are usually less
impressive following wet weather periods and particularly poor during the cooler seasons.
Surface waters surrounding metropolitan areas are generally polluted with the wastes from
these concentrated centers of human activities. To provide maximum protection to water quality,
intakes for metropolitan water supplies were originally located miles upstream from the discharge
of domestic wastes and industrial effluents. This approach permitted receiving waters to serve as
a buffer zone and provide some measure of stream self-purification to effect water quality
improvement. In recent years the development of satellite residential areas and the rapid growth
of cities have largely reduced the effectiveness of this natural self-purification zone and brought
new sources of sewage and stormwaler runoff to degrade the source water quality. For example,
there are 56 water supply intakes along the 97.8 miles of the Ohio River and 43 of these are within
five miles of an upstream wastewater treatment plant effluent. While low flow conditions are
minimized by navigational dams on this river, other surface water sources are not so protected. In
a study of surface water supplies used by 20 cities serving a total population of seven million
-------
3
people, it was estimated that the wastewater component of the source water ranged from 2.3 to
16 percent and increased to predominantly wastewater for several municipal intakes during tow
flow periods (Swayne, et al., 1980).
Bacteriological examinations of raw water quality at the Omaha, St. Joseph and Kansas City
water treatment plant intakes on the Missouri River (Table I) frequently revealed fecal coliform
densities in excess of 2,000 organisms per 100 ml (U.S. EPA, 1971). This fecal pollution load Is the
end product of varying inputs of raw sewage, effluents from primary and secondary treatment
plants of differing efficiencies, cattle feedlot runoff, and discharge from meat and poultry
processing plants. The implication is that pathogens ore also present, their numbers and kinds
being related to the diseases prevalent in upstream human and farm animal populations. To
demonstrate this point, various serotypes of Salmonella were isolated at each water plant Intake.
Virus examinations performed on raw water from the Intake at St. Joseph resulted In the reported
recovery of poliovirus types 2 and 3 and ECHO virus types 7 and 33.
In another study done on Missouri River water quality at public water intakes, total coliform,
fecal coliform, virus and turbidity levels were quantified at the Lexington intake during 1976-78
(O’Connor, 1981). Data In Table 2 were grouped by results obtained during water temperature
periods that closely corresponded to seasonal periods. During the winter period, the resulting fecol
contamination was derived largely from municipal wastewater effluent discharges and meat
processing plants upstream of th water supply intake. Spring thaws brought Increasing flood
conditions arid associated Increases In turbidity and movement of fecal contamination in runoff to
the main-stem portions of the river. Occasional high turbldltles in the summer were related to
major storm periods that increased river flow and turbiditles periodically to maximum values.
During wet weather periods, stormwater runoff included bypasses of raw sewage In treatment
facilities and movements of fecal wastes from numerous feedlot operations. Autumn periods
showed a decline in turbiditles, river flows and focal coliforms as a reflection of reduced rainfall
periods. The percent occurrence for enterovirus in these samples examined suggest ‘that virus
levels correlate inversely with coliform levels, temperature and turbidity. This unexpected finding
could be due to more rapid virus dieoff at higher water temperatures or more likely to decreased
recovery efficiency for viruses because of interference with virus detection methodology from the
higher stream turbidities created in stormwater runoff silts. Regardless of methodology limita-
-------
4
Table I. Fecal Coliform ensities and Pathogen Occurrence at the
Missouri River Public Water Supply lntakes*
Raw Water
Intake
River
Mile
Date
Fecal
per
Coliforms
100 ml**
Pathogen Occurrence
Omaha, NB
626 .2
Oct 7-18, 1968
Jan 20-Feb 2, 1969
Sep 8-12, 1969
Oct 9-14, 1969
Nov 3-7, 1969
8,300
4,900
2,000
3,500
1,950
N.T.***
N.T.
Salmonella enteritidis
Salmonella anatum
N T.
St.Joseph,MO
452.3
Oct 7-18, 1968
Jan 20-Feb 2, 1969
Sep 18-22, 1969
Oct 9-14, 1969
Nov 3-7, 1969
6,500
2,800
4,300
N.T.
6,500
N.T.
N.T.
N.T.
Salmonella montevideo
N.T.
Jan 22, 1970
Apr 23, 1970
N.T.
N.T.
19 virus PFLJ; Polio types
2,3: Echo types 7, 33
3 virus PFU, not yet typed
Kansas City, MO
370.5
Oct 28-Nov 8, 1968
Jan 20-Feb 2, 198 I
Sep 18-22, 1969
6,500
8,300
3,800
N.T.
N.T.
Salmonella newport
Salmonella give
Salmonella infantis
Salmonella poona
Sep 25-29, 1969
3,800
*Data from Report on Missouri River Water Quality Studies, WOO, Region VII, EPA, K.C. MO.,
Feb. 1971.
**Geomefrjc means
- No test for pathogens done
-------
5
Table 2. Raw Water Qualify at Lexington, MO Water Treatment Plant °
Total . Fecal Virus (PFU)
Temp.’ / No. Turbidity (NTU) Coliforms C Coliforms C/ Per 100 gal
Range Samples Range Per 100 mL Per 100 mL Range % Occurrence
0- 10°C
68
0.05-200
93,000
14,600
0-38
79.4
I0.5-2O°C
32
32-600
I
14,000
7,900
0-9.1
53.)
20.5-28°C
42
40-1,100
I
62,000
8,600
0-JO
42.9
(a) Data from O’Connor, Hemphill and Reach, Jr. (10)
(b) Sampling Period December 3, I 976 to December 27, 1978.
(c) Geometric Mean Values
(d) Spring and autumn periods combined
-------
6
tions, the data in Tables I and 2 illustrate some of the potential health hazards that become more
prevalent and more challenging to a water treatment system as surface water qualities fluctuate.
Impounded waters constitute the source water for approximately 1,700 municipal supplies
serving more that 55 million people, or 30.6 percent of the 180 million total population served by
public water supplies (Statistical Summary of Municipal Water Facilities in U.S., 1963). Although
storage and sedimentation of source water in a reservoir often improves the chemical (Symons, el
ai., 1970) and bacterial quality of the raw water (Dzyuban, 1975; Romaninko, 1971; and Geldreich,
et al., 1980), adverse quality changes can be introduced by stormwater runoff from rural and urban
areas (Olivieri, el al., 1977). Inflow of poor quality waters from small streams may introduce fecal
wastes from agricultural practices (Horvath, 1975; Geldreich, 1972; and Gaufin, 197k), effluents
from malfunctioning septic systems and leachates from landfill operations (Reid, 1966 and
Comptroller General of the U.S., 1978). RecreatIonal uses of reservoirs (Gaufin, 1974; King and
Mace, 1974; Lee, el al., 1970; and Council on Environmental Quality, 1975) and wildlife populations
(Hibler, et al., 1975; Aikird, et al., 1977; Kirner, et al., Kirner, el al, 1978; Alter, 1954; Ketchlkan
Laboratory Studies Disclose Gulls are Implicated in Disease Spread, 1954; and Fennell, etal., 1974)
native to the watershed may introduce addItional contamination into water supply impoundments.
Among these quality changes, pesticides, fertilizers, pathogens and turbidity released from an
agricultural watershed are of serious concern In water impoundments. Often these impoundments
are used by small water treatment systems whose limited treatment capabilities may, at times, be
inadequate to cope with high ammonia and turbidity In springtime raw source water. These
conditions may result in some collform or even pathogen penetration into the distribution system.
Groundwater Is used extensively as a source of potable water in arid regions and in areas
where surface supplies are Insufficient In volume, poor In quality and require extensive purification
and treatment. Seventy-five percent of approximately 60,000 public water supplies in the United
States use groundwater as the sole raw water source, and 7 percent of these public systems use a
mixture of ground and surface water (McCabe, et ol. , 1970). Private domestic water supplies, the
vast majority of which rely on groundwater, serve an estimated 33 million consumers. Since
groundwater has long been considered to be of unquestionably excellent quality, treatment is
frequently non-existent or is limited to water hardness reduction, taste and odor removal or simply
to disinfection.
-------
7
Groundwaters derived from deep aquifers are generally of good bacteriological quality
because vertical percolation of water through soil results in the removal of much of the microbial
and organic population. By contrast, the waters from shallow wells are frequently grossly polluted
with a variety of wastes. As greater land areas are used to accommodate increasing populations
and industrial growth, there will be an increased risk of polluting the high quality groundwater
resources. Waste products enter groundwater supplies by direct injection through wells, by
percolation of liquids sprayed over the land or leached from soluble solids at the surface, by leaking
or broken sewer lines, seepage from waste lagoons, infiltration of polluted surface streams, inter-
aquifer leakage, irrigation return waters, leachates from landfills, septic tank effluents, seawater
encroachment, and upwelling of salt water into fresh water aquifers. As a result of improper
source protection and inadequate treatment, it is not surprising that many small public water
systems have the poorest records for compliance with the Federal Drinking Water Standards.
A variety of waterborne disease outbreaks have been attributed to untreated or poorly
treated groundwater containing pathogenic forms of bacteria, virus or eucaryotic organisms, i.e.,
amoebas, Giardia , worms, etc. Specific bacterial pathogens that have been isolated from well
waters include enteropathogenic E. coli, Vibrio cholera, Shigella flexneri , S. sonnei, Salmonella
typhimurium, Yersinia enterocolitica and Campylobacter enteritis (Woodward, et al., 1974; LIndel
and Quinn, 1973; Center for Disease Control, 1973; Center for Disease Control, 1974; Center for
Disease Control, 1980; Greenberg and Jongerth, 1966; Highsmith, et al., 1977; Lassen, 1972;
Schiemann, 1978; Mentzing, 1981; Schroeder, et al., 1968; Dragas and Tratnik, 1975; and Evison and
James, 1973). Poliovirus and enterovirus have been isolated from a well water used to provide
water to restaurant patrons (Vander Velde and Mach, 1973). A brief review of completed
groundwater studies which includes a national survey of community water supply systems,
Tennessee-Georgia rural water supplies, interstate highway drinking water systems (Kansas,
Oregon, Virginia), and the Umatilla Indian Reservation groundwater survey (Allen and Geldreich,
1975), illustrates some bacteriological problems in groundwaters (Table 3). In these four surveys, 9
to 51 percent of the samples examined contained total cotiforms, and 2 to 27 percent of these same
samples were positive for fecal coliforms. Compared to surface water systems, the lower
percentage of contamination for groundwater systems in the community water supply study, reflect
better protection of the source and in many cases, application of disinfection. The high
-------
8
Table 3. Microbiological Summary of Completed Groundwater Surveys
Number Percent Percent
Survey of Samples Coliforms* Fecal Coliforms*
Community Water Supply Study
621
9.0
2.0
Tennessee-Georgia Rural Water Supplies
I
,257
51.4
27.0
Interstate Highway Drinking Water Systems
241
15.4
2.9
Umafilla Indian Reservation
498
35.9
9.0
*1 or more organisms per 100 ml
-------
9
percentages of rural supplies contaminated with coliforms were attributed to faulty construction
practices and inadequate sanitary safeguard for source waters of marginal quality.
WATER QUALITY
Water treatment plants built O to 80 years ago are frequently hard-pressed to adequately
treat waters of increasingly poorer quality and other treatment plants are being operated beyond
their design capacity. More recently, some water plant operations have modified treatment proc-
esses to achieve reductions in trihalomethane concentrations. These conditions may decrease the
protective barrier against waterborne pathogens in surface source waters that are not always of
uniform microbial quality. In the treatment trade-off to reduce chlorinated compounds in the
water by delaying disinfection, the possibility of more frequent occurrences of coliforms later into
the treatment chain must be accepted. As such, greater reliance must be placed on a continuous
final disinfection process prior to release of the finished water into the distribution system
(Symons, et al., 1970).
The treatment changes most likely to alter the transport and fate of microorganisms within
the treatment chain involve: a) changing the point of chlorination to follow clarification, b) use of
granular activated carbon (GAC) adsorption to achieve organic chemical reductions, and (c) altera-
tions in the type of disinfectant and application.
Changing the point of chlorine application (Ohio River Valley Water Sanitation Commission,
1980) was studied at the Cincinnati, Ohio Water Works in a series of two week study periods
(Table 4). The results of both the routine and modified treatment schemes showed that 48-hr
source water storage with alum treatment reduced the total coliform densities by approximately
97 percent, and the turbidities by approximately 90 percent. The coagulation and settling process,
however, had little further effect on turbidity reductions and only 50 percent additional reduction
of the coliform population occurred. Locating the point of chlorination after coagulation and
settling (Table 5) resulted in an intrusion of coliforms into the early stages of water treatment and
placed increased importance on maintaining on effective disinfection process at this stage to
reduce the burden on filtration. The apparent persistence of a residual standard plate count into
the filtration stage, regardless of the point of disinfection, illustrates the chlorine resistant nature
of some of these organisms.
-------
lO
Table 4. Before Treatment Modification
Chlorine
Application
- Point Ohio River Source Water°
Parameter
Sample Point (mean valuesb)
C i , Application to Stored Source Wa er
Sobrce Water Temperature 24°C (75 F)
Source
Stored Coagulated
Source & Settled Filtered
Finished
Flow time, hrs
0
48 52 52.5
55.5
Turbidity, NTU
32
1.0 1.2 0.11
0.10
Total Coliform per
100 mL
9,600
200
-------
II
Table 5. After Treatment Modification
Chlorine Application Point - Ohio River Source Water°
Parameter
Sample Point (mean valuesb)
Cl 2
Application to
Source Water
Stored
Coagulated anddSetfle Water
Temperature 24 C (72 F)
Coagulated
Source
Source
& Settled Filtered
Finished
Flow time, hrs.
0
48
52 525
55.5
Turbidity,NTLJ
14
0.80
1.1 0.07
0.06
Total Coliform per
100 mL
84,000
2,400
1,400
-------
12
By adsorbing organic substances, granular activated carbon particles concentrate bacterial
nutrients and also provide suitable attachment sites for microbial habitation. Data obtained from a
study of the Beaver Falls, Pennsylvania Municipal Water plant (Ohio River Valley Water Sanitation
Commission, 1980) indicated that at temperatures above 10°C (50°F), total coliform densities in
the effluent from activated carbon filter adsorber beds (Table 6) exceeded influent densities of less
than one organism per 100 ml . When the seasonal water temperatures dropped below 10°C,
effluent total coliform densities returned to below detectable levels in 100 ml. High initial total
coliform occurrences may also be attributed to the difficulty of disinfecting adsorption beds when
placing them info service. These field data suggest that occasional coliform penetration past the
early stages of treatment and before filtration can occur arid that these organisms become
temporarily established in the activated carbon filter/adsorber effluent.
Chloramines, chlorine dioxide, ozone, and ultra-violet light have each been proposed as a
practical alternative to free chlorine for water disinfection. Because of the desire to maintain a
disinfectant residual in distribution water, chioramines and chlorine dioxide have received the most
attention. Although monochloramine is definitely a less effective disinfectant than free chlorine
when evaluated at comparable low dose concentrations and short contact periods, it may be
practical in many plant operations when longer contact times and application of high concentra-
tions are feasible.
WATER SUPPLY DISTRIBUTION
Water transmitted through the distribution network is subject to quality changes. Sediments
develop in sfandpipes and slow-flow sections of the distribution network as a result of variations in
treatment application, breakthrough of source water turbidity arid corrosion. Following the
intrusion of microbial contamination into the system, colonization by selected organisms in the
population may occur (Allen, et at., 1980). The persistence and possible regrowth of organisms in
the pipe network are influenced by a variety of conditions that include physical and chemical
characteristics of the water, system age, variety of pipe materials, and availability of suitable
sites for colonization.
Key factors in microbial colonization within the distribution systems are a source of
nutrients, a favorable water temperature and a protective habitat (Geidreich, et at., 977 and
Water Research Centre, 1979). Nat all organisms entering the water supply distribution system
-------
3
Table 6. Granular Activated Carbon Study - Beaver River Source Water*
GAC Influent GAC Effluent**
Total Standard Total Standard
T ,mp. Turbidity Coliform Plate Count Turbidity Coliform Plate Count
Week C NTU per 100 mL per mL NTU per 100 mL per mL
21 5.6 < I NR 0.44 64 NR
2 21 4.8 < I NR 0.36 75 NR
3 15 2.3 < I NR 0.31 98 1,000
4 I I 2.9
-------
14
persist or become adapted to this environment and regrow (Water Research Centre, 1979; Reasoner
and Geldreich, 1979; and Van der Kooij and Zoeteman, 1978). Within the total coliform indicator
group, Kiebsiella and Enterobocter strains are most often noted in distribution systems (Geldrech,
etol., 1977; Water Reserach Centre, 1979; Reasoner and Geldreich, 1979; Van der Kooij and
Zoeteman, 1978, and Ptak, et al, 1973) and require minimal nutrients for growth. Among the other
organisms encountered in bacterial regrowth, Pseudomonas, F I avobacter ium, Acinetobacter , and
Arthrobocter are the most troublesome because of potential interference to coliform detection and
their acknowledged role as opportunistic pathogens (Hutchinson, et al., Fischer, 1950; Weaver and
Bolter, 195 I; Geidreich, et al., 1978; and Herson and Victoreen, 1980).
These bacterial problems in water distribution systems are not new but recognition of this
fact has been slow to evolve. In the past, the traditional concept was that coliform occurrence in
water with measurable chlorine residual was Impossible. Reporting of coilform occurrences was
frequently considered erroneous and was attributed to poor sampling technique, use of non-sterile
sample bottles, or laboratory error in analysis. Much of this attitude has been overcome through
certification of laboratories, maintenance of a quality assurance program and better training of
sample collectors (Geldreich, 1975 and Water Supply Quality Assurance Work Group, 1978).
Furthermore, the recent requirements for verification of all coliform occurrences in drinking
water, research into the mechanisms of disinfection and factors that modify disinfection
effectiveness in field situations, have done much to prove that coliform occurrences are not
spurious results that can be ignored.
While treatment processes are a selective force that could influence the chlorine resistance
of bacterial survivors in drinking water, there is no data available to indicate the sudden
emergence of new coliform strains of greater resistance. The three most frequently isolated
coliforms in drinking water have been identified as Klebsiella pneumoniae, Enterobacter aerogenes ,
and Enterobacter cloocae (Geldreich, et al., 1977). All three can survive on minimal nutrients and
will form a thick-walled capsule surrounding each cell in the pipe environment. This structure
gives these organisms some additional protection from disinfection exposure not usually observed in
laboratory test tube experiments (Reilly and Kippin, 1980). In most cases of coliform occurrence in
distribution water, generally only a single strain was found to have colonized the pipe network for
varying periods of time.
-------
15
Generally speaking, these occurrences in water distribution result from coliform regrowth but
should be considered an early warning that microbial penetration occurred in treatment, or that
there was loss of integrity of the distribution network through changes in water distribution flow or
adjustment of water pH to control taste and color complaints by consumers. All physical and
chemical manipulations in the distribution network carry the risk of changing the equilibrium and
structure of sediments which have entrapped viable organisms that could become released and
reenter the main flow of water. Such occurrences must not be ignored solely because they are
coliforms since the same pathway may result in pathogen intrusions into a water supply.
MONITORING
Since drinking water can act as a passive vehicle for the transmission of a number of serious
infectious diseases, the bacteriological quality of potable water is of paramount importance and
monitoring must be given highest priority. For those water supply systems that have laboratory
capability, there Is considerable advantage to utilizing the available laboratory resource to near
maximum capacity so that the operation achieves optimum cost effectiveness (Geldreich and
Kennedy, I 982). Monitoring for virus removal in modified water treatment operations involving
trihalomethane reductions, is beyond the capability of the average laboratory at this time and
would increase cost per test by a factor of 100 times or more over the cost of a coliform analysis.
Development of a surrogate virus indicator, such as a coliphage test, which would be simpler to
perform and more reasonable in cost is a desirable objective.
Effective monitoring requires careful consideration of sampling frequency based on many
factors including quality of source water, type of treatment utilized, and long-term integrity of the
distribution system. Furthermore, microorganisms are rarely randomly dispersed in the pipe net-
work; there are variations in flow due to demand and the distribution network configurations are
not simple. Therefore, sampling points must be selected carefully because the examination of a
single sample can indicate no more than the conditions prevailing at the moment of sampling at
that one site in the system.
If contamination from an external source occurs In the transmission or distribution main, It
may be carried Into one or more sections of the system and should be detected In more than one
sample In the sampling pattern. Contamination occurring In a distribution manifold or street
lateral typically would be detected only within one limited section of the pipe network in a large
-------
16
system. Inconsistent frequency of coliform occurrences in the area around elevated storage tanks
and standpipes is often reflected in water flow reversals that occur during different periods of the
day (Pipes and Christian, 1981).
Since there is no practical way to obtain a number of samples from all areas of a water
system at a given instant, periodic sampling is done and the accumulated information averaged
monthly. The aim of such a program is to demonstrate that the lower the mean coliform density,
the better the water quality and for a given mean coliform density, the less sample to sample
varioblity. Unfortunately, monthly averaging of data places the greatest burden of water supply
acceptability on small water systems (serving less than 4,100 people) where only one to three
samples are analyzed per month.
Sample transit time from the site of collection to the laboratory for processing continues to
present a problem in achieving a representative measure of any coliforrn occurrences. Current
recommendations specify that potable water samples must be examined within 30 hours of
collection (American Public Health Association, 980). New evidence (Figure I) supports the
argument that samples should be analyzed promptly and data reliability is optimized when these
samples are analyzed within three hours of collection (Nash and Geldreich, 1980). Apparently
cotiform recovery is a result of the interaction of factors manifested singly and collectively during
storage. These factors include storage temperature, initial coliform density, density and regrowth
of the other organisms in the microbial flora, specific occurrence of organisms antagonistic to
coliform survival and coliform stress from disinfectant exposure.
Monitoring of coliform populations in drinking water reveals that 74 to 86 percent of the
organisms that survive in dead-end areas of distribution lines are stressed by a number of factors
Including copper toxicity and high populations of standard plate count organisms and disinfectant
residual. Recovery of these organisms is not only impacted by sample transit time but also by
sample processing procedures in the laboratory (Lamka, et al., 1980; Evans, et al., 1981; and
McFeters, et al., 1982). Bile salt compounds added to media for selective recovery of coliforms
adversely affect coliform detection efficiency. Physiological studies suggest that amino acids and
vitamins ore required for recovery of the injured cells on a minimal medium. Hopefully,
investigations into development of revised test protocols and a new coliform medium for multiple
tube and membrane filter procedures will produce a significant breakthrough towards optimizing
-------
FIGURE 1
17
STORAGE TIME (hours)
EFFECT OF STORAGE AT 22°C ON COLIFORM AND
SPC DENSITIES DURING 1978
1 O
Mar-June 13
June 27-Sept 27
Oct- Dec
•—. 10.6°C
‘—.21.8°C
—a 14.5°C
Symbols
Average Sample Temperatures
— 1O
E
N
C.)
C l )
1
102
E
0
0
I-
N
E
C
1*.
I —
C
C.)
U.
\
101
4
* £
.— - — —
•
— —
a
U
—
—
1
E
0
0
I-
N
E
0
0
ic.)
z
a.
24 30
-------
18
coliform detection to complement improved sample preservation techniques for better monitoring
of all water supplies.
SUMMARY
The greatest impact of water pollution on man’s health comes through drinking water, the
source of which may be degraded by municipal sewage and industrial effluents, landfill leachates,
stormwater runoff, food processing wastes, and agricultural activities on the watershed. Water
treatment technology can successfully process poor quality source waters for chemical and
microbial contaminants. However, in the trade-off to decrease trihalomethane concentrations by
delaying disinfection, some later critical reductions of microbial populations in the treatment train
must be accepted. Greater reliance must therefore be placed on effective, continuous final
disinfection to minimize microbial risks. Water transmitted in distribution networks may also be
subject to quality changes introduced by microbial colonization of sediments in standpipes and slow
flow sections. Regrowth problems may be controlled using an effective main flushing program that
permits a maintenance of a free chlorine residual to all points of the distribution lines. Finally, the
logistic problems involved in monitoring water quality must be resolved through more effective
system sampling, prompt analysis of all water, samples and improved laboratory procedures to
detect coliform occurrences and new emerging waterborne pathogens.
LITERATURE CITED
Allard, J., D. A. Champaign, R. Delisle, H. Mires and E. Lippy. 1977. Waterborne Giardiasis
Outbreaks - Washington, New Hampshire. Morbid. Mortal. Weekly Rpt. 26:169.
Allen, M. J. and E. E. Geldreich. 1975. Bacteriological Criteria for Groundwater Quality.
Groundwater 13:45-52.
Allen, M. J., R. H. Taylor, and E. E. Geldreich. 1980. The Occurrence of Microorganisms in Water
Main Encrusfations. J. Am. Water Works Assoc. 72:614-625.
Alter, A. J. 1954. Appearance of Intestinal Wastes in Surface Water Supplies at Ketchikari,
Alaska. Proc. Fifth Alaska Sci. Conf. AAAS, Anchorage, Ak.
American Public Health Association. 1980. Standard Methods for the Examination of Water and
Wastewater . 15th Edition. American Public Health Association, Inc., Washington, D. C.
Center for Disease Control. 1973. Morbid. Mortal. Weekly Rpt. 22:77-78.
Center for Disease Control. 1974. Morbid. Mortal. Weekly Rpt. 23:134.
Center for Disease Control. 1980/ Waterborne Disease Outbreaks In the United States - 1978.
Morbid. Mortal. Weekly Rpt. 29:46-48.
-------
19
Comptroller General of the United States. 1978. Report to the Congress: Waste Disposal
Practices - A Threat to Health and the Nation’s Water Supply. General Accounting Office
CED-78- 120 Washington, D.C.
Council on Environmental Quality. 1975. Recreation on Water Supply Reservoirs: A Handbook for
Increased Use. USGPO S/N 04 1-01 1-00027- I. Washington, D.C.
Dragas, A-Z and M. Tratnik. 1975. Is the Examination of Drinkable Water and Swimming Pools on
Presence of Enteropathogenic E. col Necessary? Zbl. Bakt. Hyg. I. Abt. Orig. B 160:60-64.
Dzyuban, A. N. 1975. The Number and Generation Time of Bacteria and Production of Bacterial
Biomass in Water of the Saratov Reservoir. Gidrobiol. Zh. 11:14.
Evans, 1. M., C. W. Waarvick, R. J. Seidler, and M. W. Lechevallier. 1981. Failure of the Most-
Probable Number Technique to Detect Coliforms in Drinking Water and Raw Water Supplies.
AppI. Environ. Microbiol. 41:130-138.
Evison, L. M. and A. James. 1973. A Comparison of the Distribution of Intestinal Bacteria in
British and East African Water Sources. J. Appl. Bacteriol. 36:109-I 18.
Fennell, H., D. B. James, and J. MorrIs. 1974. Pollution of a Storage Reservoir by Roosting Gulls.
Water Treat. Exam. 23:5.
Fischer, G. 1950. The Antagonistic Effect of Aerobic Sporulating Bacteria on the Coli-Aeroqenes
Group. Zeit. Immu. u. Exp. Therm., 107:16.
Frost, W. H. and H. W. Streeter. 1924. A Study of the Pollution and Natural Purification of the
OhIo River: II. Report on Surveys and Laboratory Studies, Sec. VI, Bacteriological Studies,
P.HIS. Bull. 143, U.S. Pub. Health Service, Washington, D.C.
Gauf in, A. L. 1974. A Study of Natural Water Systems in Western Montana. Montana Univ. Joint
Water Resources Res. Center. OWRT A-OIl-Mont. (I), Bozeman, Mont.
Geldreich, E. E. 1972. Buffalo Lake Recreational Water Quality: A Study In Bacteriological Data
Interpretation. Water Res. :9I3.
Geldreich, E. E. 1975. Handbook for Evaluating Water Bacteriological Laboratories, U.S.
Environmental Protection A ency, EPA-670/9-75-006, Cincinnati, Ohio.
Geldreich, E. E. and H. Kennedy. 1982. The Cost of Microbiological Monitoring. In: Water Quality
and Health Significance of Bacterial Indicators of Pollution , W. 0. Pipes, ( 1J, CRC Press,
Boca Raton, Ha. (In Press) .
Geldreich, E. E., H. D. Nash, and D. Spino. 1977. Characterizing Bacterial Populations in Treated
Water Supplies: A Progress Report. Proc. Amer. Water Works Water Quality Technol. Conf.
2 B-5 Kansas City, Mo.
Geldreich, E. E., H.D. Nash, D. F. Spino, and D. J. Reasoner. 1980. Bacterial Dynamics in a Water
Supply Reservoir, A Case Study. J. Am. Water Works Assoc. 72:31-40.
Geldreich, E. E., M. J. Allen, and R. H. Taylor. 1978. Interferences to Coliform Detection in
Potable Water Supplies. In: Evaluation of the Microbiology Standards for Drinking Water , C.
W. Hendricks, (ed.), U.S. nvironmental Protection Agency, EPA-570/9-78 -cj02, Washington ,
D.C.
Greenberg, A. E. and J. Jongerth. 1966. Salmonellosis in Riverside, Calif. J. Am. Water Works
Assoc.58:l 145-1150.
Hendricks, C. W. 1974. Bdellovibrio bacteriovorus-Escherichia coIl Interactions in the Continuous
Culture of River Water. Environ. Letters 2:31 F.
-------
20
ierson, D. S. and H. T. Victoreen. 1980. Hindrance of Coliform Recovery by Turbidity and Non-
Coliforms. U.S. Environmental Protection Agency, EPA-600/2-80-097, Cincinnati, Ohio.
Hibler, C. P., K. MacLead, and D. 0. Lyman. 975. Giardiasis-In Residents of Rome, N.Y., and in
the U.S. Travelers to the Soviet Union. Morbid. Mortal. Weekly Rpt. 26:169.
Highsmith, A. K., J. C. Feeley, P. Skaliy, J. G. Wells, and B. 1. Wood. 1977. Isolation of Yersinia
enterocolitica from Well Water and Growth in Distilled Water. AppI. Environ. Microbiot.
34,745-750.
Horvath, K. et al. 1975. Studies on the Water Quality of the Rohod-Reservoir (N.E. Hungary).
Acta Biol. Debrecina 12:119.
Huff, L. L. 1981. The Economic Analysis of Health Risks and the Environmental Assessment of
Revised Fecal Coliform Effluent and Water Quality Standards, Document No. 81/15, Illinois
Institute of Natural Resources, Chicago, II.
Hutchinson, D., R. H. Weaver, and M. Scherago. 1943. The Incidence and Significance of
Microorganisms Antagonistic to Escherichia coli in Water. J. Bacteriot. 45:29.
Ketchikan Laboratory Studies Disclose Gulls are Implicated in Disease Spread. 1954. Alaska’s
Health 11:1.
King, J. G. and A. C. Mace, Jr. 1974. Effects of Recreation on Water Quality. J. Water Poll.
Contr. Fed. 46:2453.
Kirner, J. C., J. D. Littler, and L. Angelo. 1978. Giardiasis in Camas, Washington. J. Am. Water
Works Assoc. 70:35.
Kittrell, F. W. and S. A. Furfari, 1963. Observations of Coliform Bacteria in Streams. Jour. Water
Poll. Contr. Fed. 35:1361.
Lamka, K. G., M. W. LeChevallier, and R. J. Seidler. 1980. Bacterial Contamination of Drinking
Water Supplies in a Modern Rural Neighborhood. Appi. Environ. Microbiol. 39:734-738.
Lassen, J. 1972. Yersinia enterocolitica in Drinking Water. Scand. J. Infect. Dis. 4:125-127.
Lee, R. E., J. M. Symons, and G. G. Robeck. 1970. Watershed Human-Use Level and Water
Quality. J. Am. Water Works Assoc. 62:412.
Lindel, S. S. and P. Quinn. 1973. Shigella sonnei Isolated from Well Water. AppI. Microbiol.
26:424-425.
McCabe, L. J., J. M. Symons, R. D. Lee, and G. G. Robeck. 1970. Survey of Community Water
Supply Systems. J. Am. Water Works. Assoc. 62:670-687.
McFeters, G. A., S. C. Cameron, and M. W. LeChevallier. 1982. Influence of Diluents, Media and
Membrane Filters on the Detection of Injured Waterborne Coliform Bacteria. AppI. and
Environ. Microbiol. 43:97-103
Mentzing, L-O. 1981. Waterborne Outbreaks of Campylobacter enteritis in Central Sweden.
Lancet, 8:352-354.
Nash, H. D. and E. E. Geldreich. 1980. Effect of Storage on Coliform Detection in Potable Water
Samples. Amer. Water Works Assoc. Water Quality Technol. Confr. 123-136.
O’Connor, J. T., L. Hemphill, D.C. Reach, Jr. 1981. Removal of Virus from Public Water Supplies.
Project Report, U.S. EPA Grant #R-804903 (In Press).
Olivieri, V. P., C. W. Kruse, and K. Kawata. 1977. MicroorganismS in Stormwaler.
EPA-600/2-77-081, U.S. Environmental Protection Agency, Cincinnati, Ohio.
-------
21
Office of Drinking Water. 1978. Guidance for Planning the Location of Wafer Supply Intakes
Downstream from Municipal Wastewater Treatment Facilities. Contract No. 68-01-4473,
U.S. Environmental Protection Agency, Washington, D.C.
Ohio River Valley Water Sanitation Commission. 1980. Water Treatment Process Modifications
for Trihalomethane Control and Organic Substances in the Ohio River. EPA-600/2-80-028,
U.S. Environmental Protection Agency, Cincinnati, Ohio.
Pipes, W. 0. and R. R. Christian. 1981. Sampling Frequency - Microbiological Drinking Water
Regulations. U.S. Environmental Protection Agency, Contract EPA R-805-637, Office of
Drinking Water, Washington, D. C. (In Press).
Ptak, D.J., W. Ginsburg, and B. F. Willey. 1973. Identification and Incidence of Klebsiella in
Chlorinated Water Supplies. J. Am. Water Works Assoc. 65:604.
Reasoner, D. J. and E. E. Geldreich. 1979. Significance of Pigmented Bacteria in Water Supplies.
Amer. Water Works Assoc., Proc. Water Quality Technol. Confr., Philadelphia, Pa.
Reid, L. W. 1966. Wastewater Pollution and General Eutrophication of a Hydroelectric
Impoundment. J. Water Poll. Conir. Fed. 38:165.
Reilly, J. K. and J. F. Kippin, 1980. Encapsulated Coliform - Chlorine Resistant. J. New England
Water Works Assoc. 94:24.
Romaninko, V.1. 1971. Total Bacterial Number in Rybinsk Reservoir. Microbiolgia (USSR) 40:707.
Schiemann, D. A. 1978. Isolation of Yersinia enterocolitica from Surface and Well Waters in
Ontario. Can. J. Microbiol. 24:1048-1052.
Schroeder, S. A., J. R. CaIdwell, 1. M. Vernon, P. C. White, S. I. Granger, and J. V. Bennett. 1968.
A Waterborne Outbreak of Gastroenteritis in Adults Associated with Escherichia coli. 1968.
Lancet, 6:737-740.
Shilo, M., and B. Bruff. 1965. Lysis of Gram Negative Bacteria by Host-Independent Ectoparasitic
Bdellovibrio bacteriovorus Isolates. J. Gen. Microbiol. 40:317.
Statistical Summary of Municipal Water Facilities in United States. 1963. PHS Pub. 1039.
Swayne, M. D., G. H. Boone, D. Bauer, and J. Scott Lee. 1980. Wastewater in Receiving Waters at
Water Supply Abstraction Points. EPA-600/2-80-044, Municipal Environmental Research
Laboratory, Cincinnati, Ohio.
Symons, J. M., J. K. Carswell, and G. G. Robeck. 1970. Mixing of Water Supply Reservoirs for
Quality Control. J. Am. Water Works Assoc. 62:322.
Syrnons, J. M., A. A. Stevens, R. M. Clark, E. E. Geldreich, 0. T. Love, Jr.,and J. DeMarco. 1981.
Treatment Techniques for Controlling Trihalomethanes in Drinking Water .
EPA 60O7 8l-I56. Cincinnati, Ohio.
U.S. Environmental Protection Agency, Office of Water Quality, Region VII. 1971. Report on
Missouri River Water Quality Studies, Kansas City, Mo.
Vander Velde, T. L. and W. M. Mack. 1973. Poliovirus in Water Supply. J. Am. Water Works
Assoc. 65:345-348.
Van der Kooij, D. and B. C. J. Zoeteman. 1978. Water Quality in Distribution Systems. Special
Subject 5. Internat. Water Supply Assoc. Cong. Kyoto, Japan.
Water Research Centre. 1979. Water Distribution Systems: Maintenance of Water Quality and
Pipeline Integrity. Conf. Papers and Proceedings. Medmenham Laboratory, Medmenham,
Eng lui J.
-------
22
Water Supply Quality Assurance Work Group. 1978. Manual for the Interim Certification of
Laboratories Involved in Analyzing Public Drinking Water Supplies: Criteria and Procedures.
EPA 600/8-78-008, Office of Monitoring and Technical Support, U.S. Environmental Protec-
tion Agency, Washington, D. C.
Weaver, R. H. and 1. Boiter. 1951. Antibiotic-Producing Species of Bacillus from Well Water.
Trans Kentucky Acad. Sci. 13:183.
Woodward, H. E., N. Hirschhorn, R. B. Sack, R. A. Cash, I. Brownlee, G. H. Chickogong, L. K.
Evans, R. H. Shephard, and R. C. Woodard. 1974. Acute Diarrhea on an Apache Indian
Reservation. Amer. Jour. Epidemiol., 99:281-290.
Wuhrmann, K. 1972. Stream Purification. In: Water Pollution Microbiology , R. Mitchell, (ed.),
Wiley lnterscience, N.Y.
-------
EXECUTIVE SUMMARIES OF ISSUES PAPERS
-------
TABLE OF CONTENTS
EXECUTIVE SUMMARIES
Title Page No .
Microbial Agents of Waterborne Disease I
Measurement of Microbial Quality 12
Monitoring of Microbial Water Quality 14
Analytical Methods for Microbial Water Quality 20
Source, Treatment and Distribution 25
Compliance and Policy Issues 29
-------
MICROBIAL AGENTS OF WATERBORNE DISEASE
EXECUTIVE SUMMARY
Introduction
This paper identife.s the disease-causing microbiological agents or their. toxic products which
are potentially present in drinking water or its sources. A brief review of the current state of
knowledge regarding each agent is presented. information on the potential for these agents to be
in drinking water is analyzed and interpreted. Recommendations are developed for regulations for
microbial drinking water standards and practices and for critical research needs.
Regulations for Source Water Quality
Generally increasing wastewater pollution burdens to both ground and surface waters and the
continued occurrence of waterborne disease outbreaks where there has been obvious source water
contamination Indicate the need for regulations to evaluate source water qualIty. For example, in
the perIod 1971-1977, 67% of the largest waterborne disease outbreaks were due to source water
contamination where treoment was either Inadequate or nonexistent. It is recommended that
regulations be developed for continuing programs of surveillance that are based upon on-site
sanitary surveys.
Analyses of information onwaterborne pathogens and fecol Indicator bacteria such as total
and fecal coliforms and E. coil demonstrate a number of different relationships between Indicators
and various pathogens. For some enteric pathogens such as Salmonella and Shigella , the presence
of the pathogen In fecally contamInated waters is generally predictable using traditional fecal
indicator bacteria. This relationship exists because these pathogens ore almost Invariably
associated with fecal contamination and their survival In the oquatic environment and their
reduction by conventional water treatment processes are similar to that of fecal Indicator
bacterIa.
For other enteric pathogens, their presence in raw source waters or finIshed water supplies
may not be adequately predicted by traditional fecal Indicator bacteria. Although Yersinla
enterocolltica, Campyiobacter jejuni , enteric viruses and Glardia lamblia are associated with fecal
contamination, some of these pathogens, notably Yersinia, Compylobocter , and Giardia , have
-------
2
animal reservoirs, and their ability to survive and persist in natural waters or their resistance to
removal and destruction by water treatment processes may be greater than that of coliform
bacteria. Therefore, traditional fecal indicator bacteria levels in source waters and finished
waters are quantitatively less reliable and predictive for these enteric pathogens. It should be
noted, however, that for these microbial disease agents, as exemplified by the enteric viruses,
there is a general relationship of increasing pathogen frequency and perhaps concentration with
increasing indicator bacteria levels. The relationships between these disease agents and fecal
indicator bacteria are more qualitative than quantitative.
There are also disease agents that are not associated with fecal contamination and may occur
“naturally” in aquatic environments, such as the opportunistic bacterial pathogens, nontubercular
mycobacferjcj, Legionella pneumophila and cyanobacteria (blue-green algae) toxins. For these
agents, coliform bacteria are clearly inadequate indicator organisms and therefore, other
approaches are needed to protect and evaluate water quality for them.
Regulations for On-Site Sanitary Surveys
Regulations are recommended for on-site sanitary surveys. For source waters, such surveys
will identify: (a) potential fecal and industrial waste sources that could degrade raw water quality,
(b) potential animal reservoirs of pathogens, and (c) other potential conditions that could lead to
pathogen contamination, such as thermal enrichment which could encourage Legionella prolifera-
tion or toxic blue-green algae blooms.
Treatment Regulations
Treatment regulations are recommended because of (a) continued deterioration of raw source
waters, (b) continued occurrence of waterborne outbreaks, (c) the recognition of waterborne
pathogens that are either not predictable or only marginally predictable by coliform analysis, and
the resistance of these pathogens to natural self-purification and water treatment processes.
Minimum treatment for all groundwater supplies should be chlorination with an effective free
residual or an adequate alternative disinfectant. In the absence of excessive turbidity this
treatment should effectively reduce such problematic pathogens as Yersinia and Campylobacter as
well as bacterial pathogens of traditional concern ( Salmonella and Shigella ) to acceptably low
levels. Minimum treatment for surface water supplies should be filtration and chlorination with an
-------
3
effective free residual or an adequate alternative disinfectant. in the absence of excessive
turbidity this treatment should effectively reduce such pathogens as Yersinia, Campylobocter , and
Giardia .
Regulations for Finished Water and Distribution Systems
During the period 1971-1977, 27% of all large waterborne disease outbreaks were due to
disfribution system contamination. Therefore, it appears that additional effort must be made to
protect finished water quality once it leaves treatment plants and before it reaches the consumer.
Furthermore, there is a growing body of physical, chemical and microbial evidence indicating that
finished water quality often degrades during transport through distribution systems. For example,
a number of studies have shown that standard plate counts and turbidity levels of drinking water
increase as water travels through distribution systems. Based on current knowledge, the reasons
for such increases appear to be: (I) bacterial regrowfh and colonization, (2) corrosion of
distribution lines, and (3) the entrance of external contaminants by such means as undetected
broken or cracked pipes, maintenance and repair operations on mains and cross connections.
Because of water quality problems related to distribution systems, including the presence of
opportunistic pathogens, regulations are needed for monitoring finished water quality. Two
specific issues must be addressed. These are: (I) monitoring programs that are designed to
accurately assess finished water quality throughout the distribution system, and (2) better
indicators of finished water quality during distribution. The former requires the establishment of
rational monitoring programs in which the location of sampling points and the sampling frequency
are cap8ble of achieving this objective. The latter requires the use of additional water quality
indicators other than the coliform test. Two additional indicators for this purpose should be an
appropriate standard plate count and turbidity. Therefore, routine monitoring of total coliforms,
standard plate count organisms and turbidity in finished drinking water is recommended at both the
treatment plant and at various locations in the distribution system that are truly representative of
the range of conditions in the distribution system. This monitoring should be done within the
framework of a vigorous surveillance program for evaluation and protection of distribution system
integrity.
-------
Guideline and Research Recommendations for Specific Disease Agents
Legionella pneumophila and other Legionefla-like bacteria.
Guideline Recommendations
(I) When potable water is used for evaporative type cooling systems, adequate measures such
us chlorine disinfection should be taken to prevent the proliferation of Legionella to excessive
levels.
(2) Excessive levels of Legionella in the piping system of new public use facilities such as
hospitals, hotels and workplaces should be controlled by disinfection according to the recommended
AWWA procedure for new pipes or an equally effective procedure.
(3) Cooling tower exhausts in such facilities should be directed in such a way that they are
not readily drawn into fresh air intakes. Institutions with populations at high risk, such as
hospitals, should be particularly made aware of the potential presence of Legionella In finished
water. For example, if the institution is located at the end of a water distribution system line or
has a self-contained water supply system, disinfection within the facility and monitoring of hot
water sources for jIa by appropriate methods should be considered.
(4) Utilities having water supplies with large holding reservoirs or towers should be Informed
that Legionella can potentially grow in these storage systems and that they should take necessary
actions to prevent Its proliferation.
(5) In public places where high population densities ore likely to be near such “amplifiers” of
Legionella as recirculating fountains, humidifiers (or dehumidifiers), cooling towers and air wash
systems, guidance should be given concerning the possibility that such sources may oct as
Legionella habitats from which aerosol dissemination can occur.
Research Needs
(I) Ecology . Information is needed on the physical, chemical and biological factors that
influence Legionella occurrence, survival and proliferation (growth) in both source and finished
waters, especially under storage conditions. Information is also needed on Leglonella occurrence In
groundwaters, including deep aquifers. Research is needed to determine the relationships, if any,
between Legionella and traditional indicator bacteria including fecal indicators and standard plate
count organisms.
-------
5
(2) Source wafers and stored finished waters . Although quantitative information is available
on Legionella occurrence in natural water in the eastern U.S., data are more limited for other parts
of the country. Furthermore, because most of these data have been based only on direct
fluorescent antibody analysis (DFA), there is a need to corroborate these findings using additional
methodology based upon viable counts.
(3) Treatment . The reduction of Legionella in conventional water treatment processes must
be determined with particular attention given to its destruction by free chlorine and a variety of
alternative disinfectants.
(4) Distribution . Research information is especially needed on Legionella occurrence and
proliferation in distribution and storage systems, with particular emphasis on the role of dead ends
and other areas of distribution systems having long residence times.
(5) Transport studies . Research is especially needed on the aerosolization of Legionella from
amplifier systems supplied with finished water, such as water fountains, shower heads and
evaporative type cooling systems in environments where high risk exposure of the population may
occur.
(6) Pathogenicity . Information is needed on the pathogenicity, virulence and human infective
doses of Legionella from natural and finished waters. Such studies should include the identification
of pathogen ici ty/vi rul ence factors.
(7) Health Effects . The relative significance of drinking water in the transmission of
Legionella should be evaluated by determining the potential for infection by ingestion and
inhalation of contaminated drinking water.
Opportunistic Pathogens
Guideline Recommendations
(I) Hospitals, nursing homes and other institutions with high risk populations should be
informed of microbial water quality changes in distribution systems that may lead to conditions of
increased exposure to opportunistic pathogens within these facilities.
(2) Water distribution systems in newly constructed hospital facilities should be disinfected
according to AWWA procedures for new distribution lines or an equally effective procedure. The
adequacy of disinfection should be verified by standard plate count measurement.
-------
6
Research Needs . The relationship between waterborne opportunistic organisms and the
occurrence of infection and disease in special risk populations such as infants, the elderly and
1 mmunocornpromi sed individuals should be delineated.
Enter c Viruses
Guideline Recommendations . There exists epidemiological evidence of waterborne outbreaks
ue to such enteric viruses as hepatitis A, Norwalk-type agents and rotaviruses. Furthermore,
eriteric viruses that are capable of causing human infection have been isolated from drinking
waters that met standards for coliforrns and turbidity and contained free chlorine residuals.
Because there appears to be on association between virus occurrence and fecal coliform levels in
row water, it is recommended that where source waters contain >500 fecal coliforms/lOO ml,
monitoring be conducted to establish treatment process efficiency. Evaluation of treatment
process efficiency shall be based upon the measurement of an appropriate viral indicator in the
following samples: intake wafer, water just prior to final disinfection, and finished water at the
treatment plant.
Research Needs
(I) Water quality Indicator(s) for enteric viruses . Monitoring the effectiveness of specific
water treatment processes for enteric virus removal is recommended when source water quality Is
poor (>500 fecal coliforms/l00 ml). Research is needed to develop a suitable Indicator parameter
or virus marker for enteric viruse that could be readily monitored in the field. The suitability of
coliphages, turbidity, and possibly other potential enteric virus indicators should be determined.
(2) Hepatitis A and gastroenteritis viruses . Hepatitis A and gastroenteritis due to Norwalk-
type agents and perhaps rotaviruses are the most frequent waterborne viral diseases In the U.S.
Research on these agents is needed in the following areas:
(a) improved methods for their detection, cultivation and assay:
(b) studies on their removal and destruction in water and wastewater treatment processes
and their survival in natural waters;
(c) field studies on their occurrence in raw waters used as source water.
(3) Enteric virus field surveys . Although extensive field studies have been conducted on
enteric virus occurrence in row source waters and finished water supplies, new field studies are
needed. The reasons for this are: (I) the availability of improved virus detection methods and (ii)
-------
7
recent evidence from limited field studies showing that enteric virus removals in a full-scale,
conventional wafer treatment plant are not nearly as great as the predicted removals based on
previously reported laboratory and pilot plant studies. Such field studies should be done using the
best available methods and they should be done on a variety of different water supplies that
represent a wide range of possible conditions for source water quality, type of treatment system,
and finished water quality.
(1k) Enteric virus dose-response studies . Some dose-response information is available on the
likelihood of infection (or illness) from ingesting different amounts of enteric viruses. However,
more studies using additional virus types, lower virus doses, repeated exposures and different
human host populations are needed.
(5) Epidemiological studies . Epidemiological studies are needed to determine if drinking
water is a significant source of virus infection in the population and if this route of transmission is
important for virus entry into and persistence within communities.
(6) Enterlc virus detection methods . Although methods to detect and quantify enteric viruses
in water have been greatly improved in recent years, It is likely that only a small and variable
proportion of the total viruses actually present in a field sample are being detected. Further
Improvements in virus detection methods are needed in order to make them more efficient and
reliable. The best available methods should also be systematically evaluated in carefully designed
collaborative studies and quality assurance tests. In addition, methods are needed to monitor virus
recovery efficiency when processing field samples, possibly by using marker viruses or virus analogs
In field samples.
Glardia lambUa
Guideline Recommendations . Although filtration and chlorination with a free residual are
capable of extensive Giardia reductions, it is advised that systems also employ appropriate
continuous in-line turbidity monitoring for filtration performance. This additional parameter is
needed to detect minor turbidity increases which may be indicative of cyst breakthrough.
Furthermore, it is suggested that filter back wash water be wasted and that water also be wasted
during filter ripening periods to prevent cyst contamination of the filtered water.
Research Needs
(I) In order to accurately determine the effectiveness of disinfection, improved methods for
indicating cyst viability are needed.
-------
8
(2) Research is needed on the effectiveness of alternative disinfectants (to free chlorine) to
destroy Giardia cysts.
(3) Improved methods are needed to detect and quantify Giardia cysts in natural and finished
wafers.
(L) Stool surveys of watershed animals in different regions of the country are needed in order
to determine the potential for contamination of water supplies by other than human wastes.
(5) Epidemiological studies are needed to determine if drinking water is a significant source
of Giardia infection in the population and if this route of transmission is important for Giardia
entry into and persistence within communities.
Nontubercular Mycobacteria
Guideline Recommendations . Nontubercular mycobocteria appear to be a newly emerging
group of potential waterborne pathogens. However, at the present time, there is too little
information available to recommend specific guidelines based on sound scientific evidence.
Because of the potential of this group to produce disease in compromised hosts, and the possibility
of waterborne transmission, research in this area is needed.
Research Needs . Research Is needed to determine the occurrence, persistence and
proliferation of mycobocteria in source waters, through conventional treatment processes and
within distribution systems. Such studies require the development of quantitative methods for the
recovery of these organisms. Furthermore, epidemlological studies on the waterborne transmission
of these disease agents should be conducted and surveillance should Include skin testing of the
exposed population. Research is also needed on the pathogenicity and virulence of waterborne
nontubercular mycobacteria.
Toxic Freshwater Cyanobacteria (Blue-Green Algae)
Guideline Recommendations
(I) Municipalities using surface waters in which toxigenic algae are present in amounts
exceeding I0 3 cells/mI by direct microscopic analysis should monitor for the presence of a toxk
strain via mouse bioassay and/or free toxin. The water should be monitored both before and after
treatment.
(2) Use of finished waters containing toxins at an LD 50 of JOOO mg/kg body weight (l.a. by
mouse bioassay) should be discontinued until toxin bioassay indicates an LD 50 of 2,000 mg/kg.
-------
9
Research Needs
(I) Research is needed to develop chemical methods for detecting and quantifying algal toxins
in water.
(2) Quantitative information is needed on the effects of water treatment processes on algal
toxins.
(3) There is a need for better epidemioiogical surveillance information on the occurrence of
waterborne algal toxin outbreaks. Such information could be gathered through cooperative joint
efforts by the EPA, CDC and appropriate state agencies.
Panel Conclusions for Other Agents in Water
Acceptability of Existing Regulations for Pathogens. The Panel concluded that adequate
protection of drinking water from public health risks due to Salmonella, Shigelia and entero-
pathogenic E. coIl is provided by current regulations and technology.
Agents Lacking Public Health Sigrdf (canoe in Drinking Waters. The following agents were
Judged by the panel to be of Inconsequential public health significance In drinking water,
endotoxins and antibiotic resistance factors in bacteria.
Agents for Which There Is No Current Evidence of Public Health Risk in Drinking Water. The
panel concluded that there was no evidence for current public health risks via U.S. drinking water
for the following agents: Entamoeba histolytica, Noegleria sp., Acanthamoeba , and the
helmlnths. However, the potential public health significance of these agents In drinking water
should not be Ignored and should be periodically reevaluated. This is particularly important for
Entamoeba histolytica , because of the documented historical evidence for its transmission via
drinking water.
Agents for Which There is a Suggestion of Public Health Risk. Nontubercular mycobocteria
in drinking water do not appear to pose a risk to healthy consumers. However, they may pose a risk
to certain segments of the population that have existing health problems, especIally if the
organisms colonize special environments. The panel established no regulatory recommendations for
these organisms, except those far opportunistic pathogens in general.
-------
l0
Additional Research Recommendations for Disease Agents in Source Water,
Treatment Processes and Distribution Systems
Source Waters. Quantitative information is needed on the occurrence and survival of a
number of disease agents in source (natural) waters under a variety of environmental conditions.
These agents are: the bacteria Yersinia enterocolitica, Campylobocter jejuni, Legionella
pneumophila , other Legionella-like bacteria (LLB), and nontubercular mycobocteria; the enteric
viruses, especially rotaviruses, Norwalk-type agents and hepatitis A; and the protozoan Giardia
lamblia .
Water Treatment. Information on the effectiveness of water treatment processes for many
disease agents is lacking. This is particularly evident for the more recently recognized disease
agents, such as Yersinja enterocolitica, Campylobocter jejuni, Legionella pneumophila and other
LLB, nontubercular mycobocteria, enteric viruses and algal toxins.
Spec f Ic Research Recommendaffans far Other Disease Agents
Salmonella and Shigefla . Better quantification methods are needed for these two genera of
bacteria in natural and finished waters. The development of these methods should emphasize ease,
reliability and sensitivity.
Yersinia enterocolltica . Quantitative methods for detecting Yersinia in water are needed.
Quantitative information is needed on the survival of Yersinia in natural waters under a variety of
environmental conditions and on the reduction of Yersinia by such water treatment processes as
chemical coagulation, flocculation, filtration, and disinfection by free chlorine and alternative
disinfectants. There is a need for information on pathogenicity and virulence of environmentally
isolated Yersinia organisms compared to clinical isolates. Such studies should include information
on the relationship between pathogenicity/virulence and the presence of endotoxin, heat stable
enterotoxin and invasiveness factor.
Cam pylobacter jejuni . Improved methods are need for isolating and quantifying Campyl-
obocter in natural and finished waters. Complete identificatIon of the animal reservoirs for this
agent and their impact on source water is needed. The relationships between traditional fecal
-------
I I
indicator bacteria and Campylobacter in water should be quantitatively determined with respect to
their survival in source waters and their reduction by conventional water treatment processes.
In order to determine the role of water in the transmission of Campylobacter , better
information is needed on the endemic occurrence of this agent in the population and on its role in
gastrointestinal illness.
-------
12
MEASUREMENT OF MICROBIAL QUALITY
EXECUTIVE SUMMARY
The panel on the measurement of microbial quality felt that the existing coliform and
turbidity regulations are appropriate and provide a substantial but not absolute safeguard from
disease. Minor modifications in the quantification and formulation of the coliform determination
will not alter the utility of the measurement. At present, in drinking water, total coliforms are
still the best indicators available. However, the relationship of coliforms to the risk of disease
transmission through contaminated water is not well established. The panel felt strongly that the
importance of the sanitary survey be re-emphasized by specific incorporation into the primary
regulations. This will permit better interpretation of the microbial quality of water. The sanitary
survey should be conducted In all water systems and updated annually. Complete surveys should be
conducted at least every 5 years. Sanitary surveys should be conducted In all new, modified or
expanded water systems before use. Since the sanItary survey has been neglected, the committee
recommends that a set of guidelines be developed In the form of a handbook or similar document
that can be used to conduct the survey.
Human pathogens should be absent from finished drinking water such that the level of disease
Is below an acceptable risk. However, It should be noted that a test for the presence of any such
agent In a speclf led volume of finished water cannot prove the absence of the agent from the water
from which the sample was taken. On the basIs of what Is now known, any standard requiring
testing to demonstrate that a given pathogen is not detectable in a specified volume of finished
water seems unlikely to add to protection of public health significantly, so as to justify the added
cost of the required sampling and testing. The measurement panel, thus, took the classic position
of recommending application of surrogate determinations.
Three distinct aspects of the use of surrogates to measure microbial quality were recognized.
The surrogates may serve as: I. IndIcators of feces (the classical role of indicators), 2. Indicators of
water treatment efficiency, 3. Indicators of deterioration, degradation and/or recontamination.
No single measurement adequately fulfills all the requirements for these aspects of indicators
throughout the water system from source to tap.
-------
‘3
The panel evaluated the selected surrogate measurements in drinking water for the above use
categories. Escherichia col was the most useful fecal indicator. Total coliform, heterotrophic
plate count and free chlorine residual were considered the best indicators of treatment efficiency
and post treatment degradation, deterioration, and/or contamination.
The following guidelines were suggested for source water, treatment trains and distribution
systems. Three levels of testing were recommended: I. routine, 2. periodic, and 3. diagnostic.
Chlorine demand, turbidity and any fecal indicator were recommended for routine testing of source
waters. Chemical and physical measurements are proposed because of their ability to provide rapid
information on the quality of source water and the fecal indicator is recommended because it can
provide information on long term water quality changes. Free chlorine residual, turbidity, total
coliform and aerobic heterotrophic bacteria were recommended for routine testing of treatment
efficiency, and any deterioration, degradation or recontamination of the distribution system.
Enterococci and Clostridium perfringens spores were suggested for periodic testing in the
treatment train in order to provide a record of inactivation and removal of indicators known to be
more resistant to chlorination than the more commonly used total coliform indicator group. These
two organisms were also suggested as indicators of water quality in the distribution systems.
The sanitary survey was recommended as the most useful approach to diagnosing water
quality problems in the water system. Ancillary microbial identification tests should be used in
conjunction with the sanitary survey in problems occurring in treatment trains and the distribution
system.
Each of the surrogate measurements proposed above should be employed in the same manner
that a physician uses disease symptoms, laboratory analysis and patient history to diagnose illness
to prescribe a remedy. Each determination represents an important piece of information that can
be used to interpret the water quality. The more information the better the interpretation.
The heterotrophic plate count received serious consideration by the measurement panel.
While it was agreed that the recommendation of a regulation with inherent legal requirements and
enforcement provisions would not be useful, the application of the plate count for water treatment
quality control and distribution system monitoring remains desirable. Every water network should
be encouraged to apply a consistent heterotrophic plate count methodology. The plate count
results should be recorded on a quality control chart and observed values exceeding established
limits should motivate modification of the water system operation.
-------
14
MONITORING OF MICROBIAL WATER QUALITY
EXECUTIVE SUMMARY
The objective of microbiological monitoring is to provide a measure of the reliability of
protection of potable water against the transmission of waterborne disease. This protection
depends upon multiple barriers which include selection of the highest quality source water, use of
various water treatment processes, disinfection of treated water, and protection of the integrity of
the water distribution system. Ideally, the monitoring program should provide data which can be
used to evaluate the reliability of protection over long periods of time and over the entire area
served by the distribution system. However, monitoring programs required under the National
Interim Primary Drinking Water Regulations fall far short of the ideal, particularly for small water
systems which examine only a few samples each month for microbiological contamination. The
purpose of this issues paper is to point out some possibilities for badly needed improvements in the
present approach to microbiological monitoring of water distribution systems.
Statistical Parameters for Evaluation
Indicator bacteria ore seldom randomly dispersed In a water distribution system; that is, the
variance of the indicator bacteria density is usually much greater than the mean. The estimation
of the mean density of indicator bacteria in a water distribution system is subject to extremely
large errors because of this variability and also because of the occasional occurrence of densities
which are too large to be measured by the currently used examination techniques. The Monitoring
Panel strongly recommends that the average sample count be dropped as a parameter for
evaluation of microbiological contamination of water flowing through a distribution system.
The fraction of the samples from a water distribution system which have some indicator
bacteria present can be readily modeled using the binomial distribution without reference to the
parameters of the frequency distribution of the indicator bacteria density. The Monitoring Panel
recommends that the fraction of the samples with indicator bacteria present be used as the
parameter for assessing the level of microbiological contamination.
-------
15
Protection Relithility Standard
The Monitoring Panel recommends that a protection reliability standard (PRS) be adopted.
The PRS recommended is no more than 5% of the samples examined during any consecutive
12 month period with coliform bacteria present in 50 ml of water. The water utility shall include
in each monthly report the results of the bacteriological examinations for the previous II months
to demonstrate compliance or non-compliance with the PRS. Violation of the PRS would not
require public notification but should cause the state to conduct an evaluation of the water system
for the purpose of determining the cause of the violation and stipulating remedial action.
Minimum Number of Samples Required
The Monitoring Panel recommends that enough samples be collected to give at least 95%
confidence that less than 10% of the water passing through c i distribution system in any
12 consecutive month period is contaminated. For the smallest systems covered by the regulations,
this is 60 samples per 12 months or 5 samples per month, in which case 3 positive samples in
12 months would meet the PRS and give 95% confidence that less than 10% of the water is
contaminated.
Maximum Contaminant Levels
If is recommended that the MCL should be rio more than 20% of the samples collected in any
month with coliform bacteria present in 50 ml in water. This would allow the smallest systems
with 5 samples per month to have one positive sample in a month without violating the MCL.
Violation of the MCL requires public notification.
Number of Samples Required
The minimum number of original samples per month recommended for the smallest systems Is
5. The Monitoring Panel recommends that the minimum number of samples per month for the
largest systems remain 500. It is also recommended that the parameter for determining the
minimum number of samples per month for intermediate size systems should be the total length of
pipes making up the system.
-------
16
Location of Sampling Points
Samples shall be collected from points which are well dispersed throughout the entire
distribution system. Sampling sites should be selected in consultation with the state but should not
be fixed from month to month. The sites should be varied in such a way that all sections of the
system including areas of low flow are sampled in the course of c i 12 month period. There should be
no part of the distribution system where contamination could persist indefinitely without any
possibility of its being detected by the monitoring program.
Check Samples
When coliform bacteria are found in any sample from a distribution system, check samples
should be collected and examined as soon as possible. The normal expectation is that check
samples will not have coliform bacteria present. If a check sample has coliform bacteria present,
this is evidence of a localized source of contamination and an investigation should be undertaken to
identify and eliminate the contamination. The results of the bacteriological examination of check
samples should be used in calculating the fraction of the samples with coliforms present for the
monthly report. This will allow a water utility to avoid violation of the MCL and public
notification when the contamination is not extensive.
Remedial Action
If co liforrns are found in more than 20% of the approved number of samples per month at any
time before the end of the month, the regulatory agency should be notified within 48 hours. This
potential violation of the MCL should cause the water utility, in conjunction with the state, to
initiate immediate investigative and corrective action which should include more extensive
sampling of the distribution system in areas which the samples containing coliforms were collected,
sampling of the raw water source, treated water, and distribution system storage, identification of
the bacteria in the positive samples, review of records of treatment including disinfection, review
of records of disfribution system pressure variations, and formulation of a plan of action to
eliminate the contamination.
-------
17
Monitoring Treated Water
Microbiological monitoring of treated wafer is a separate problem from monitoring of a
distribution system. In order to eliminate continual seeding of the distribution system with
coliform bacteria, it is necessary to demonstrate that at least 99% of the treated water entering a
distribution system has a coliform density of less than I per 500 ml. This can be accomplished by
collecting and examining 360 samples (500 ml) of finished water over a period of approximately
12 months. If no more than one of these samples is positive, this provides 99% confidence that no
more than 1% of the water has a coliform density as high as I per 500 ml. If there are
interruptions of treatment or alterations in treatment processes, the demonstration should be
repeated by monitoring of the treated water for another 12 month period.
Stondcrd Plate Coimt Monitoring
The Monitoring Panel recommends that the standard plate count be considered for InclusIon
in the monitoring program and that plate counts be made on some samples collected for coliform
tests for an Initial time period and on a seasonal basis to establish a data base. If a comparison of
standard plate counts from the points of entry to any points in the distribution system
demonstrates a one hundred fold Increase, then the standard plate count should be included in the
monitoring to provide information on treatment effectiveness, potential interference with coliform
tests, and the possible presence of opportunistic pathogens. This recommendation Is for a guideline
rather than a regulation.
Turbidity Monitoring
The Monitoring Panel recognizes that the presence of high amounts of turbidity, whether
inorganic or organic, can interfere with the coliforms test and/or protect microorganisms from
disinfectant action. Therefore, it is recommended that the present MCL of I TU be retained but
that monitoring of the treated water entering the system be increased to once each 8 hour shift
unless continuous monitoring is provided. A filtered water turbidity of
-------
18
be a filtered water turbidity of 0.2 TU, each filter should be monitored for possible breakthrough,
and continuous monitoring and recording should be provided, especially at locations where plants
operate unattended for portions of each day.
Free Chlorine Monitoring
Under the present Regulations it is possible to substitute free chlorine measurements for
some of the microbiological monitoring samples. The Monitoring Panel is aware of data which
show that the presence of free chlorine (above 0.2 mg/I) is not equivalent to the absence of
coliform bacteria but is not aware of any evidence to show which method of monitoring is more
effective in protecting public health. The Monitoring Panel recommends that the chlorine
substitution option be retained and that studies to determine its effectiveness be undertaken.
Needed Reseorch
The recommended changes in the Drinking Water Regulations should cause a fundamental
Iteration in the approach to microbiological monitoring of water distribution systems. That such
an alteration is needed is clear especially for small water distribution systems. However, there is
very little experience with the approach to monitoring embodied in the recommendations of this
Issues Paper and research is needed to demonstrate how these concepts can best be applied.
Among the projects which should be carried out are:
I. Development of a data base for small systems (population <9,400 examining <10
samples per month under the present Regulations) to show the probabilities of
exceeding the present and recommended MCL’s. How will the percent compliance
change if the recommended MCL is adopted?
2. Investigations to determine the most effective and least cost methods (media, sample
collection and handling, etc.) of detecting coliforms if the attempt to measure the
coliform density of individual samples is eliminated.
3. Studies of the advantages of testing positive samples for a variety of indicator bacteria
(Clark’s P-A test) and the costs associated with the use of such procedures.
4. Development of sampling approaches and programs for localization of contamination in
a disfribution system.
-------
‘9
5. Relatively long term projects (3 to 5 years) to determine normal variability of the
microbiological quality of water in a distribution system and to identify those events
(e.g., changes in water source, treatment alterations, construction of new mains, etc.)
which cause changes in the microbiological quality of the water.
-------
20
ANALYTICAL METHODS FOR MICROBIAL WATER QUALITY
EXECUTIVE SUMMARY
Definition of Coliform
The committee spent considerable effort in reaching a consensus on a new definition of
coliforms. The following is a more general definition based on taxonomic considerations. This
definition should encourage the development of new and innovative methods for finding more rapid
and economical means to detect these bacteria. The definition of coiiforms is to include those
lactose-fermenting, gram-negative, non-sporeforming rods which comprise the taxonomic
assemblage of species now recognized as coliform indicator organisms. These species are members
of the genera Cifrobacter, Enterobacter, Klebsiella , and Escherichia and should also include
metabolically similar species, i.e., strains of Aeromonas, Hofnla, Serratia , and Yersinia that
conform to this definition.
Recommendations. Investigate validity of 35°C temperature for coliform enumeration.
Support research for pursuing new, Innovative, modern directions for coliform detection and
enumeration.
Fermentation T be Proc.éw.
Conclusions. Numerous problems were noted with the current coliform enumeration proce-
dures. With regard to the fermentation tube technique, the panel:
I. Indicated fewer cases of synergistic lactose fermentors and fewer, if any, gram-positive
rods will grow in lauryl tryptose broth (LTB) compared with lactose broth (LB).
2. Agreed that masking of coliform detection has been demonstrote4 in presumptive
media. LTB presumptive medium is superior to LB since fewer turbid, gas-negative
presumptive tubes are found.
Regulations. Gas-negative turbid presumptive tubes should be processed to the confirmatory
step. All turbid-only tubes from any given distribution system should be processed until It is
confirmed that coliform masking does not occur. The panel did not determine how long or how
many samples should be processed. If evidence of coliform masking is obtained, 10% of all future
turbid-only presumptive tubes should be processed.
-------
21
Recommendations.
I. The minerals-modified glutamate medium should be evaluated in the U.S. as an
alternative to LTB.
2. Studies should be conducted to find a suitable alternative to the BGLB confirmatory
medium. BGLB is too inhibitory (selective) and false negative reactions can be obtained.
Suggested current alternatives to BGLB include:
EC broth
m-Endo agar LES
3. The bias in the current MPN Tables is acknowledged. Studies are underway to correct
this problem.
Membrane Filtration Procedure
Conclusiort.s. Membrane filtration media and techniques result in significant underestimates in
coliform numbers. This appears to be due to at least the following factors:
I. Physiological injury results in a loss of up to 50-90% of the total coliform population.
2. Lack of differential colony color results in the appearance of atypical (non-sheen)
coliform colonies which may not be verified.
3. Use of BGLB verification suppresses gas formation by certain coliforms.
Regulations. The panel considers a regulation is necessary to require verification of atypical
colonies. A minimum of one plate per month or 10% of all plates examined on a monthly basis
should be tested. Five atypical colonies minimum should be tested whenever typical colonies are
absent. The panel also recommends that BGLB be deleted and verification be performed using
LTB.
Recommendations. Research efforts are required to solve these deficiencies in crucial
monitoring areas. Specifically research is necessary to:
I. Develop alternative primary recovery media to minimize loss of injured cells. Research
is currently underway In this area.
2. Develop improved and/or innovative coliform detection and verification procedures.
Verification procedures may be based upon detection of substances unique to coliforms
-------
22
(specific enzymes) or may involve improved media designed to promote rapid gas
production.
Turbidity ond Mernbrcvw Filtration
Recommendations. When turbid surface water supplies are not chemically treated or filtered,
the membrane filter coliform enumeration technique can underestimate the level of coliform
contamination. Therefore, serious consideration should be given to monitoring coliforms by the
fermentation tube procedure when finished water turbidities exceed 5 NTLJ.
Presence-Absence Test (P-A)
Recommendations. Further evaluation of the P-A test as an alternative coliform isolation
technique is a strongly recommended research need . Revisions to the previously published
procedure have been made by the Ministry on the Environment (Canada) and have enhanced the
semi-quantitative aspects of this concept. Additional advantages include the I) possibility of
examining a larger number of samples, 2) supplemental analyses for a variety of indicators and/or
pathogens from a single sample providing a wider data base and 3) amenability to automation.
Stonckwd Plate Count (SPC)
Conclusions. An SPC regulation is justif led for the following reasons:
I. Provides additional information relating to treatment efficiency.
2. Indicates changes/problems in the distribution system, including growth promoting
substances.
3. Indicates possible problem of coliform inhibition (suppression).
Regulations. SPC should be monitored at a sampling frequency coinciding with coliform
requirement.
Recommendations. Specific density limitation is not really feasible, given the variety of
circumstances in different systems and seasonal changes in the same system. The panel
recommmends the following as general guidance: <100/mi, achievable goal for all systems;
100-500/mi, anticipated during seasonal increase or at certain locations in the system (dead end,
-------
23
low residual) where flushing would be indicated; >500/mi, poor microbiological quality. If single
sample results show a 5-10 fold increase over commonly achieved levels, the cause should be
investigated. Greater than tO-fold increases should require reporting. For SPC analyses, any
method in Standard Methods is appropriate (e.g., pour plate or spread plate), but a 48-hr incubation
at 35°C should be used.
Fecal Coliform FC)
Recommendations. The committee could identify no basis for the evaluation/enumeration of
fecal coliforms in drinking water. Rather, in those circumstances that raise concern regarding the
sanitary qualify of a water supply (i.e., increased turbidity, treatment breakdown, possible cross-
connections, etc.), increased monitoring with total coliforms and SPC should be indicated.
Other Detection Procedures
Conclusions. The committee felt that at present no other detection procedure involving
bioassays, chemical analyses, or automated electronic systems was yet available to replace
“conventional” indicator organism analyses.
Recommendations. The panel strongly suggests continued research to find rapid, economical,
and efficient testing procedures for monitoring biological agents of water quality. Automated
procedures have already been de eloped to monitor biological agents important to the pharmaceu-
tical industry and to monitor fecal coliforms in wastewater effluents. These breakthroughs should
encourage research initiatives in the field of potable water monitoring.
Chlorine Residual Measurements
ConcLusions. Precision and accuracy of residual chlorine analyses described in Standard
Methods revealed most to be satisfactory. However, one commonly used commercial field kit was
reported to overestimate chlorine concentrations. This potential problem deserves further
investigation.
-------
24
Collection and Transport of Drinking Water Samples
Recommendations. The scientific evidence illustrates the complex and apparently unpredict-
able nature of coliform survival, die-off or regrowth in potable water. The panel recommends that
the current 30-hr maximum transport time be retained and all efforts be devoted to sample
analysis in the shortest possible time. The panel rejects the concept of discarding samples that
exceed the 30-hr limit, but within reason. It is not the panel’s intent to allow analysis of samples
more than 3-4 days old. Late samples should be analyzed for contamination immediately and a
second sample requested. If the late sample contains coliforms, a “check” sample will already be
enroute to the laboratory. Suppliers of chronically late samples requesting a variance must
indicate any special conditions which might allow for or justify continued violation of transport
time requirements. Research efforts must be devoted to finding chemical agents capable of
stabilizing coliform counts during periods of transport.
-------
25
SOURCE, TREATMENT AND DISTRIBUTION
EXECUTIVE SUMMARY
Source
In selecting a raw water source, either a surface water or a groundwater, it is normal
practice to use that source(s) which provides the best quality water to meet the demand for the
required quantity. A groundwater will generally have a more consistent and better microbial
quality than a surface water, which often contains a wide range of microorganisms, in terms of
both type (bacteria, algae, protozoa, fungi, etc.) and density. Some microorganisms are indigenous
to the water environment while others may be transients, gaining entrance to the water from the
air or by way of wostewater discharges or runoff from urban or rural areas. The major sources of
microbial pathogens affecting a surface water are domestic and industrial wastewater discharges,
both treated and untreated; agricultural and urban runoff; septic tank seepage; and leachate from
solid waste disposal sites. Depending upon soil type, depth and associated factors, the microbial
quality of natural waters found in many aquifers (groundwaters) is acceptable because of the
removal of microorganisms by filtration and adsorption as the water percolates through the soil.
On the other hand, groundwoters are subject to microbial contamination on a site-specific basis, as
Influenced by the local topography, type of soil, nature of the contaminant, rote of water
withdrawal, type of well structure, etc. Further, the survival of human pothogens in soil and
aquifers may be on the order of several months which may be greater than in surface waters.
It is likely that the quality of source waters will continue to deteriorate, unless a deliberate
effort is made to improve and/or protect the quality. The production of a microbiologically safe
water supply must begin by controlling continuous and intermittent sources of contamination
affecting both surface and groundwaters. If this is not successfully accomplished, a greater burden
will continue to be placed on the other barriers associated with produdng a microbiologicolly
acceptable water supply, i.e. the natural purification process and water treatment facilities. In
general, the quality of the raw water dictates the processes included in the treatment system.
Thus, a raw water having a poor microbiological quality must receive more extensive treatment if
the objective of producing a finished water free of detectable pathogens is to be achieved.
Considering raw water sources with respect to producing a potable water supply, it is
recommended that:
-------
26
I. Regulatory agencies consider the row water quality relative to real and potential
sources of contamination, turbidity, bacterial quality, etc. when specifying the required
treatment of a particular source water.
2. Continuous monitoring of raw surface water sources for turbidity and pH be practiced.
3. Research be undertaken to evaluate the effect of further quality deterioration of
source waters on the efficiency of removing/inactivating microorganisms by the various
unit processes employed in water treatment and to determine whether a reliable
relationship can be established between raw water quality and the required degree of
water treatment.
Treatment
A variety of unit processes and operations, including raw water storage, aeration, pre-
chlorination, coagulation-flocculation and sedimentation, hardness reduction, filtration, activated
carbon treatment, and disinfection are available for the treatment of water, so as to render it
acceptable for public use and consumption. Of these, only disinfection is practiced specifically to
meet microbial quality objectives, i.e. absence of any detectable pathogenic microorganisms. The
other unit processes and operations, on the other hand, either remove microorganisms from water
in oddition to achieving their primary objectives or render the water more susceptible to effective
disinfection. A deliberately planned, properly designed, constructed and operated water treatment
system, consisting of an appropriate series of unit processes and operations, is capable of providing
the degree of treatment required to meet existing microbial standards.
The most widely used water disinfectant today is chlorine. It is now recognized, however,
that the use of free chlorine as a disinfectant can lead to the formation of triho omethanes (THM)
with waters containing certain forms of organic matter. THM production may be eliminated or at
least reduced, without compromising the microbial quality of the finished water, by altering the
point of chlorine addition, by removing precursor organic matter before applying chlorine, or by
using a disinfectant other than free chlorine.
In view of the importance of water treatment in meeting the microbial quality standards for
potable water supplies, it is recommended that:
I. All water supplies be disinfected unless it can be shown through a documented sanitary
survey or otherwise, that microbial contamination is not a problem.
-------
27
2. All water supplies using a surface water source be pretreated and filtered unless it can
be shown that this degree of treatment is not required.
3. All water supplies using a surface water source be monitored continuously for
disinfectant residual at a representative entry point into the distribution system; for
small systems, compliance with this regulation may not be justified for economic
reasons.
4. The microbial quality of the finished water not be compromised in meeting the MCL
standard established for THM. If the microbial quality of the finished water is
adversely affected by control measures taken to meet the MCL for THM, the microbial
standard should take precedence over the MCL established for THM. It is not
suggested, however, that compliance with the established MCL for THM be abandoned.
5. Research be initiated to determine the significance of various mechanisms of natural
die-off (or growth) of microorganisms through raw water storage, to identify conditions
associated with certain treatment processes and operations which might promote the
colonization and growth of pathogenic microorganisms, to establish the effect(s) of
dissolved inorganic substances and particulate matter on disinfectant kinetics and
efficiency, to evaluate the potential for microorganisms developing an inherited
resistance to a particular disinfectant, to determine the effectiveness of certain water
treatment processes and operations in removing/inactivating specific pathogenic micro-
organisms, and to evaluate the potential water quality problems associated with the use
of free chlorine, other than the formation of THM, and the alternative disinfectants
such as chioramines, chlorine dioxide, and ozone.
Distribution
The water distribution system, consisting of mains, pipes, and pumps and storage tanks,
provides many opportunities for microbial degradation and contamination of the wafer through such
incidents as main breaks, bock siphonage, regrowth of microorganisms within the system,
introduction of microorganisms through open reservoirs, etc. It should be recognized that the vast
majority of a water supply system (in terms of bulk, equipment, and investment) is associated with
the distribution of the water and that only a minor amount of surveillance, relative to the
treatment system, is possibie within the distribution network. On the other hand, concern has
-------
28
always been expressed about the microbial quality of the wafer in the distribution system, but
regulations have avoided any type of specific requirements. Regulatory action is necessary in
several areas (and to varying degrees) to stimulate action and surveillance at the state and/or local
levels.
To minimize the impact of distribution systems on the microbial quality of water supplies, if
is recommended that:
I. The water in all portions of a distribution system contain a disinfectant residual.
2. All new finished water reservoirs be constructed with a cover. Similarly, the placement
of covers on existing open reservoirs should be encouraged.
3. All new mains, pipes and appurtenances be disinfected before being placed in service,
utilizing the AWWA standard or an equivalent procedure. Disinfection of repaired
mains, pipes, etc. should also follow AWWA disinfection procedures where applicable.
4. An active and effective cross-connection control program be implemented for every
public water supply, taking into account the specific character of the distribution
system.
5. Research be implemented to define the condition(s) which permits microorganisms to
colonize distribution systems and to identify preventative measures, and to establish
the relationship between turbidity, physical and chemical characteristics of suspended
matter and the occurrence of microorganisms and their density.
-------
29
COMPLIANCE AND POLICY ISSUES
EXECUTIVE SUMMARY
A review of existing microbiological standards of drinking water quality and operating
experiences with the enforcement of those regulations has been completed. On the basis of this
assessment, recommendations for specific changes have been made although, it should be noted,
the regulations generally serve to protect the public health. In addition to a revised micro-
biological maximum contaminant level (MCL), it is suggested that EPA adopt regulations or provide
formal guidance with respect to analytical or evaluation details.
I. The suggested MCL requires that coliform bacteria should be absent from at least 95%
of the water samples collected and analyzed.
2. Exceeding the MCL or the single sample maximum trigger an immediate investigation
and corrective action.
3. The minimum number of samples collected per month from a community or a large
noncommunity water system should be 2. Increased emphasis should be given to using
the physical examination of the water system (sanitary survey) to evaluate water
quality.
4. The regulations should specify acceptable sampling taps and that samples should be
analyzed as soon as possible after collection but not later than 30 hours (unrefrigerated)
or 54 hours (refrigerated).
5. If the membrane filter technique is used for analysis, any evidence of coliform bacteria,
even if a count cannot be made, should be considered as a positive result. If the
fermentation tube method is used, results should be reported as percent of tubes,
confirmed test, positive.
6. The suggested MCL obviates the need for averaging coliform test results but com-
pliance should be assessed on the basis of the most recent 20 samples.
7. Where complete treatment is not provided for surface water, or disinfection for
groundwater, a standard plate count MCL of 500 colonies per milliliter should be
instituted.
-------
30
8. The chlorine substitution rule is little used (even though it is potentially valuable);
consequently, if may be eliminated from the regulations. Guidance should be provided
to define acceptable criteria for individuals making field measurements for chlorine.
9. If an acceptable presence - absence test for coliform bacteria can be developed, its use
should be incorporated into the suggested regulations.
10. The MCL for turbidity should not be changed.
II. The regulations should require that surface waters should be treated by coagulation,
sedimentation, and filtration, or equivalent, and disinfection.
12. The regulations regarding small water systems should not be revised to include
treatment requirements, but they should require meaningful sanitary surveys.
-------
ISSUES PAPERS
-------
MICROBIAL AGENTS OF WATERBORNE DISEASE
Mark Sobsey and Betty Olson
Panel Members: Martin Blaser, Gabriel Bitton, Wayne Carmichael, Gunther Craun*, Gary
duMoulin, Martin Favero, Carl Fliermans, Charles Gerba, William Hausler, George Healy, Walter
Jakubowski, Theodore Wetzler
ABSTRACT 2
INTRODUCTION 3
BACKGROUND 3
Disease Agents; Classification, Source, Illness, Pothogenesis and Epidemiology 3
Classiflcat ion of Microbial Agents 5
Waterborne and Water-Washed Pathogens 5
Water-Based Disease Agents 5
Epidemiology of Microbial Disease Agents Transmissable by Contaminated Water 6
Occurrence of Indicator Organisms and Disease Agents in Natural and Finished Waters 8
Survival of Pathogens in Natural and Finished Waters 8
Indicator Bacteria and Pathogen Reductions in Water Treatment Processes 12
ANALYSIS 22
Bacteria 22
Salmonella 22
Shigella 23
Yersinia enterocolitica 24
Campylob acter Jejunt 25
Legionello pneumophila and Other Legionella Species 27
Non-tubercular Mycobocteria 29
Opportunistic Pathogens 31
Drug resistance 32
Enteric Viruses 33
Hepatitis A Virus 34
Norwalk and Related Gastroenteritis Viruses 35
Rotoviruses 35
Adenoviruses 36
Enterovlruses 36
Reovlruses 36
Overall AnalysIs 37
Prof ozoans 37
Giordia lamblia 37
Entamoeba histolytlca 39
Bolantldlum colt 40
Noegleria fo 1 i 40
Acanthamoeba 41
Helmlnths 41
Toxins of Microbial Origin 41
Exotoxlris 41
Eridotox lris 42
CONCLUSIONS FOR SPECIFIC AGENTS 43
BacterIa 43
Salmonella and ShI ella 43
YersFnla enterocolitica and Campytobocter JeJuni 44
Leglonella 44
OpportunistIc Pathogens 44
Enter Ic VIruses 45
Protozoans 46
G lordia 46
*Gurither Craun was unable to participate in the workshop but did provide material that was
used in the preparation of this report.
-------
2
OVERALL CONCLUSIONS 46
Source Water Quality and Treatment Requirements 46
Finished Water and Distribution Systems 50
RECOMMENDATIONS FOR REGULATIONS AND GUIDELINES 50
Regulations for Source Water Quality and Treatment 50
On-Site Sanitary Surveys 50
Treatment 51
Regulations for Finished Water and Distribution Systems 51
Guidelines Recommendations for Specific Disease Agents 52
Legionella pneumophila and Other Legionella-Like Bacteria 52
Opportunistic Pathogens 52
Enter Ic Viruses 52
Giardia lamblia 53
Toxic Freshwater Cyanobocterio (Blue-Green Algae) 53
Panel Conclusions for Other Agents 53
Acceptability of Existing Regulations 53
Agents Lacking Public Health Significance 53
Agents for Which There is No Current Evidence of Public Health Risk 53
Agents for Which There is a Suggestion of Public Health Risk 54
RESEARCH RECOMMENDATIONS FOR DISEASE AGENTS 54
Introduction 54
Research Recommendations for Disease Agents in Source Water, Treatment
Processes and Distribution Systems 54
Source Waters 54
Water Treatment 54
Research Recommendations for Specific Disease Agents 55
Salmonella and Shigella 55
Yersinia enterocolitica 55
Campylobacter jejuni 55
Legionella pneumophila and other Legionella-Like Bacteria 55
Nonfuberculcir mycobocteria 56
Enteric Viruses
Opportunistic Pathogens 58
Giardia lamblia 58
Cyanobocteria Toxins 58
LITERATURE CITED 59
ABSTRACT
Recommendations for regulations and research needs are delineated for waterborne disease
agents. Specifically, four categories of agents are addressed: bacteria, enteric viruses, protozoans
and microbial toxins. Within each of these categories, the organisms or agents of importance are
described in terms of pathogenicity, epidemiology and public health significance in relation to water
supplies. The occurrence and survival of these disease agents are discussed in relation to source
water, water treatment processes and distribution systems. In addition, the ability of traditional
indicator bacteria (e.g. coliforms) to predict the presence of these pathogenic agents is discussed
and alternatives which may provide a better indication of the presence of certain disease agents
are presented.
-------
3
INTRODUCT ION
Disease-causing microbial agents and toxic products of aquatic microbes have been and will
continue to be important factors in establishing the quality of drinking water. Such agents can be
and often are present in natural waters used as raw sources for drinking water supplies, and in some
cases they are present in relatively high concentrations.
Considerable reliance is now placed upon water treatment processes to remove or destroy
microbial disease agents in raw source waters and produce finished water of acceptable microbial
quality, regardless of source water quality. Current water treatment practices are generally
capable of achieving this end, even when incoming microbial levels are high. However, it is
technically impossible to produce drinking water that is absolutely free of disease agents, arid it is
unrealistic to either assume that this is so or to even establish this as an objective.
It is generally believed that drinking water meeting current bacteriological standards and
other standards pertinent to microbial quality, such as turbidity, does not pose a significant health
risk due to disease-causing microbial agents, regardless of source water quality or the nature and
extent of treatment. However, recent discoveries of previously unrecognized potential waterborne
diseas gents, documented isolations of specific disease agents from finished waters meeting
current bacteriological and other drinking water standards, and the periodic occurrence of disease
outbreaks caused by wafer supplies meeting current standards have raised concerns about the
adequacy of present water supply practices and regulations.
In this paper we will attempt to: (I) identify the disease-causing microbial agents or their
toxic products potentially present in water, (2) briefly review the current knowledge about these
agents, (3) analyze and interpret this information in relation to current practices and regulations
and (4) recommend future directions for the development of revised regulations.
BACKGROUND
Disease Agents: Classification, Source, Illness, Pathogenesis crid Epidemiology
The important waterborne disease-causing microbial agents and toxins in the U.S. are listed
in Table I. Also listed are their major reservoirs or primary sources and estimates of typical
maximum concentrations in these reservoirs or sources. In addition, basic information concerning
human infection by these agents is presented in Table I. Specifically listed are: dIsease name or
-------
Tthle I. Surrwnci-y of T
Ncine of Orguauas or Group t’ . Tppes
wu omic, Clinical m c ]
M4or
Epidemiotogical Featu
M ReservoinJ
Primiey Soosces
res of Potenti
Conc. tn
Purnaey
5o. ce
al Drinkin
Infect.
Dose
g Water
Prevolesce
(Avg. %
Escretiurc)
Disease Agents
Durotics,
Of
9wdding
C c srier
Stole
Bacteria
Enteroviruses
Poiovirtises
Cossack icyiruses A
Coxsaclcievsruses B
Echoviruses
Other enteroi iruses
Reoviruses
Rotcwiruses
Adenoviruses
Hepatitis A virus
N rwatk uad related
Gl viruses
Algae ( blur-. eeri )
acute
2
3
23 ic.. .,.tic
os iIic ....,;Iis
31 NweIfl t
AHc. t tie
elM IJ seid Q
>3
37 lB sesi Q Mfr.
I -
, e..I.’ J witiS
l•
sLJ .IL .
flwesnilly eewidsed
.1 sd water
aeheed watei
nua teces
iseoou teces
nou feces
13sngalsum feces
wa tece
Isesses feces
teces
sewa es
?
?
7
[ l-5% ] t -3 sLeks
? -2 weeks
Iwee lc
7
? 3 weeks
? 2,900
4
> I.
Salmonella !x i
Sdmonelto Pt!:P!3 !i
Otiser sc,lm flo
Shge lla
Vibrio efioleree
i opathogenic E. coB
Yersinsci enterocoWtkss
Conipylobacl isi
Lecjicnelta pneim iiki
ond relatnsfl,acterio
Mycobacterium htherculosis
O her ( otypicmyc iO
O çortuaistk helena
Enleric Viruses
— kver
— fr
13auon teces
hi.rici /asimoI feces
135.5,. feces
13,. , . feces
t j,v o ind feces
twasi/ouisnd C?) feces
low
high
high
medium
high
high
high
9
1 1 r bpicc’lly
I ‘0.l I J 4wee lcs.
I J Locc.>lye r
I week
9
1.2—153
? ?,
? 1-3-weeks
I
Protozociss
Acenthiunotha costellor.i
Balonti . usn coil
E ntGnoebo f lytico
Giordia k, .t4ia
11 i
Hefrninths
N edes (muadesems)
As is kiis*,nlcoides
rric+..e s trichkxo
lks*wornw
Ai ytostomo aodenàe
Necator snsekocsse
5ircm 4oldes stercoralss
l0 ’Ig
l0’ig
10 Ig
tO /g
10 /g
9
7
7
. 5
7
.5
io6i
109 pizt/g
I0 Ig
l B /g
9
I0- 10/g
10-10 lq
lO-IO /g
10-10:19
10-10 /g
IO iO
106
wnlMc . .,Lidtn
I l4a ( 4. ..tm p )
eeoB k
I *entu)
ic
I
I .ewa
+
9
.5
7
9
*
.5
S
*(rnonths to years)
4
*
*
Mthaeou
Miceocystis arrugloss
Azomawa
S .izoItwix cilcicolo
wal aid waler
twasms & alma feces
suit aid waler
tuisas (& almal) feces
13a feces
13 , .wa feces
lessee. & w*nud feces
alseol
very low
very low
low 3-10
low 1 .5-2.0
? very low
I c i
-------
5
typical disease syndrome, typical human infectious dose, estimated overall prevalence of infection
in the U.S. population, duration of excretion by infected persons and potential for a carrier state
(long-term excretion). information on infectious dose is presented in terms of three dose
categories: low, medium, and high, which are defined as viable organism levels of 100,
100-10,000, and >10,000 per infectious dose, respectively.
Classification of Microbial Agents
In describing the microbial agents of public health concern in drinking water, Feachem’s
(I 977) classification of water-related disease agents is useful. Four major groups of water-related
infectious disease agents are distinguished: (a) waterborne, (b) water-washed, (c) water-based, and
(d) water-related insect vector-borne. The first three groups are especially relevant to drinking
water and the agents in Table I are in one or two of these three categories.
Waterborne and Water-Washed Pathogens. Many of the potential drinking water pathogens
ore excreted or shed by infected humans and/or animals (Table I). This category includes agents
infecting the gastrointestinal tract and excreted in feces, agents infecting the respiratory tract
and/or eyes and therefore present in exudates from these sites, and agents infecting and, hence,
shed from the skin. With the possible exception of the pathogenic E. coli, Klebsiella pneumoniae
and some of the opportunistic pathogens such as the pseudomonads, these pathogens are not normal
inhabitants of the enteric or respiratory tracts, or the skin. However, in any sizeable population,
some people are likely to be infected with, and therefore excreting or shedding at least some of
these agents.
Surface or groundwaters that become contaminated with excreta and exudates from such
sources as raw or partially treated sewage, septic waste systems, surface runoff and stormwaters
and leachates from domestic solid waste landfills can become contaminated with these pathogens.
The levels of such pathogens in excreta may vary geographically, seasonally and with the hygienic
and socioeconomic status of the contributing population. Under certain conditions, some of the
waterborne and water-washed pathogens may be able to persist for long periods and even multiply
in natural and finished waters.
Water-Based Disease Agents. Also shown in Table I are potential drinking water pathogens
and toxins that are water-based. These organisms can be “naturally” present in raw and finished
-------
6
waters under certain environmental conditions. Because they are not necessarily associated with
human or animal excreta, the potential presence of these pathogens in water is not predictable
using traditional bacterial indicators.
Epidemiology of Microbial Disease Agents Tronsmissible by Contaminated Water
Some of the agents that can be transmitted by contaminated water cause diseases that are
legally required to be reported to the U.S. Public Health Service Centers for Disease Control
(CDC). It should be emphasized, however, that this disease reporting system is largely passive and
relies on voluntary reporting by individual states of data submitted by hospitals, physicians and
other health agencies and workers. Although reporting is a legal requirement for some diseases, it
is generally acknowledged that appreciable underreporting occurs (Mausner and Bahn, 1974). In
addition to the general problem of underreporting, special studies have shown particular def i-
ciencies in both recognizing and reporting waterborne disease (Craun, 1978; Batik et al., 1983).
Despite these limitations, however, the reported data are still useful for discerning trends of
disease occurrence.
In Table 2 are shown data compiled by the CDC and the U.S. EPA on diseases that have been
transmitted by drinking water in waterborne outbreaks (Craun, l979a; 198la; l981b; Lippy, 1981).
For most of the diseases that can be transmitted by water as well as other routes, especially
fecally-transmjtfed diseases, the available data indicate that drinking water accounts for only a
small percentage of the reported cases (Table 2). Hence, it appears that other routes of exposure,
primarily direct and indirect person-to-person contact, are more important than the water route.
This conclusion must be tempered by the fact that the source of exposure is not specifically
determined for many of the reported cases of disease. Water may be the source of exposure in
some of these cases, but the extent to which the water vehicle is responsible for these cases is not
known. Furthermore, additional waterborne disease outbreaks that are not listed in Table 2 have
occurred during the period 1971-1979. Some of these additional outbreaks by specific agents are
discussed elsewhere in this paper.
Although available data suggest that contaminated drinking water accounts for only a small
proportion of the total reported cases of disease caused by many of the agents given in Table I, it
is important to recognize that common-source exposure routes such as water and food are directly
amenable to a variety of specific control measures to protect sanitary quality. In contrast,
-------
7
Table 2. Waterborne Disease Agents: Total Reported Cases and Waterborne Outbreaks
Waterborne Outbreaks, 1971-79° Average Annual Casesb
Illness/Agent No. No . Annual Woterborne All Causes
Outbreaks Cases Cases 1979
Acute gostroenteritis,
unknown etiology 1148 26,826 3,353 unknown
Giordiasis 31 17,822 2,228 5,619
Shigellosis 22 5,174 647 20,135
Infectious hepatitis 114 398 50 30,407
Salmonellosis I I 1,370 171 33,666
Norwalk-type viral
gastroenteritis 5 1,233 154 unknown
Campylobacter
gastroenteritis I 3,000 375 unknown
Toxigenic E. coil
gostroenterT W I 1,000 125 unknown
°Documented waterborne outbreaks, 197 1-79, reported to CDC and EPA; excludes outbreaks due to
chemical agents.
bCenters for Disease Control, Morbidity and Mortality Weekly Report, Annual Summary, 1979.
-------
8
transmission routes involving direct and indirect person-to-person contact are less easily con-
trolled. Therefore, continued efforts to minimize exposure to disease agents via drinking water by
vigorous application of sound engineering practices and other measures is to be encouraged.
Occurrence of Indicator Organisms and Disease Agents In Natural and Finished Waters
Data on the occurrence of disease agents and indicator bacteria in natural and finished
waters are presented in Table 3. This table is not an extensive or complete compilation of such
information. The purpose in presenting the data is to demonstrate that indicator bacteria such as
total coliforms and E. ccii can occur in high concentrations in natural surface waters and that
disease agents can also be present in raw source waters and even some finished drinking waters.
Data on the detection of microbial pathogens and/or coliform bacteria in drinking water
implicated in disease outbreaks in the U.S. are shown in Table k. This table is not intended to be a
complete listing of all waterborrie outbreaks in which pathogens and/or coliforms were detected in
the incriminated water. The table is representative of the prevailing situation with regard to
microbial analysis of drinking water implicated in outbreaks.
In many outbreaks, especially those involving smaller water supplies, coliform analyses are
not done very often until cases of illness appear in the population and water becomes the suspect
vehicle. In many instances, such analysis shows that coliforms are present at excessive levels.
Until recently, efforts were rarely mode to detect pathogens in incriminated drinking water
because of the technical difficulties of such analyses. When such efforts were made, pathogens
were sometimes detected and quantified. In some cases, it was found that pathogen levels were
low and coliform levels were high. In some other cases, pathogen levels were low and coliforms
were at low levels or were nondetectable. These latter findings suggest that for some pathogens,
coliforms may be inadequate indicators of fecal contamination under certain conditions.
Survival of Pathogens in Natural and Finished Waters
The ability of disease-causing microbes to survive and persist in natural and finished waters is
important for drinking water supplies for several reasons. Storage of raw water is often used to
improve water quality, including microbial quality, prior to further treatment. In addition to
sedimentation of large organisms such as protozoan cysts and smaller microbes associated with
particulate matter, thermal effects, ultraviolet light-induced inactivation, and predation, are
-------
TobIe 3. Occr xrence of Indicator Bacteria and Microbial Disease Agents In Natural and Finished Waters
Agent Type of Water Geographic Location Ambient Concentration Reference
(per
tOO ml)
Total coliforms river (Severn) U.K. 0-57,000 Morris and Waite, 1981
Total cotifxms river (Wcwfe) U.K. (4-27,000 Morris and Waite, 1981
Total coliforms river (Leani) W(. 100-14,000 Morris and Waite, 1981
Total coliforms river (Blithe) U.K. 0-I 10,000 Morris and Waite, 1981
Total coliforms river (Tcine). U.K. 1.1-20 x l0 Morris and Waite, 1981
Total cotiforms river (Stour) U.K. 3.9-100 x lO Morris and Waite, 1981
Total coliforms river (Wye) U.K. 120-51,000 Morris and Waite, 1981
Total coliforrns river (Diwe) U.K. 100-170,000 Morris and Waite, 1981
\0
Total coliforms river (Derwent) U.K. 10- 100,000,000 Morris and Waite, 1981
Total cotiforrns river (Avon) U.K. 1,400-59,000 Morris and Waite, 1981
Fecol coliform ri’ er Colorado 2.1-2.4 Brickler & Ttjnnicliff, 1980
Fecol coliform river tributeies to Colorado River 3.6-8.0 Brickler & Tunnicliff, 1980
E. coil river (Severn) U.K. 0-72,000 Morris and Waite, 1981
E. coil river (Worfe) UK. 100-3,300 Morris and Waite, 1981
E.coti river (Leucin) U.K. 100-2,200 Morris and Waite, 1981
E. coil river (Blithe) U.K. 0-580 Morris and Waite, 1981
E. coli river (Tanie) U.K. 1.2-16 x IO Morris and Waite, 1981
E. coil river (Stour) U.K. 4.6-tOO x IO Morris and Waite, 1981
E. coil river (Wye) U.K. 0.6-160 x (02 Morris and Waite, 1981
E. coil river (Dove) U,J<. 0-32,000 Morris and Waite, 1981
E. coil river (Derwent) U.K. 0-70,000 Morris and Waite, 1981
E.coti river (Avon) U.K. 1,000-18,000 Morris and Waite, 1981
-------
Table 3 (ConVd). Occurrence of Indicator Bacteria and Microbial Disease Agents In Natural and Finished Waters
Agent Type of Water Geographic Location Ambient Concentration Reference
(per IOU ml)
Salmonella sp. river Nancy, France 0.0-7.0 Roland & Block, 1980
Salmonella sp. river Mississippi River 43 Brickler & Tunnicliff, 1980
Salmonella sp. river Ohio River .5-300 Brickler & Tunnicliff, 1980
Salmonella sp. creek t mile below sewage outfall 4.5-12 Brickler & Tunnicliff, 1980
Salmonella sp. cistern (home) unknown 0.26-1.1 Brickler & Tunnicliff, 1980
Pseudornonas and 7
P. picketil hospital unknown 10 -10 Black etal., 1979
Opportunistic pothogens hospital respiratory unit faucet, i o2 _ 108 Reinbard, 1980
Pseudornonas sp. aerators, taps of distilled,
Acinetobocter sp. deionized drinking water.
flavobocterium sp.
Pseudomonas oeruqinosa stream unknown 0-2 x I0 Wheater, 1980
Aeromonas hydrophila river Anacosta River 2,000-20,000 Seidler etal., 1980
Enteric viruses river River Wear, U.K. 0-2.5 Edwards & Wyn-Jones, 1981
Enteric viruses river River Wear, U.K. 0-2.5 Edwards & Wyn-Jones, 1981
Enteric viruses surface GDR 0-1.8* Dobberkau et al., 1981
Enteric viruses river (bayou) Houston, TX Avg. 3.5 Grinstein ef al., 1970
Enteric viruses river U.K. 0-57.4 Morris and Waite, 1981
Endotoxin tc water Mexico City, Mexico 80.0 DiLuzio & Friedman, 1973
Endotoxin reservoir Sewickley, Penn. .03I 320.0l Sykoro, et ol., 1980
Endotoxin distribution Pennsylvania .087-.5i Sykora, et al., 1980
Endotoxin finished Pennsylvania .032-.3 I Sykora, et al., 1980
*jn virus positive samples
ml
-------
Table 4. Pathogen and Indicator Detection in Drinking Water Supplies Implicated in Outbreaks.
Etiological Agent
Type of Water Supply
Type of Treatment
Concentration of:
Reference
Pathogen
Indicator
Salmonella typhimurium
ground
untreated
17/L
1.4 total coliforms/L
Boring elal., 1971
Shigella sonnei
ground
none
>1/1.6 L
1,250 total coliforms/L
Baine etal., 1975
Salmonella typhi
ground
none?
unknown
present
Craun, 1979b
Enterotoxigenic E. coIl
surface
chlorination
unknown
present
Craun and Gunn, 1979
Yersinia enfercolitica
ground
none
unknown
unknown
Eden et al., 1977
Gkrdia lamblia
Gkrdin lamblia
underground cistern
surface
unknown
chloramination
unknown
> I cysf—
1.1 x tO
L
unknown
low w/occasional high
Craun, 1979a
Croun, 1979a
Gicrdia lamblia
surface
filtration & chlorination
unknown
unknown
Craun, l979a
Gicrdia lomblia
surface
chlorination
unknown
unknown
Craun, l979a
Gicrdio lamblio
surface
fi lfration/chlorinolian
(coogulotion)
unknown
absent
Craun, I 979a
2 raw water
two source waters; two treatment systems
-------
12
believed to contribute to the inactivation or destruction of microbial pathogens during storage.
However, if pathogens survive for considerable periods in raw water, then the only reliable means
for their control is the use of engineered treatment processes. The survival of fecally-associated
pathogens relative to the survival of indicator bacteria is also an important consideration. If these
pathogens survive no better than indicators and the latter are present in higher concentrations,
then the absence of indicators very likely reflects the absence of pathogens.
As shown by the data in Table 5, the survival of some gram negative bacterial pathogens such
as Salmonella, Shigella , and Campylobacter is generally similar to that of indicator bacteria such
as E. coh. However, other microbial pathogens, such as Yersinia enterocolitica, Giardia lamblia
and enteric viruses, survive considerably longer than do indicator bacteria and the other gram
negative bacterial pathogens. Thus, water storage cannot be relied upon to consistently produce
appreciable reductions of certain microbial pathogens unless storage times are extremely long.
Indicator Bacteria ond Pathogen Reductions in Water Treatment Processes
Water treatment is the most important and direct means to control the microbial quality of
drinking water. Treatment schemes consisting of chemical coagulation and flocculation, filtration
and disinfection aire usually used for surface waters, although high qualify surface waters are
sometimes treated by direct filtration and disinfection or just disinfection. Groundwaters, being
generally lower in such contaminants as microbial agents, particulates and organics, often receive
no treatment other than dIsinfection.
A number of laboratory and field studies have shown that under optimum conditions, these
treatment processes can substantIally reduce the levels of microbial agents and other contaminants
in water. Information on the reduction of specific agents by coagulation-flocculation, filtration
and disinfection is presented In Tables 6, 7, and 8, respectively. It should be noted that enteric
viruses and protozoan cysts are less effectively reduced by certain treatment processes than are
enteric bacteria. These findings have raised concerns about the possibility of producing drinking
water that meets current bacteriological standards but still contains sufficient viral and protozoan
pathogens to pose a health risk to consumers. Such situations are most likely to occur when
treatment is minimal or only marginally effective and raw water pathogen levels are high. At the
present time there is inadequate epidemiological information available to show that this is a
significant or widespread problem.
-------
Table 5. Survival of Pathogens in Natural and Finished Waters.
Initial
Concentration
Agent (m i 1 ) Type of
Temp.
Water (°C)
Time
(days) for
Indicated
Infectivit
y Loss:
T 1 Reference
T
T
T
T 999
Salmonella thompson I0 environmental 17-18 — — 28 Dutka and Kwan, 1980
S. enteritidis I0 well 9.5-12.5 0.79 — — McFetersetal., 1974
ser. pcratyphi D
5. enteritidis tO well 9.5-12.5 0.66 — — McFeters etal., 1974
ser. pa-atyphi A
S. enteritidis l0 well 9.5-12.5 0.66 — — McFetersetal., 1974
ser. typhimurium
Salmonella typhi IO 5 well 9.5-12.5 0.25 — — — — McFeters etal., (974
Salmonella enteritidis l0 well 9.5-12.5 0.08 — — — — McFeters etol., (974
ser. pa-otyphi B
Salmonella typhi IO Polluted river 37 —
-------
Table 5. (Cont’d) Survival of Pothogens in Natural and Finished Waters.
Shigella flexneri
Shigella flexneri
Shigella flexneri
Poliovirus 2
Coxsockievirus 63
Echovirus 7
Coxsackievirus A4
Coxsackievirus A4
Echovirus
6,11,30,33
Echovirus
6,11,30,33
Poliovirus I
Coxsockievirus A9
Poliovirus I
Coxsockievirus A9
Echovirus 7
Echovirus 7
— — >21
— — >21
— — 14
— — 4
-- Mohadjer & Mehrabian, 1975
— Mohadjer & Mehrabian, 1975
Mohodjer & Mehrabian, 1975
Pozhar, 1973
Pozhcr, 1973
Pozhcir, 1973
Simkova & Woilnerova, 1973
Simkovo & Walinerova, 1973
Welke et ci., 1973
— — Herrmann et ci., 1974
— — Herrmcnn et ci., 1974
— — Herrmcnn et ci., 1974
Herrmann et ci., 1974
Kokinoetal., 1977
Kokinoetal., 1977
Initial
Concentration
Agent (ml ’) Type of
Temp.
Water (°C)
Time
(days) for
Indicated
Infectivity
Loss:
T 1 Reference
T
T 90
T 99
T 999
river
filtered river
tap
river
river
river
river
river
diluted river
L I-’
4—6
4—6
16—20
16-20
16-20
4-8
20-22
8
io6.7
io6.7
io67
0.32
32
, 0 5I_io8J
106
I o6
I o2.3
io23
7
21
21 —
— — — — 29-35
— — — — 29-35
— — — -- 29-35
— — — — >150
— — — — >45
— — — — >560
diluted river 20
lake
lake
filtered lake
filtered lake
well
well
— 70-322 Welke et ci., 1973
19-25
19-25
21-33
21-33
20
10
— 66
— Il3
-------
Table 5. (Cont’d) Survival of Pathogens in Natural and Finished Wafers.
Initial
Concentration
Agent (m l ’) Type of
Temp.
Water (°C)
Time
(days) for
Indicate
d Infectivit
y Loss:
T 1 Reference
T
T
199
1999
Coxsockievirus B3 io2*3 well 20 — “-‘66 Kokinaetal., 1977
Coxsockievirus B3 io23 well (0 — — “ - ‘ 113 kokina etol., 1977
Vibrio cholerae unknown well 9.5-12.5 0.29 — — McFeters et aL, 1974
Vibrio cholerae l0 spring room — <0.04 Ld -uiri et al., 1939
Inabo --
Vibrio cholerae io6 tap (Calcutta) room — <0.75 Lohiri etal., 1939
Inaba
Vibrio cholerae 106 river room — <0.75 Lahiri ef al., 1939
Inoba ——
Vibrio cholerae unknown tank unknown — 1-2 Neogy, 1965
Vibrio cholerae unknown tank unknown — <8 Neogy, 1965
El Tar
Yersinia unknown sterile unknown — 157 Dominowska md Mallotka, 1971
enterocolitico
Campylobocter jejuni l0 -l0 stream 25 — — -- 2-3 4 Bkiser etaL, 1980
(autoclaved)
Campylobocter jejuni stream 4 — — — 3-18 17-33 Blaser etal., 1980
(autoclaved)
Echovirus 7 IO 3 river 4-6 -- 90 Bogdosarym & Abieva, 1971
Echovirus 7 1O 3 tap 4-6 — 75 Bogdasaryan & Abieva, 1971
Poliovirus 102.3 well 20 — 33 Kokina etal., 1977
Poliovirus I0 2 well (0 — 55 Kokino etol., 1977
-------
Table 5. (Cont’d) Survival of Pathogens in Natural and Finished Waters.
Initial
Concentration
Agent (mH) Type of
Temp.
Water (°C)
Time
(days) for Indicated
Infectivity
Loss:
T Reference
T 50
T
T 99
T 999
1977
1977
1977
(977
1977
Reovirus
lO
surface
9- 15
—
—
—
—
200
Mahnel,
etal., 1977
Poliovirus 1,3
IO
river
23-27
—
—
hO &
1.8
—
—
O’Brien
& Newman,
Coxsockievirus A13
l0
river
23-27
—
—
0.3
—
—
O’Brien
& Newman,
Poliovirus I
IO 4
river
4-8
—
—
2
—
—
O’Brien
& Newman,
Coxsockievirus BI
IO 4
river
4-8
—
—
2.4
—
—
O’Brien
& Newman,
Poliovirus I
filtered river
23-27
—
—
2
--
—
O’Brien
& Newman,
SA-l I (rotavirus)
JO 23
unpolluted fresh
20
10
—
Hurst &
Gerba, 1980
SA-l I (rotaviros)
io2.3
heavily polluted fresh
20
--
—
—
(4
—
Hurst &
Gerba, 1980
Coxsackievirus B3
io2.3
unpolluted fresh
20
—
—
—
6-8
—
Hurst &
Gerbo, 1980
Coxsockievirus B8
iol.3
heavily polluted fresh
20
—
—
—
6-8
—
Hurst &
Gerba, (980
Echovirus 7
unpolluted fresh
20
—
—
—
3
—
Hurst &
Gerba, (980
Echovirus 7
,o2.3
heavily polluted fresh
20
—
—
—
(4
—
Hurst &
Gerba, (980
Entomoeba
histolytica
various waters
4
—
—
—
55-60
—
Chang,
(943
Entarnoeba
histolyticci
various waters
—
—
—
38-42
Chong,
(943
Entamoebo
histolytica
lo -lo
various wafers
21-22
—
—
—
7-8
—
Chang,
1943
0 ’
‘No data.
-------
Table 6. Pathogen cud Indicator Bacteria Reduction by Chemical Coagulation
Agent Type of Water
Coagulation Conditions
Turbidity
Removal
(%)
% Removal
Reference
Coagulant
Dose
(mg/I)
a
T 1 —
Tot4l colifarm River A1 2 (S0 4 ) 3 (2.6 14 96 97 Curnmins & Nash, 1978
ToftI co!iform River A1 2 (50 4 ) 3 20 8 140 74 MaIlman & Kahler, 1948
Total coliform River A 1 2 (S0 4 ) 3 25 140-255 96-99.6 99.4 Chang etal., 1958a
E.co li River A 1 2 ( 50 4 ) 3 10.5 $68 90 83 Streeter, 1927
E .co li Lake A 1 2 (S0 4 ) 3 12.1 40 72 76 Streeter, 1929
Coxsackievirus A2 River A 1 2 ( 50 4 ) 3 25 140-255 95 95 Chang et al., 1958o
Coxsockievirus A2 River FeC I 3 (5 5-10 80 -90 95 Chang etal., 1958a
Polio I, 2, 3 River Fe(S0 4 ) 3 40 unknown unknown 99.8 Guy ef oL, (977
Naturally occurring River Fe(50 4 ) 3 40 unknown unknown 88.3 Guy et aL, 1977
enteroviruses
Gicrdio muris Grovel pit A $ 2 (S0 4 ) 3 5 9 61 96 Arozareno, 1977
Giordia muris Grovel pit A1 2 (S0 4 ) 3 tO, 25, 50 9 89-94 99.4 Arozarena, 1977
Giordia muris Grovel pit A 1 2 (S0 4 ) 3 25 unknown unknown 58 Arozareno, 1977
Gicrdia muris Gravel pit A 1 2 (50 4 ) 3 25 unknown unknown 90 Arozarena, $977
E. histolytica Clear water A 1 2 (50 4 ) 3 30-60 unknown unknown 33-55 Spector et at., 1934
tnitial Turbidity, in Turbidity Units
-------
Table 7. Indicator and Pathogen Reduction by Filtration
Initial
Removal
Agent Type of Filter Aerial Loading Rate Concentration
(percent) Reference
Total coliforms slow sand 5 M/day unknown 70-98 Hoekstra, 1978
Total coliforms slow sand unknown unknown 96.3-99.5 Poynter and Slade, 1977
Total coliforms slow sand unknown unknown 88 (low temp) Burmcin, etal., 1962
Total coliforms slow sand unknown unknown 50 Robeck etal., 1962
E. coil slow sand unknown unknown 41 (low temp) Burman etal., 1962
Poliovirus slow sand 0.008 GPM/ft 2 I0 6 /L 98 Robeck et al., 1962
Poliovirus slow sand 12 M/doy unknown 98.25% (5°C) Poynfer & Slade, 1977
Poliovirus slow sand 4.8 M/day unknown 99.999% (11°C) Poynter & Slade, 1977
Po liovirus sand 2 GPM/ft 2 I 0 7 /L J 0-58% Robeck ef cii., 1962
Poliovirus sand 2-6 GPM/ft 2 10 7 /L 98* Robeck ef cii., 1962
Coxsackievirus A5 sand 0.2 GPM/ft 2 8/L 98 Gilcreas and Kelly, 1955
Coxsackievirus A5 sand 2.0 GPM/ft 2 8/L 10 Gilcreas and Kelly, 1955
Poliovirus I sand unknown 10 4 /L l9 37.5** Guy etal., 1977
E.histolytica rapid sand 2 GPM/ft 2 41 I/L 99.99 Jokubowski and Hoff, 1979
E. histolytica rapid sand 6.4 or 9.5 GPM/ft 2 838-10,000/gal 80.3-99.86 Jakubowski and Hoff, 1979
E. histolytica rapid sand 6.4 GPM/ ft 2 56-10,000/gal 82.4—99.95 ,Jakubowski and Hoff, 1979
E. histolytica diatomaceous earth 0.2-4.2 GPM/ft 2 1,935-350,000/gal 99.998 Jokubowski and Hoff, 1979
Gicrdia lamblia diotomaceous earth unknown l8,000 * > 99.99 Jakubowski and Hoff, 1979
Gicrdia larnblia diatomoceous earth unknown I 60,000*” > 99.999 Jakubowski and Hoff, 1979
‘Alum pretreatment
“Ferric pretreatment
***TotoJ cysts detectable
-------
Table 8. Indicator and Pathogen Reduction by Disinfection
Concentration
Agent Disinfectant (mg/liter)
Contact Time Temp.
(mm.) pH (°C) % Reduction* Reference
E. coil HOd 0.1 0.4 6.0 5 99 Scarpino et ci., 1974
E. coil OCI .0 0.92 10.0 5 99 Scorpinoetal., 1974
1.0 175 9.0 5 99 Sidersetal., 1973
E. coil NI-1 2 C 1 1.0 64’ 9.0 15 99 Siders et ci., 1973
6iT NH.,CI .2 33.5 9.0 25 99 Sidersetal., 1973
E.coii(ATCC 11229) NHe I 2 1.0 5.5 4.5 IS 99 Esposito et ci., 1974
E. coil oc l: 0.3 10 10.0 20-25 100 Butterfield, 1948
E. coil OCI 0.4 10 10.0 20-25 100 Butterfield, 1948
1T dO 2 0.25 I .8 6.5 5 99 Cronier et cii., 1978
E.co li C 10 2 0.25 1.2 6.5 10 99 Cronier etal., 1978
E.coIi C 10 2 0.25 0.68 6.5 20 99 Cronier etal., 1978
E.coii C 10 2 fL25 2 0.27 6.5 32 99 Cronier et ci., 1978
E.coli (ATCC 11229) UV - 3 x 10 iW-S/cm N/A 7.0 20 99.9 Rice ondi:1 f, 1981
Salmonella typhi OCI 0.3 10 10.7 20-25 100 Butterfield, 1948
Pseudomonas pyocyaneae OCI_ 0.75 10 10.7 20-25 100 Butterfield, 1948
Pseudornonas pyocyoneae OCI 0.4 10 10.0 20-25 100 Butterfield, 1948
Legionella pneumophila HOCI 3.3
-------
Table 8 (Cont’d). Indicator and Pathogen Reduction by Disinfection
Concentration
Agent Disinfectant (mg/liter)
Contact Time Te, mp.
(mm..) p1-i (“C) % Reduction’ Reference
CoxsackieB3 03 0.095 10 7.0 25 99 Evison, 1977
Polio 3 03 0.082 10 7.0 25 99 Evison, 1977
Polio 1 03 0.042 10 7.0 25 99 Evison, 1977
Echo I 03 0.44 10 7.0 25 99 Evison, 1977
Coxsockie B5 03 0.053 10 7.0 25 99 Evison, 1977
Polio 2 0. 0.039 10 7.0 25 99 Evison, 1977
Echo I (Fcrouk) HOCI 0.5 0.5-%.0 6.-tO 5 99 Engelbrecht etal., 1978
Echo 5 (Noyce) HOCI 0.5 1.3-27.0 6-10 5 99 Engelbreeht ef ci., 1978
Polio I (Mahoney) HOCI 0.5 2.1-21.0 6-10 5 99 Engelbrecht etal., 1978
Coxsockie B5 (Faulkner) HOCI 0.5 3.4-66.0 6—b 5 99 Engelbrecht etcl., 1978
Coxsackie A9 (Griggs) HOCI 0.5 0.3-1.5 6-10 5 99 Engelbrecht ef ci., 1978
Polio 2 (Lansing) HOC 1 0.5 1.2-64.0 6 -tO 5 99 Engelbrecht et at., 1978
Potiovirus 3 C 10 2 0.5 5.0-0.25 5.6-8.5 20 99 Warriner, 1967
Poliovirus I C 10 2 0.5 12.0-2.0 7.0 5-25 99 Cronier et at., 1978
Coxsackievirus A9 ClO 0.5 0.5 7.0 IS 99 Cronier et al., 1978
Entomoebahistolytica HOC?I Ii 15 7.0 27 99 Chong c Fair, 1941
Entornoebo histolyticc HOCI 50.0 10 7.0 27 92 Spector et ci., 1934
Entamoebahistolytica I) 6.0 tO 5.0 30 99 Stringer and Kruse, 1971
Entomoebahistolytica liOl 22.0 10 8.0 30 99 Stringer and Kruse, 1971
Entamoeba histolytica HOCI 1.5 10 5.0 30 99 Stringer and Kruse, 1971
Entamoebahistolytica OCl 4.0 10 9.0 30 99 Stringer and Kruse, 1971
Entamoebo hisfolytica NH.,CI 8.0 10 8.5 30 99 Stringer and Kruse, 1971
Entamoebohisfolytica NH I 2 6.0 tO 5.5 30 99 Stringer and Kruse, 1971
Entamoebohistolytica 0. 0.3 5 6.5-8.0 10-27 98->99 Newton and Jones, 1949
Entamoebahistolytica I-fOCI, 0C 1 3.4 >20-70 7.5-8.0 23-26 99 Brady et at., 1943
Enfamoebahisfolytica 1-tOCI 2.0 10 6.0 30 99 Kruseetal., 1970
Entomoebahistolyfico OCI 2.0 10 9.0 30 90 Kruseetol., 1970
Entamoebohistolytica NH. ,CI 8.0 10 9.0 30 97 Kruseetal., 1970
Entamoebahistolytica NHC 1 2 8.0 10 5.0 30 99.9 Kruseetol., 1970
Entamoebohistolytica I, 9.0 10 7.0 30 99 Stringer etal., 1975
Entamoebohistolytico liOCI 2.0 10 5.0 30 99.9 Stringer et ci., 1975
Entamoebo histolytica F+ 1 2 C 1 11.5 tO 8.0 30 99.9 Stringer et ci., 1975
-------
Table 8 (Cont’d). Indicator and Pathogen Reduction by Disinfection
Agent
Disinfectant
C
oncentration
(mg/liter)
Contact Time
(nun.)
pH
T mp.
rC)
% Reduction*
Reference
Gicrdia larnblia
UV
6 x
IO 4 pW-s/cm 2
N/A
7.0
20
75
Rice ond Hoff, 1981
Gkrdia Iamblia
Gicwdia tcn’nblia
Giardia tonthlia
Gkrdio lamblia
Gkwdia Ianthlia
Gicrdia larablia
Gicrdia tarablia
Gkwdia Ian,blio
Gkrdia krnblia ( )***
Gicrdia kimblia (a)***
Gicrdia muris
Giardia lomblia(s)
Gicrdia Iamblia(a)
Gicrdia muris
Gkrdici Iamblia (s)
Gicrdia lorablia (a)
Gicrdia muris
HOCI
HOCI
HOCI
HOCI
HOCI
HOCI
HOCI
HOCI
HOCI
HOCI
HOCI
HOCI
HOCI
HOCI
HOCI
HOCI
HOCI
1.5
1.0
2.0
4.0
8.0
2.0
4.0
8.0
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
10
60
60
60
60
10
10
10
12
15
25
IS
tO
23
22
19
40
6-8
6-8
6-8
6-8
6-8
6-8
6-8
6-8
6
6
6
7
7
7
8
8
8
25
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
99.7
99.7-90
99.8-99.9
99.8-99.9
99.8-99.9
90-89
97-90
99.8-97
90
90
90
90
90
90
90
90
90
Jarroll etoL, 1981
Jarroll etal., 1981
Jarroll et E., 1981
Jarroll etal., 1981
Jarroll etol., 1981
Jarroll etat., 1981
Jorrott etal., 1981
Jorroll etal., 1981
Rice et ol., 1982
Rice etal., 1982
Rice etal., 1982
Riceetal., 1982
Riceetal., 1982
Rice etal., 1982
Rice et at., 1982
Rice et at., 982
Rice et ol., 1982
*First number corresponds to the lowest designated pH value given.
= Not reported or unknown
= cysts from symptomatic case; (a) = cysts from asymptomatic case.
-------
22
Information on the effects of water treatment processes on specific pathogens will be
discussed In subsequent sections of this paper.
ANALYSIS
Bocterla
Salmoneua . There are over 2,200 known serotypes in this group, all of which are pathogenic
to humans, causing mild to acute gastroenteritis, and, very occasionally, death (Table I). Typhoid
and paratyphoid are both enteric fevers that only occur in humans. The other Salmonella serotypes
cause acute gastroenteritis, and the organisms can infect humans and a variety of other animals.
At any given time, 0.1% of the population will be excreting Salmonella . Water supplies implicated
in Salmonella outbreaks have been confined mostly to smaller systems, and the greatest risk in this
regard is from untreated, unprotected surface water supplies. In the U.S. there were II
waterborne outbreaks of salmonellosis involving a total of 1,370 cases during the period 197 1-79
(Craun, 1981b).
The 1965 outbreak in Riverside, California is the largest to date with 16,000 cases (Boring et
ol., .1971). This outbreak is especially notable because coliforms were absent before the outbreak,
and during the epidemic, coliforms were measured at only l.4/L while S. typhimurium was
measured at I ilL. There are several other instances in which Salmonella were isolated from water
supplies in the absence of coliform recovery (Seligmann and Reitler, 1965).
Although many studies have examined the occurrence of Salmonella in natural waters,
relatively few have quantified the presence of these organisms (Table 3). Brickler and Tunnicliff
(1980) reported between 1.5 and 300 organisms/IOO ml in the Ohio River, while Roland and Block
(1980) reported only 43 organisms/lOO ml for the Nancy River in France. The quantification of
Salmonella spp. in natural waters has been difficult, because most of the methods are more
qualitative in nature.
Several investigators have examined the survival of Salmonella species in natural waters.
The data in Table 5 indicate that there is some variation in reported die-off rates from different
studies for waters of similar quality. These variations may be due, in part, to differences in assay
systems and other experimental methods. Similarly, these limitations also apply to the data for
other organisms. At fairly high doses of l0 organisms/mi, Salmonella die-off rates are more rapid
-------
23
as temperature increases (Feachem et at., 1978). Water quality also influences the survival of
Salmonella , with increcised survival in water of higher quality. For example, S. fl p j die-off rates
at 18-21°C in tap water compared to those in Thames River water were 211 and 18 days,
respectively. Because of their similarity to E. coil and other coliform bacteria, it is generally
believed that Salmonella are reduced in water treatment processes to the same extent as
coliforms. The results of disinfection studies support this view. Butterfield (1948) found that a
tO minute contact time was necessary to produce a 99.9% kill of both S. typhi and E. coIl at pH tO
using 0.3 mg hypochiorous acid per liter (Table 8). Post-treatment contamination of finished water
is also of concern because microbial slimes in distribution systems have been shown to support the
growth of Salmonella .
Although the occurrence of salmonellosis in the U.S. is primarily due to contaminated food,
and the dose required to produce infection is high (except perhaps for Salmonella typhi) , the
intermittent occurrence of waterborne salmonellosis outbreaks makes these organisms a continuing
concern for drinking water quality.
ghigeua . This bacterial genus is divided into four main subgroups based on biochemical and
serological characteristics. Shigella are transmitted by the fecal-oral route and produce mild to
acute bacillary dysentery in humans. Because this syndrome is limited to humans and primates,
reservoir hosts in the environment are unimportant in its spread. Of the reported U.S. isolates in
1980, 69.4% were S. sonnel ; 27% S. flexneri ; 1.9% 5. boydii ; and 0.8% S. dysenteriae (Centers for
Disease Control, l98 1a). The dose of Shigella required to produce human infection is lower than
that for many other enteric bacteria.
Waterborne shigellosis is most often the result of obvious contamination from one identifiable
source, and the majority of outbreaks are from inadequately maintained and monitored semi-public
systems. For example, improper disinfection of fecolly contaminated well water resulted in a
shigellosis outbreak involving 1,200 people in Florida in 1974 (Weisman, et al., 1976). Waterborne
outbreaks of shigeltosis in the U.S. increased somewhat from 1961 to 1977 although this may be due
to increased reporting rather than a real increase in outbreak occurrence. As shown in Table 2,
from 197 1-79 there were 22 waterborne outbreaks of shigeltosis resulting in over 5,000 cases. On
an annual basis this number of documented waterborne cases is small compared to the annual total
number of cases reported (e.g., 20,135 cases in 1979). However, for most reported cases, the source
of exposure or route of transmission is not identified.
-------
24
Because of the difficulty in isolating Shigello from natural water, no quantitative data on
their occurrence in such waters are presented. However, Shigelki have been isolated from water
sources, especially during outbreaks (Table 4).
Several studies have examined the die-off rates of Shigella in both natural and finished
waters. These studies showed that Shigella survival times ore lower than those for other enteric
bacteria. However, they have been reported to persist for as long as 28 days in filtered river
water, 14 days in unfiltered river water and 6 days in tap water at 23-25°C when initial
concentrations were high (lO organisms/mi) (Table 3) (Mohadjer and Mehrabian, 1975).
Few water treatment studies, including disinfection studies, have used Shigella as a test
organism. This may be due to its relatively poor survival in water, the effectiveness of chlorine in
killing it (Table 8), and its perceived lack of importance as a waterborne pathogen. Because of its
similarity to coliforms, it would be expected that removal of Shigello by coagulation and filtration
would be similar to them.
The survival of Shigella in water and their Susceptibility to disinfection and other water
treatment processes appear to be similar to coliform organisms. Therefore, routine monitoring of
coliforms should be adequate to protect drinking water from most contamination situations
involving these organisms.
Yersinia enterocoUtjca . This gram-negative bacterium can cause acute gastroenteritis and
can be found in water in cold or temperate areas of the United States. Many wild, domestic and
farm animals are known to be reservoirs of this organism, including wild animals associated with
water habitats (beavers, minks, muskrats, nutrias, otters, and rocoons) (Wetzler and Allard, 1977;
Wetzler, et ol., I 978a). The potential of waterborne disease is enhanced by the increased virulence
of the organism of 20°C and its ability to grow at temperatures as low as 4°C. At this
temperature, the organisms will grow with a generation time of 3.5 to 4.5 hours,, if at least trace
amounts of organic nitrogen are present (Wetzler, 1981). There were two reported incidents of
waterborne gastroenteritis in the U.S. during the period (971-78 possibly caused by Yersinia (Eden,
et al., (977; Keet, (974), although the documentation for a Yersinia etiology in both outbreaks is
poor.
Yersinia has been isolated from untreated surface and ground waters in the Pacific
Northwest, New York and other regions of North America, with highest isolations during the colder
-------
25
months (Harvey, et al., 1976; Wetzler, et at., 1979; Prior, 1976; Shayegani, et at., 1981).
Concentrations of the organism in positive surface water samples have ranged from 3 to 7,900 per
100 ml. Yersinia isolations from water are poorly correlated with levels of total and fecal
coliforms or total plate count bacteria (Wetzler et ci., 1979; Prior, 1976).
There is little information on Yersinia survival in natural waters and water treatment
processes. As shown in Table 5, one report indicated that Yersinia survived as long as 157 days in
sterile water (Dominowska and Mallotka, 1974).
In field studies on Yersinici occurrence in chlorinated-dechlorinated secondary effluent and
receiving (river) water, the organism was isolated in 27% of the effluent samples, 9% of the
upstream samples and 36% of the downstream samples (Turnberg, 1980). Mean total and fecal
coliform reductions in effluent chlorination were 99.93 and 99.95% respectively, thus indicating
effective coliform disinfection. In a study of treated (chlorinated) and untreated drinking waters,
50% of the samples from untreated waters contained potentially virulent Yersinia strains and 50%
contained “environmental” (i.e., non-virulent) strains (Fulgham, 1980). From treated water,
however, Isolation frequencies of virulent and environmental strains were 66.7 and 33.3%,
respectively. These findings suggest that virulent Yersinia strains may be more chlorine-resistant
than environmental strains. Unpublished preliminary data Indicate that Yersinia resistance to free
chlorine is similar to that of E. coil at pH 5.5 and 7.0 but somewhat greater at pH 8.5. At
temperatures< 20°C, Yersinia is more resistant to monochioramine than is E. coil (Wetzler, 1981).
In a survey of untreated and treated (chlorination or filtration plus chlorination) drinking
water supplies, the occurrence of Yersinia-positive samples was 14.0 and 5.7%, respectively
(Wetzler, et ci., 1979). Water samples with<2.2 coliforms per 100 ml were 6.9% Yersinia-positive,
and those with >2.2 coliforms/l00 ml were 15.9% Yersinia-positive. Yersinia isolation did not
correlate with total or fecal coliforms in this study.
Considering the existence of animal reservoirs, the widespread occurrence and persistence of
Yersinla in natural and treated water in at least some geographic areas, the evidence for possible
waterborne outbreaks, and the lack of definitive information on its reduction by treatment
processes, this pathogen is of potential importance in drinking water.
Campylobacter JeJuni. Campylobacter jejuni is one of several species of gram-negative,
curved, rod-shaped bcicterici that infect humans or other animals. C. j j j has recentt ’ been
-------
26
recognized as a common bacterial cause of acute gastroenteritis, usually producing diarrhea,
abdominal pain and fever, with a typical incubation period of 2-5 days and about one week of
illness. Untreated cases may continue to excrete the organism for as long as several weeks.
C. jejuni also infects a variety of wild and domestic animals including dogs, cats, sheep, goats,
cattle, swine, fowl and rodents.
Several waterborne outbreaks of C. jejuni gostroenteritis have been reported in the U.S. and
other countries (Blaser and Peller, 1981). One U.S. outbreak among campers in Wyoming was
associated with consumption of untreated mountain stream water, and the organism was sub-
sequently isolated from streams in the incriminated area (Taylor, et cii., 1981). Two other U.S.
outbreaks involved municipal water supplies. In one of these outbreaks, the water system appeared
to be intact, there was no obvious source of fecal contamination and 33 of 34 water samples were
coliform negative and low in standard plate count bacteria. One sample had 8 coliforms/lOO ml
(Kornblatt et cii., in press). However, chlorine residuals in the distribution system were low,
averaging 0.02 and 0.05 mg/I during the two successive months of the outbreak, and two possible
cross-connections were identified. The other municipal outbreak involved inadequately disinfected
surface water where sanitary surveys revealed several possible sources of feccil contamination of
the watershed (Tiehan and Vogt, 1978). However, routine coliform tests preceeding and during the
outbreak were neCative.
Little information is available on ambient levels and survival of C. jejuni in natural water.
The organism has been detected in fecally contaminated natural waters in New England, the
Roekies, the Pacific Northwest, and it may occur elsewhere in the U.S. Widespread distribution of
C. jejurii in natural wafers is indicated by the results of a field survey in the South of England in
which the organism was isolated from about 50% of the water samples tested (Knill, et al., 1978).
Positive samples also contained coliforms, thus demonstrating the ability of indicator bacteria to
detect fecal contamination of natural waters associated with the presence of C. jejuni .
Information on C. jejuni survival in natural waters is limited. Recent laboratory studies
(Table 5) indicate that the time for 99.9% inactivation of the organism ranged from about
3- 18 days in autoclaved stream water kept at 4°C to about 2-3 days in the same water kept at
25°C (Blaser etal., 1980).
-------
27
There is only limited information on the effects of water treatment processes on C. jejuni ,
except for disinfection. Recently, Wang, et al. (1982) found that 100% kill of “low” concentrations
(10 -10 /ml) of C. jejuni was obtained with 2.5 and 0.625 mg/I chlorine in I and 30 minutes,
respectively. 100% kill of “high” concentrations ( 10 6 - 10 7 /ml) of C. jejuni was obtained with
2.5 mg/I chlorine in 30 ninutes, and 2 of 3 strains tested at high concentrations were completely
killed with 5 mg/I in I minute. These findings suggest that conventional chlorination of adequately
pretreated water will probably destroy C. jejuni . However, its relative resistance to chlorine
compared to coliform bacteria is not known, and waterborne outbreaks of gastroenteritis due to
C. jejuni have been attributed to water supplies that were negative for coliforms when sampled
several weeks after the outbreaks.
Considering the recent documentation of waterborne outbreaks, the possible widespread
distribution of C. jejuni in natural waters of at least some geographic areas, the uncertainties
about the ambient levels, sources and survival of the organism in natural waters, and inadequacy of
information on its removal and destruction by water treatment processes, this pathogen is of some
concern in drinking water supplies.
Legionella pneumophila and Other Legionella Species. Seven species of this gram negative
rod have now been recognized as etiological agents of Legionnaire’s Disease (LD). The clinical
features of this syndrome do not differentiate this disease from other acute types of pneumonia.
The incubation period is 2 to 10 days and it is a multisystem illness which can also involve the
gastrointestinal tract, kidneys, and central nervous system. In some outbreaks, infected individuals
develop pneumonic LD, while in others the illness is milder, self-limiting, nonpneumonic and of
1-2 days duration. The latter illness is known as Pontiac Fever. The incubation period of LD is 2
to 10 days and for the mild form it averages about 1.5 days.
Between 1976 and 1979, nearly 2,000 epidemic-associated cases from separate outbreaks of
LD were reported in the U.S. Most epidemics have occurred in the summer months, but some have
also occurred during other seasons. In recent years, 200 to 3,000 cases of LD have been reported
annually in the U.S., but the true number of annual cases is estimated to be about 25,000 (Foy, et
al., 1979).
Most LD outbreaks have been associated with aerosols from cooling towers and other
evaporative type cooling water systems where high concentrations of organisms were present in the
-------
28
cooling water. Thus, potential Legionella exposure may occur in a variety of situations where
waters containing high concentrations of the organism become aerosol ized. No outbreaks of LD
have been attributed to the consumption of drinking water, but there hove been outbreaks
associated with water aerosols from hospital shower heads (Tobin, et al., 1981; Cordes, et al.,
1981). L. pneumophila has been isolated from potable water sources in association with outbreaks
of LD in resort hotels in other countries, but there is no conclusive epidemiologic evidence
establishing potable water as the vehicle or identifying the precise route of water exposure
(Centers for Disease Control, 1980; 1981b).
Legionella pneumophiki is apparently a naturally occurring and widely distributed organism
that presumably grows in the aquatic environment. Environmental isolates of Legionella have
essentially the same morphological, physiological, biochemical and serological characteristics as
clinical isolates. In a recent survey of 67 lakes and rivers in the U.S., primarily in the southeast,
Legionelki was found in virtually all of the waters at concentrations ranging from 9.1 x l0 to
3.0 x io6 cells per liter by direct fluorescent antibody (DFA) analysis (Fliermans, et al., 1979;
1981). Legionella isolates were also obtained from 15% of sample concentrates tested by direct
inoculation of guinea pigs. Isolation frequencies from samples were significantly higher from
warmer (36 to 70°C) than cooler (0 to 36°C) waters and did not correlate with other water quality
parameters, including conductivity, dissolved oxygen, pH, chlorophyll a, pheophytin and Secchi disk.
Guinea pig isolations from monthly samples from a single water source showed a marked seasonal
variation, with higher isolations in the spring and summer than in the fall and winter (Fliermans, et
al., 1981).
There appears to be a relationship between Legionella occurrence in natural waters and algal
growth, as Legionelki has been isolated from algal-bacterial mat communities (Tison, et al., 1980).
Using mineral salts medium in laboratory batch cultures, Legionella readily grows in the presence
of a number of algae, apparently by using algal extracellular products as substrates. Berendt (1981)
reported that the survival of L. prieumophila is enhanced by a low molecular weight substance from
the alga Fischerella.
Legionella can survive for relatively long periods in water, with survival times of 250 days in
pond water, L l5 days in tap water and 69-139 days in distilled water (Skaliy, et al., 1979).
Relatively little is known about Legionella removal in water treatment processes, except for
-------
29
disinfection. In laboratory studies (Table 8) with chlorine demand-free tapwoter at 25°C, Skaliy, et
al., (1980) found that 3.3 mg/I of free chlorine gave a 6.1 log decrease in Legionella at “0 time”
(i.e., samples taken just after agitating the reaction mixture to disperse the added cells). Lower
chlorine concentrations were not tested in this study. Recent field studies on chlorination of
cooling water systems containing high levels of Legionella (108 organisms/I) indicate that free
chlorine residuals of 1.0-1.5 mg/I produce 99% reductions of the organism in 19 hours (Table 8).
Although information on Legionella occurrence in raw source waters, removal and inactiva-
tion by water treatment processes and persistence in distribution systems is limited, there is no
evidence that ingestion of finished drinking water is a significant route of LD. However,
widespread distribution of the organism in aquatic environments and the occurrence of outbreaks in
hospitals and other public use facilities suggesting potable water as a vehicle, indicate the need for
further information on the occurrence and possible proliferation of this organism in water supply
systems and on the risks to the population of LD from various types of potable water exposures,
especially aerosols.
Nontubercular Mycobacteria. The mycobacteria are gram-positive aerobic, rod shaped
organisms with thick cell walls and a high lipid content that confers their acid-fast characteristic.
These organisms are characterized by filamentous growth, a preference for lipids and relatively
slow growth rates. The nontubercular mycobacteria were previously considered saprophytic and
therefore, of no public health concern. However, it is now recognized that some of these
organisms are opportunistic pothogens that can cause pulmonary and other diseases in individuals
with preexisting health problems.
In a review of nontuberculous mycobocterial disease Wolinsky (1979) emphasized that two
species, Mycobacterium kansasii and Mycobacterium avium—intracellulare were the predominant
cause of such disease in this country. M. avium-intracellulare is the most frequently isolated
mycobacterium other than M. tuberculosis , and its reported incidence has increased markedly in
recent years (Good, 1980; Stottmeier, et c i ., 1980). Individuals at the greatest risk from infection
due to M. avium-intracellulare are young children and the aged. The type of illness manifested in
each population is different: lymphadenitis in younger age groups and atypical pulmonary
tuberculosis in older individuals.
-------
30
Two epidemiologic observations stand out with respect to M. avium-intracellulare and
M. kansasii . First, there is a distinct geographic pattern of disease distribution, with most cases of
M. kansasii appearing in the midwest, and most cases of M. avium-intracellulare appearing in the
southeast and midwest. However, both organisms appear to be more prevalent than was previously
recognized. Second, there is little or no evidence that these organisms are transmitted from
person-to-person by the respiratory route. This suggests that infection is acquired by exposure to
contamination from a presently unknown environmental source.
Recent studies have shown that potable water can support the growth of these organisms and
transmit them through municipal distribution systems. In addition, waterborne M. avium-
intrcicellulare has been epidemiologically linked to nosocomial infections (duMoulin, et al., 1981).
During an investigation of a pseudoepidemic of M. avium-intracellulare , the organisms were
observed in hospital water supplies. Sputum and urine specimens sent for acid fast bacillus (AFB)
analysis were found to contain these organisms that had no apparent relevance to the clinical
conditions for most patients. Because of the epidemic characteristics, a common source of
contamination was suspected. Subsequent AFB cultures of hospital water mains showed abundant
growth of M. avium-intracellulare identical to the contaminating strain in clinical specimens. The
organisms were subsequently found in a survey of bedside carafes and nebulizer reservoirs, as well
as the hospital water mains. Water supplies from other hospitals also contained these organisms
(duMoulin and Stottmeier, 1978). Mycobacterium avium-intracellulare and M. kansasii have been
isolated recently from a variety of water sources in the U.S., and some investigators believe that
aerosols of contaminated waters may be responsible for the spread of the disease (George, et ol.,
1980; Powell and Steadham, 1981).
The acid fast bacilli that are capable of causing skin infections, such as Mycobocterium
marinum ( balnei) , are clearly capable of being waterborne and may be transmitted by this route,
especially in swimming waters (Dailloux, et al., l980a; 1980b). However, for the nontubercular
mycobocteria causing respiratory infections, the significance of waterborne transmission remains
uncertain. For individuals with preexisting health conditions, especially pulmonary disease or
neoplasms, there may be a risk of infection by aerosol transmission from contaminated water. At
present, however, there is no evidence of risk to healthy humans who might ingest or respire water
containing nontubercular mycobacteria.
-------
31
Opportunistic Pathogens. This heterogeneous group of gram-negative bacteria is ubiquitous
in the environment, causing disease infrequently and under unusual circumstances. Several
bacterial genera, including Pseudomonas, Aeromonas hydrophila, Edwardsiella farda, Fiavo-
bacterium, Kiebsiel Ia, Enterobacter, Serratia, Proteus, Providencia, C itrobacter and Aci ne-
tobacter , are identified as opportunistic pathogens. Post-operative patients, cancer or leukemia
patients, newborn babies, and the elderly and infirm are major segments of the population at risk
f or infection. The numbers of organisms needed to establish infection in these individuals are
generally lower than in healthy children and adults. Therefore, the occurrence of high levels of
opportunistic pathogens in water supplies at institutions where the host population is more
susceptible deserves special consideration. In addition, certain routes of exposure to endotoxins
and pyrogens produced by these bacteria can pose a significant health hazard to compromised
individuals in specialized environments such as hemodialysis units.
A number of recent studies have shown that opportunistic pathogens are commonly isolated
from and can grow in finished water supplies (Colwell, ef al., 1978; Olson and Hanami, 980). For
example, P. aeruginosa is capable of growing in distilled and other water with concentrations
reaching I 0 8 /ml (Favero, et al., 1971; 1975; Carson, et al., 1975). Ingestion of or contact with high
concentrations of this organism in water can cause enteric, eye, ear and upper respiratory tract
infection. Enteric infections caused by these organisms are usually mild in nature and are often
not reported to medical practitioners. In fact, approximately 10% of the population are carriers of
P. aeruginosa and shed the organism in the feces (Hugh and Gilardi, 1974).
The occurrence of these opportunistic bacteria in institutional drinking water is common.
For example, routine analysis of tapwater from a newly completed hospital addition yielded
between 3,000 and 4,000 colony forming units per 100 ml of water (Eichhorn, 1977). Sub-
sequently, water samples from new isolation rooms showed total plate counts of between 3-4 x l0
colony forming units per IOU ml of water. Pseudomonas species were the predominant bacteria
recovered. Ice from new ice machines was equally contaminated. Opening of the facility was
delayed one week so that a purging protocol could be instituted. This involved the running of fully
open taps for at least 15 hours with subsequent retesting for bacteria content. The retesting
showed a significant drop in colony forming units to a level of between 4-5 x 102 colony forming
units per 100 ml of water. Recent reports of hospital-acquired ! 9 ionella pneumophila infections
-------
32
in new buildings suggest that standing water in piping systems of new facilities may encourage the
growth of Legionella and opportunistic pathogens to unacceptable levels.
While not demonstrating the virulence of the pseudomonads, members of the genus Flavo-
bacterium were observed as frequent colonizing organisms of the upper airways of seriously ill
hospital patients. During 70 months, 8.4 percent of alt admissions to on intensive care unit were
colonized with Flavobacterium (deMoulin, 1979). Because of its unusual pigmentation, biochemical
properties, and unique antibiotic susceptibility pattern, the organism was ideally suitable as an
epidemiologic marker. Flavobacterium was recovered from the tap water, sink drains, ice
machines, water baths, and from the hands of intensive care unit personnel. The source of these
bacteria was ultimately traced to five of the nine finished water reservoirs serving the
metropolitan Boston area. Because this group of bacteria is relatively chlorine-resistant, it travels
unimpeded to the consumer. In the hospital environment widespread antibiotic use appeared to
select for these organisms.
The public health significance of opportunistic pathogens to the population at large is
unknown. These organisms usually cause infection in individuals already compromised by
underlying illness. Increasing evidence indicates that this group of organisms is important in
nosocomial infections. Because these bacteria usually predominate in finished water, the standard
plate count (SPC) provides a good measure of their occurrence. High quality supplies, meeting the
coliform standard, usually contain between I and 500 SPC organisms per milliliter.
Because many of these organisms are able to colonize distribution lines, increases in plate
count organisms in distribution water indicate a deterioration of water quality as the water travels
from the plant to the consumer. The opportunistic pothogens are of some concern in finished
drinking water because of their ubiquitous nature, their ability to grow in finished water and their
potential to produce disease in certain high risk segments of the population in specialized
environments. The presence of these bacteria in drinking water does not appear to pose a
significant risk of infection when ingested by healthy individuals.
Drug Resistance. Drug resistance in waterborne bacteria was first described through
differential resistance patterns among coliforms organisms in rural and urban streams of the
United Kingdom (Smith, 1970). Since this initial work, drug resistant bacteria have been shown to
occur in a variety of aquatic environments, including drinking water (Gonzal, etal., 1979; Grabow,
-------
33
et al., 1974; Hinshaw, et ci., 1969; Kelch and Lee, 1979; Armstrong, et aI., 198 I). Drug resistance
is not restricted to coliform bacteria and is common among most bacterial genera.
Information on the likelihood of increased bacterial resistance to antibiotics with increased
antibiotic usage is still somewhat inconsistent. Antibiotic selective pressure has not been shown to
increase enferotoxigenic E. cofl in a population. The extent to which antibiotic usage contributes
to the emergence of drug resistance in bacteria) populations appears to be related to the
therapeutic regimes being practiced in a given locale (Meyer and Lerman, 1980).
Armstrong, et ci. (1981) reported an increase in drug resistance to two or more antibiotics in
plate count bacteria in distribution water compared to raw water. The question of increased
introduction of drug resistant organisms info drinking water through widespread use of antibiotics
remains a concern. However, there is no information to date which suggests that the occurrence of
drug resistant organisms in water supplies is a health threat. In addition, because many of these
bacterial genera have the ability to proliferate In the distribution system itself, a vigilant program
to minimize bacterial regrowth would suggest an effective means of control of this potential
problem. A drinking water standard for plate count organisms would, further Insure minimal
exposure of the consumer to high numbers of drug resistant plate count organisms.
Enteric Viruses
The viruses of major concern In drinking water are the so-called enteric viruses that Infect
the gastrointestinal tract of humans and In some cases other animals and are excreted In the feces
of infected persons or animals. As shown In Table I, the enterlc viruses belong to several animal
virus taxonomic groups and in total there are more than 100 different serological types.
The enteric viruses are most commonly transmitted by direct and indirect person-to-person
contact via the fecal-oral route (Melnick, 1976; Lennette, 1976). Exposure is primarily among
young children with subsequent secondary exposure of older children and adults, especially within
households and other living units. This route of exposure Is so common and widespread that enteric
virus infections are endemic in virtually all sizable communities. Infections usually occur without
obvious clinical symptoms. Only 0.1 to 10% of the infected population may show signs of illness,
depending upon the type of virus, the health and immune status of the population and environ-
mental factors. Enteric viruses can also be transmitted via fecalfy contaminated water and food.
-------
34
Hepatttis Type A. The best documented evidence for enteric virus transmission by drinking
water is for hepatitis A virus (HAy). HAV causes “infectious hepatitis”, an acute illness
characterized by inflammation and subsequent necrosis of the liver. The virus is usually ingested
and probably initially infects the intestinal tract, producing a viremia and subsequent liver
infection. Major clinical symptoms are fever, malaise, and gastrointestinal symptoms such as
nausea, vomiting and diarrhea. Jaundice, due to liver dysfunction, occurs in some but not all cases.
During the last several years about 30,000 total cases of hepatitis A have been reported
annually in this country (Centers for Disease Control, 1981c), and in most cases the route of
exposure is unknown. Epidemiological evidence for HAV transmission via contaminated drinking
water is extensive (Goldfield, 1976; Mosley, 1967) and between 1971 and 1979 there were
14 reported outbreaks involving 398 cases in the United States.
As with most waterborne outbreaks of infectious disease, those due to HAV usually involve
extensive and obvious contamination of the water supply as evidenced by excessive levels of
indicator bacteria in the water, obvious sources of contamination and either lack of or deficiencies
in treatment. In a recent HAV outbreak due to contaminated well water in Georgetown, Texas
bacteriological samples of tap wafer were consistently coliform-negative and no obvious source of
fecal contamination could be identified (Hejkal, et al., 1982; Wilson, et al., 1981). However, raw
well water had high levels of fecal coliforms (up to 1,500 per 100 ml) prior to chlorination and
there were deficiencies in chlorination practice. These findings suggest that HAV is more resistant
to chlorine than are coliform bacteria. Consequently, coliforms may not adequately predict the
presence of HAV in chlorinated drinking water.
Information on HAV levels in natural and finished waters, its survival in water and removal in
treatment processes is not available due to lack of convenient laboratory hosts and assay systems
for this virus. However, based upon its morphological and biochemical similarity to enteroviruses
(Siegel, et al., 1981) and early studies on survival and persistence in water and in treatment
processes (Neefe, et al., 1945; 1947), HAV is thought to be more persistent in water and less
efficiently reduced by treatment processes than enteric bacteria. In fact, recent evidence
indicates that HAV may be somewhat more resistant to chlorination than other enteric viruses
(Peterson, et at, 1982). Because of the continued occurrence of waterborne HAV outbreaks in the
-------
35
U.S. and uncertainties about HAV levels in raw and finished water and it s removal by water and
sewage treatment processes, this pathogen is still of considerable concern in U.S. drinking water
supplies.
Norwalk and Related Gastroenteritis Viruses. The Norwalk-type viruses are a major cause of
acute, epidemic gastroenterif is (GI) in older children and adults. In temperate climates these
outbreaks occur mainly in the winter months (Holmes, 1979). There are cit feast three serologically
distinct viruses in this group, and preliminary characterization indicates that one of them, the
Norwalk agent, may be a calicivirus or a parvovirus (Greenberg, et cii., 1981). These viruses have not
been grown or assayed in cell cultures, but recently developed immunochemical methods ore now
available for detecting and assaying viral antigens and antibodies.
Several recent waterborne outbreaks of gastroenteritis in the U.S. have been attributed to
Norwalk-type agents based upon seroiogical evidence from ill persons (Table 2). Because of the
unavailability of cultivation methods for these viruses, there is no information on their occurrence
and persistence in raw and finished waters or their removal by treatment processes. It is possible
that many outbreaks of woterborne GI previously categorized as “unknown etiology” may actually
have been caused by these viruses. Documentation for their etiology in recent waterborne
outbreaks and lack of any information on their occurrence, survival and fate in water and
treatment processes make Norwalk and related viruses important waterborne pathogens.
Rotaviruses. Rotoviruses are a major cause of acute gastroenteritis, primarily in children.
Rotavirus diarrhea of infants is endemic worldwide and in less developed countries is a major cause
of infant mortality. Reinfection is possible in both older children and adults. High concentrations
of the virus are shed in the feces (up to lO particles/gram) when illness occurs and for about a
week thereafter. In temperate climates there is a peak seasonal incidence of infection during the
colder months. Human rotaviruses are difficult to grow and assay in cell cultures, but
immunochemical methods are available for detection of rotavirus antigens and antibodies. There
are at least 3 human rot avirus serotypes and a number that infect other animals. The animal
rotaviruses are antigenically similar but not identical to the human serotypes and their role in
human illness is uncertain. Because of difficulties in growing and assaying human rotaviruseS in
cell cultures, information on their occurrence and survival in water, their reduction by treatment
-------
36
processes, and their possible role as etiological agents of waterborne gcistroenteritis outbreaks is
unknown. However, rotaviruses have been detected in finished drinking water in Mexico and in U.S.
sewage and treated effluents (Smith and Gerba, 1982; Gerba, 1981), and recent laboratory studies
suggest that they are not as effectively removed by certain water and wastewater treatment as are
enteroviruses (Farrah, et cii., 1979). Because of their significance as agents of gastroenteritis and
the lack of information about their transmission via water, the rotaviruses are of special concern.
Adenoviruses. The well-characterized adenovirus serotypes are known to cause primarily
upper respiratory illness, but concurrent enteric infections with fecal shedding may also occur.
More recently, newly recognized odenovirus serotypes that were previously uncultivatable have
been implicated as causative agents of acute gastroenteritis (Yolken et al., 1982). Adenoviruses
have been detected in wastewater and contaminated natural waters but not in finished drinking
water. There are no documented drinking water outbreaks of adenovirus illness, but several
outbreaks of adenoviral pharyngoconjunctivitis have been attributed to contaminated swimming
pools. The role of adenoviruses in outbreaks of drinking water gastroenteritis of unknown etiology
has not been established.
Entarovtruses. The enteroviruses comprise a large group of enteric viruses, many of which
can infect both the enteric and upper respiratory tracts. In enteric infections the viruses are
detectable in the feces a few days before illness and for 1-2 weeks thereafter. Although
enteroviruses are readily detectable in raw and treated sewage effluents and contaminated surface
waters and have also been detected in finished drinking waters, their role in waterborne disease
outbreaks is uncertain. Drinking water has been implicated as the route of transmission for
poliomyelitis, but substantive epidemiological evidence is available for only a few reported
outbreaks and is by no means conclusive (Mosley, 1967).
Reovtruses. The role of reoviruses in human illness is uncertain but they may produce both
respiratory and enteric infections. These viruses have a wide host range that includes mice, cattle
and primates. Like enteroviruses, the reoviruses are readily detectable in feces within the first
week or two of infection and they have been isolated from sewage and contaminated water,
including drinking water. Their significance as disease-causing agents in water is uncertain but
probably not as important as other enteric viruses.
-------
37
Overall Analysis. Compared to enteric bacteria, human enteric viruses are relatively more
resistant to inactivation by physical, chemical and biological factors in the aquatic environment,
and are less efficiently reduced by water treatment processes such as chlorination and rapid sand
filtration. Although virus reductions in pilot scale water treatment systems using coagulation,
filtration, carbon adsorption and chlorine disinfection have been extensive p99.999%) (Guy, et al.,
1977), recent field studies on virus levels in both source and finished water from a conventional
treatment plant using coagulation, filtration and chlorination indicate relatively poor virus
reductions and readily detectable virus levels in finished water (Payment, 1981).
Of special significance to ground water is the fact that enteric viruses are less efficiently
removed in soils than are enteric bacteria such as fecal coliforms (Sobsey, et at., 1981).
Furthermore, recent field studies have shown a lack of statistical correlation between indicator
bacteria levels and the presence of viruses in groundwater wells (Marzouk, et at., 1979), thus
suggesting the failure of bacterial indicators to adequately represent viral contamination.
Considering the continued occurrence of waterborne viral gastroenteritis and hepatitis, the
‘documented occurrence of viruses in finished drinking water, the stability of viruses in natural
water and uncertainties about the effectiveness of virus removal and destruction in water
treatment processes, at least some of the enteric viruses, especially HAV and Norwalk-type
viruses, are of particular concern in drinking water.
Protozos s
Giardia lamblia. Giardia lamblia is a flagellated protozoan found in the small intestine of
man and other mammals. Trophozoites are 9 to 21 pm long, 5 to 5pm wide and 4 to t4pm thick.
Cysts ore ovoid, 8 to 12pm in length and 7 to 10pm in width.
The clinical syndrome giardiasis involves mechanical interference with absorption and
enzyme activity which can lead to weight loss, malnutrition and anemia. Symptoms include chronic
diarrhea, steatorrliea and abdominal cramps. Levels of Giardio cysts in feces from infected
persons may be as high as 1-5 x 06 cysts per gram and the cyst level in domestic raw sewage has
been theoretically estimated at about IO cysts per liter (Jakubowski, in press).
There has been a steady increase in the number of waterborne outbreaks of giardiasis in the
United States since 1971 (Craun, 1979). Growing recognition of this disease, through the
-------
38
identification of giardiasis in travelers to the Soviet Union, has been responsible for increased
surveillance, investigation, and reporting by U.S. public health authorities. Only one foodborne
outbreak has been reported (Osterholm, et. al., 1981) and water appears to be a major vehicle for
transmission of giordiasis. There were 31 woterborne outbreaks and 17,822 cases of giardiasis from
1972 to 1979 (Table 2). Wciterborne giardiasis has occurred primarily in mountain areas of this
country, including New England, the Pacific Northwest, and the Rocky Mountains. Traditionally,
these areas have utilized surface water sources with minimal treatment, primarfly disinfection
only. Giurdiasis in both endemic and epidemic forms in the United States suggests consumption of
untreated or inadequately treated drinking waler to be an important cause of endemic infections.
Extensive bacterial contamination of water and obvious sources of fecal contamination have
been found in some outbreaks of giardiasis. However, most outbreaks are characterized by little or
no bacterial contamination of finished water. This suggests that when certain conventional
treatment processes, such as chlorination or pressure filtration plus chlorination, are applied to
contaminated source waters, coliforms are inadequate indicators for Giardia in finished waters.
Although ambient levels of Giardj cysts in natural waters or finished water supplies are not
known, Giardia cysts have been isolated from water in 8 of 20 reported outbreaks. in one outbreak
where Giardia cysts were isolated from the raw water source, the concentration of the cysts
recovered was extremely low, I cyst/l.I x 106 L (Shaw, et cil., 1977). Using improved recovery
methods, Giardia cysts were also detected in the water of a distribution system during an outbreak
(Lippy, 1978). Generally, the bacteriological quality of drinking water incriminated in giardiasis
outbreaks has either not been measured, or when measured has met coliform standards.
Information from waterborne outbreaks and from surveys of Giordia prevalence in wild
animals suggests that animal reservoirs, notably beavers and muskrats, are a major source of
Giardia contamination of water supply systems. In addition to the role of animal reservoirs in the
contamination of raw water sources, the relatively long survival times of cysts in water (1-3
months, Table 5) further enhance the potential for Giardia contamination of water sources in
remote areas. Other factors which favor the transmission of giardiosis via water inck d relatively
high concentrations of cysts in fecal material and their low infective dose (Table I).
Disinfection studies have shown that Giardici cysts are considerably mare resistant to chlorine
than ore indicator bacteria or even enteric viruses (Table 8) (Jarroll, et c i., 1981). Although
-------
39
laboratory and pilot scale studies have shown that Giardia cyst can be efficiently removed by
diatomaceous earth filtration, investigation of waterborne outbreaks suggests that pressure
filtration using granular media may be less effective. Recent laboratory and pilot scale studies on
Giardia cyst removal y granular media filtration indicate relatively poor cyst removal when
coagulant doses are low or less than optimum, when filtered water turbidities increased during
filter runs, and during filter ripening periods. Effective cyst removal was not always achieved
when the I NTU MCL for turbidity was met (Logsdon, et al. , 1981). However, these studies also
indicate that granular media filtration is capable of achieving efficient 99%) cyst removal under
conditions of optimal coagulant doses, very low finished water turbidities (0.3 NTU) and con-
sistently stable filter operation conditions.
Coliform bacteria may be inadequate indicators of the microbial quality of drinking water in
those geographic areas where Giardia is endemic in wild animal populations. This is because of the
ability of Giardia cysts to survive for long periods in natural waters and to resist such water
treatment processes as filtration and chlorination. For these reasons, waterborne transmission of
giardiasis is of major concern in at least some areas of the United States.
Entamoeba hlstolytica. Entamoeba histolyticc is the protozoan causing amoebic dysentery.
Two forms exist: the relatively fragile trophozoite, l5-25i.tm in length, and the hardy, infective
cyst, l0-l5 im in diameter. Although many infections are asymptomatic, intestinal illness COfl
occur, with symptoms ranging from acute, fulminating dysentery with fever to mild gastroenteritis.
In addition, some infections result in invasion of the mucosal lining of the intestine and distribution
of the organism via the blood stream to other organs. The main clinical feature of this latter type
of infection is the formation of amoebic abcesses, primarily in the liver.
Humans are the only reservoir and the organism is widely distributed in the population, with
prevalence rates ranging from 3 to 10% in the U.S. The reported annual case rate for amoebiasis in
the U.S. is 2,000-3,000. Asymptomatic carriers and persons with mild illness may fecally shed
cysts for months or years.
E. histolytico is transmitted by the fecal-oral route and outbreaks due to contaminated water
have occurred. The last reported waterborne outbreak in the U.S. was in 1953. Waterborne
outbreaks have invariably been associated with obvious fecal contamination of untreated or
inadequately treated supplies.
-------
1 0
Information on E. histolytica occurrence in fecally contaminated natural waters is limited,
but concentrations in raw sewage have been estimated at 5,000 cysts per liter and measured
concentrations in treated effluents have been as high as 2.2 cysts per liter.
E. histolytica cysts ore efficiently removed by conventional filtration processes, with
removal of 80 to nearly 100% reported. Like Giardia lamblki , E. histolytica cysts are more
resistant to disinfection than are enteric bacteria or viruses. Cyst destruction of 99.999% requires
a contact time of about tOO minutes with a free chlorine residual of 0.5 mg/I and a temperature of
23°C. Cyst destruction is even less effective at lower temperatures or with combined chlorine
residuals.
Because of the absence of waterborne outbreaks in the U.S. in recent years, the absence of
animal reservoirs and the effectiveness of filtration for cyst removal, E. histolytica does not
appear at present to be a pathogen of major concern in adequately treated water supplies using
filtration. However, water supplies using only chlorination and potentially susceptible to human
fecal contamination are still of some concern with respect to E. histolytjca transmission.
Balantidium coli. This is the only known ciliated parasite that may infect humans. It is SOpm
or more in diameter and may cause acute or chronic dysentery in chronically ill or debilitated
patients. It is primarily a parasite of pigs. Transmission appears to be fecal-oral, and there has
been only one reported waterborne outbreak of balantidiasis (Center for Disease Control, 1972).
This outbreak occurred in the Truk District of Micronesia, and was attributed to contamination of
the water supply with pig feces after a typhoon.
Naeglerta fowleri . This facultative parasitic amoeba is one of the causative agents of
primary amoebic meningoencephalitis in children and young adults. The trophozoite measures 8 to
l5 .tm. It is a free-living amoeba found in soil, water and decaying vegetation. Exposure to the
organism is usually from swimming in fresh water lakes containing high concentrations of the
organism. Field surveys in southeastern fresh water lakes indicate that Noegleria is ubiquitous in
such waters. The organisms proliferate rapidly as water temperatures rise. The risk of infection
from swimming in infected lakes has been estimated to be less than I in 2.5 million exposures. The
route of entry for infection is the nasopharynx. The organism penetrates the olfactory epithelium
and enters the brain via the olfactory nerve plexus (Chang, 1978). The disease symptoms include
-------
41
headache, nausea, high fever and somnolence, almost invariably leading to death on the fifth or
sixth day. As all disease incidents have been associated with swimming and similar activities In
natural water, finished water supplies would not appear to be a significant route of transmission for
this agent.
Acanthamoeba . This organism, the other cause of primary amoebic meningoencephalitis, is a
widely distributed free-living amoeba found in both fresh water and sewage effluents. It measures
lO-25pm. All species except A. polestinensis are at least mildly pathogenic. This organism gains
entry into the host through abrasions, ulcers, or as a secondary invader during other infections and
not through any specific site of entry. Although Acanthamoeba has been detected in tapwater,
finished drinking water is not a likely route of transmission for this agent (Committee on the
Challenges of Modern Society).
H.Iminth.
The five different genera of helminths that are most likely to be found In U.S. drinking water
are listed in Table I. Although helminth Infections are still prevalent In the US. population,
especially in the Southeast, the occurrence of disease due to these agents In the U.S. has been
extremely low during the last few decades, Because these organisms are excreted with human
and/or animal waste, a potential for disease transmission by water does exist. Eggs of most
parasitic heiminths are large (>50 Pm) and are effectively removed by conventional water
treatment processes such as sedimentation, coagulation and filtration. However, the larvae of
hookworms and threadworms are motile and can migrate through sand filters. This larval stage is
also highly resistant to chlorine, The lack of helminth Infection from water supplies in the U.S. is
most likely attributable to generally adequate sewage and water treatment practices as well as the
absence of wild animal reservoirs. Despite the obvious potential for drinking water contamination
by helmlnths, these organisms are not considered pathogens of sufficient significance in U.S.
drinking wafer to suggest major changes In current water supply practices or regulations.
Toxins of Microbial Origin
Exotoxtns. It has been known for some time that at least three species of freshwater
cyanobacteria, namely Anoboena flos-aquae, Microcystis aeruginosa and Aphanizomenon flos-
-------
L 2
aguae , produce exotoxins (Carmichael, l98la). Only certain strains of these organisms produce
toxins and the non-toxic strains predominate in natural waters. These exotoxins belong to two
general classes: peptides and alkaloids. However, of about 12 different toxins produced by these
species, only one (an alkaloid) has been identified, synthesized and toxicologically characterized
(Carmichael, 1981 b).
When water-blooms of these toxic species occur in surface waters, the cells and toxins con
become sufficiently concentrated to cause illness or death in mammals, birds or fish which ingest a
sufficient dose. Such conditions occur when there is both a bloom and a mechanism for
accumulation of the cells, such as a prevailing directional wind that causes the cells to concentrate
along one shoreline. Animal losses from such exposure hove been reported, including major losses
of cattle, sheep, hogs, birds and fishes.
The acute oral toxicity of these exotoxins to humans has not been established, but there is
increasing evidence that human exposure to these toxins in recreational and possibly drinking
waters has caused contact irritations and possibly gastroenteritis (Carmichael, 1981a). Lack of
knowledge about cyonobacteria exotoxin occurrence in water, removal by treatment processes and
the potential consequences of exposure from drinking water makes these agents a continuing
concern in the management of drinking water supplies.
EndotoxIrts. Lipopolysoccharides (LPS) are the constituents of the outer cell wall of
heterotrophic gram-negative bacteria and cyanobacteria. In both groups, LPS is composed of
repeating otigosacchoride units, a bas I core polysciccharide and a hydrophobic lipid. This
lipopolysaccharide moiety causes endotoxemia and a biphasic pyrogenic fever in mammals. Other
symptoms include leucopenia, leucocytosis and shock. Numerous cases of endotoxic shock have
occurred in debilitated and immunosuppressed patients. Hindman, et al. (1975) found that an
epidemic of pyrogenic reactions among kidney patients occurred at the some time as on algal
bloom in the raw water. High concentrations of LPS accompanying high densities of Schizothrix
calcicola may have been responsible for a massive outbreak of waterborne gastroenteritis in
Sewickley, Pennsylvania (Lippy and Erb, 1976).
Because of the difficulties of measuring endotoxins in clinical specimens, the physiologic
effects and disease syndromes are only now being investigated. The relationship between endotoxiri
and immune function is complex. However, it has become clear that endotoxin complicates
-------
43
treatment and in fact can increase mortality in infected patients. Although ingested endotoxin can
easily pass through the G.l. tract, there is little evidence f or absorption through the intestinal walls
into the portal blood supply. Thus, it is unlikely that endofoxin ingestion in drinking water poses a
health risk (Jakubowski and Ericksen, 1980).
With the advent of Limulus amebocyte lysate assay, measurement of levels of endotoxin down
to picogram quantities became feasible (Yin, et al., 1972). Typically, endotoxins occur in low
concentrations and are ubiquitous in water. Levels of endotoxin measured in various water supplies
are generally on the order of 10 nanograms per millilter (Jakubowski and Ericksen, 1980). Freshly
distilled water has also been shown to be contaminated with endofoxin, indicating the relative heat
stability of these compounds at 100°C. LPS is also produced by certain cyanobacteria (blue-green
algae) including S. caicicola and Anabaena flos-aguae .
There is little evidence to date to suggest that ingestion of low levels of LPS in drinking
water has any major effect on the normal population. At present, there is no simple method
available to detect endotoxins in water, except the Limulus amebocyte lysate assay. Although the
effectiveness of conventional and advanced water treatment processes for endotoxin removal still
remains to be evaluated, there is no evidence that the ambient levels of endotoxin found in most
water supplies pose a significant health risk.
CONCLUSIONS FOR SPECIFIC AREAS
Bacteria
Salmonella and Shtgella . The continued presence of both Salmonella and ShigeIla in the
population, as well as animal reservoirs in the case of Salmonella , requires continuing control of
water quality to protect the consumer from being exposed to these disease agents via the water
route. In most instances, coliform monitoring is sufficient to indicate the potential occurrence of
these two pathogens. However, there have been instances where these pathogens were isolated
from raw or finished waters containing acceptable levels of coliforms. Coliform suppression by
high numbers of plate count organisms might explain such occurrences, since >500 plate count
bacteria/mi have been shown to inhibit coliform bacteria (Geldreich, 1972). Therefore, the
combined use of the coliform standard and a plate count standard in finished water would likely
furnish adequate prediction of these pathogens.
-------
L 4
Yersinia enterocoUtiCa and Campylobacter jejuni . Superficlaily, it would appear that these
two groups of bacteria are sufficiently similar to the Salmonella and Shigella to apply the some
analyses and recommendations to all of them. However, there are several reasons why this is not
possible. The limited information available on Y. enterocolitica and C. jejuni suggests that these
organisms may be more prevalent and persistent in natural waters than the Salmonella and Shigetia ,
although definitive information is locking on this point. However, Y. enterocolitica is more
resistant than E. j to some forms of chlorine disinfection, and it has been detected in finished
waters meeting coliform bacteria standards. In addition, gastroenteritis outbreaks due to C. jejuni
have occurred in water supplies that hove appeared to meet coliform standards.
In areas where Y. enterocolitica and C. jejuni are endemic in animal populations and
therefore likely to be present in source waters, additional efforts must be taken to assure their
absence from the finished water, besides meeting coliform bacteria and other drinking water
standards. Specifically, it is recommended that source water contamination by animals harboring
these organisms be controlled. This may be achieved by finding an alternative raw water source
that is not affected by such animals or by protecting the existing source water from extensive
fecal contamination by these animals. In addition, it Is recommended That treatment of surface
waters consist of chemical coagulatIon, filtration and disinfection with free chlorine, Such
treatment is likely to be effectIve In reducing V. enterocolitica and C. j uni to acceptable levels,
Legionefla . The presence of Leglonella In a variety of natural waters and its Isolation from
water plumbing fIxtures in consumer environments, especially hospitals, suggests that water
distribution systems may transport Legionello to plumbing fixtures that serve as a growth habitat
and then release high concentrations of the organisms into the water. At present, insufficient
information is available regarding Legion removal by conventional water treatment systems,
growth and persistence in natural waters, and ability to colonize distribution systems and plumbing
fixtures to make specific suitable recommendations. Furthermore, the pathogenicity and human
Infectious dose of environmental Isolates in drinking water need to be elucidated before the role of
drinking water in Leqionella transmission can be fully evaluated.
Opportun(stic Pathogerts. This large and varied group occurs in both source and finished
waters. Properly operated conventional water treatment processes appear to be capable of
-------
45
producing finished waters containing relatively low concentrations of these organisms. However,
they are frequently isolated from distribution systems having a chlorine residual. Increasing
numbers of these organisms in finished waters are indicative of water quality deterioration in
distribution systems, because they can colonize and grow on pipes and other surfaces. The standard
plate count organisms are representative of this group and therefore provide an excellent measure
of their occurrence. Furthermore, an estimation of the extent of distribution deterioration as well
as incidental contamination can be gained through standard plate counts of finished water samples
at the treatment plant and distribution system samples. For the aforementioned r asons, a
standard plate counts, done concurrently with coliform analyses at representative locations, are
recommended for water distribution systems.
Enteric Viruses
Because enteric viruses have been detected in bacteriologically acceptable finished drinking
waters that were either extensively treated or had free chlorine residuals, both protection of
source water quality and extensive treatment appear to be necessary. Mounting pollution burdens
to both ground and surface waters, and the occurrence of waterborne disease outbreaks when
obvious source water contamination has occurred, indicate the need for regulations to control,
evaluate and monitor source water quality. Of the largest waterborne disease outbreaks during the
period 1971-77, 67% were due to source water contamination where treatment was either
inadequate or absent altogether. It is recommended that these regulations include a continuing
program of surveillance based on sanitary surveys, microbial indicator standards for source water,
and treatment requirements based on source water quality.
The importance of source water quality is demonstrated by recent findings on the incidence
of viruses relative to indicator organisms in contaminated surface waters. In general, the
percentage of virus-positive samples increased with increasing levels of indicator bacteria (Morris
and Waite, 1981). The frequency of virus-positive samples increased substantially when total
coliforms or E. coil levels were >500/100 ml. These findings suggest that, ideally, total or fecal
coliform levels in source waters should not exceed 500/IOU ml in order to minimize the virus
burden in water treatment plants. Because coliform levels in source waters often exceed this
level, It would appear that microbial standards for source water indexed to the extent of treatment
are necessary.
-------
Li6
Protozocr s
Giordia . In areas where wild animal reservoirs for Giardia are likely to exist, such as New
England, the Pacific Northwest and the Rocky Mountain states, surface water supplies are
especially susceptible to Giardia contamination. Compared to coliform bacteria, Giordia cysts are
relatively resistant to natural self-purification processes and to such conventional water treatment
processes as pressure filtration and chlorination. Therefore, Giordia cysts may be present in
finished water supplies that meet coliform bacteria standards, especially if disinfection is the only
treatment. The low infective dose of this organism further contributes to the potential for
waterborne disease outbreaks.
Several alternative approaches to controlling Giardia contamination of drinking water are
recommended. Conducting sanitary surveys of surface waters for evidence of animal reservoirs in
states having documented waterborne outbreaks within the last several years is advised. Surface
water supplies having both filtration of appropriate design for cyst removal and disinfection as
minimum level of treatment could be exempted from such surveys. If sanitary surveys reveal
animals that might Impact source water, several actions are recommended. First would be
examinations of animals or animal excreta for Giárdia . If animals are Glardia-positive, selection of
on alternative source water, preferably groundwater, that Is not impacted by animal reservoirs is
recommended as a long-term solution. Alternatively, source water could be monitored for fecal
coliforms. When samples are fecal coliform-positive, and therefore indicative of the presence of
Glardia , a boil water order could be Issued until fecal coliform levels decline. Ideally, water
supplies with Giardla-positlve animals In the watershed should ultimately upgrade treatment to
Include filtration of appropriate design for effective cyst removal.
OVERALL CONCLUSIONS
Source Water Quality d Treatment Requirements
Generally increasing wostewater pollution burdens to both ground and surface waters and the
continued occurrence of woterborne disease outbreaks where there has been obvious source water
contamination Indicate the need for regulations to evaluate, control and monitor source water
quality. For example, In the period 1971-77, 67% of the largest waterborne disease outbreaks were
-------
1 7
due to source water contamination where treatment was either inadequate or nonexistent. It Is
recommended that regulations be developed for continuing programs of surveillance that are based
upon sanitary surveys, microbial standards for source waters, and freatment requirements based
upon source water quality.
Analysis of information on waterborne pathogens and fecal indicator bacteria such as total
and fecal coliforms and E. con demonstrates a number of different relationships between Indicators
and various pathogens. For some enteric pathogens such as Salmonella and Shigello , the presence
of the pathogen in raw source waters and in finished water supplies Is generally predictable using
coliform bacteria. This relationship seems to hold because these pothogens are Invariably
associated with fecal contamination, they are present of much lower concentrations than are
coliforms, and the survival of these pathogens in the aquatic environment and their reduction by
conventional water treatment processes are similar to that of coliforms.
For other enteric pathogens, their presence in raw source waters or finished water supplies
may not be adequately predicted by coliform bacteria. Although Yersinla enterocolitica,
Campylobacter jejuni , enteric viruses and Giardia lomblia are associated with human fecal
contamination, some of these pathogens also have animal reservoirs. Moreover, their ability to
surviVe In natural waters or to resist removal and destruction by water treatment processes may be
substantially greater than that of coliform bacteria. Therefore, traditional fecal indicator bacteria
levels In source waters and finlst ed waters are quantitatively less reliable and predictive for these
enteric pathogens. It should be noted, however, that for these pathogens, as exemplified by the
enteric viruses, there is a general relationship of increasing pathogen frequency and perhaps
concentration with increasing levels of indicator bacteria.
There are also disease agents that are not associated with fecal contamination and may occur
“naturally” In aquatic environments, such as the opportunistic bacterial pathogens, nontubercular
mycobocterla, 2 jlo pneumophlla and blue-green algal (àyanobacterlai) toxins. For these
agents, coliform bacteria ore clearly Inadequate Indicator organisms, and therefore, other
approaches to protect and evqluate water quality are needed for them.
Specific regulations for sanitary surveys of source and finished waters are recommended.
Such surveys of source waters will identify: (I) potential fecal and industrial waste sources that
could degrade raw water qualIty, (2) potential animal reservoirs of pothogens, (3) and other
-------
48
conditions that could lead to source water contamination, such as thermal enrichment or algal
blooms causing Legionella proliferation or the presence of cyanobacterial toxins.
Microbial standards for source waters are also recommended. These standards would include
maximum microbial indicator levels above which the source water is considered unusable for
drinking water supply, regardless of treatment, and a series of maximum allowable indicator levels
that designate the level or degree of treatment required for drinking water production. Several
types of microbial indicators may be needed for source water quality standards. Total and/or fecal
coliforms (or E. coil) are certainly useful and historically-proven reliable indicators of fecal
contamination that ore clearly relevant to at least some enteric pathogens of fecal origin.
Standard plate count organisms may also be useful indicators of source water quality because
fecally-contaminated source waters are also likely to have high levels of heterotrophic bacteria.
Such bacteria might overburden treatment processes and lead to excessive SPC levels in finished
waters. Futhermore, high levels of SPC bacteria can suppress coliform detection in both raw and
finished water, thereby leading to underestimation of fecal contamination of the water.
Turbidity may also be a useful indicator of source water quality. Excessive turbidity can
overburden the efficiency of such treatment processes as coagulation, filtration and disinfection
and lead to excessive turbidity levels in finished waters, in addition, at least some microbial
pathogens such as enteric viruses and at least some opportunistic bacteria tend to be preferentially
associated with turbidity particles. If source waters have high turbidity levels with associated
pathogens, such turbidity-associated pathogens are more likely to appear in the finished water.
Particle-associated opportunistic pathogens may proliferate In and colonize finished water distri-
bution systems.
The concept of indicator standards for raw source water quality is not new, and has a sound
technical basis that dates back to the 1930’s. In 1915, the U.S. Public Health Service began a series
of studies on the microbial quality of raw water that could be treated to meet the coliform
standard for drinking water used by interstate carriers, These classical studies, under the
direction of Streeter and colleagues, examined the performance of 31 municipal treatment plants
and on experimental treatment plant on the Ohio River in Cincinnati (Streeter, 1935). A major
finding of these studies was that consistent production of finished water meeting a coliform
standard of 1/100 ml by filtration and chlorination required that the raw water coliform density
-------
‘49
not exceed about 5,000/100 ml. For raw waters treated only by chlorination, a coliform density not
exceeding 50/100 ml was required to consistently meet the drinking water standard for coliforms.
The need for raw source water standards indexed to treatment requirements seems to be even
greater today in light of increased and newly-recognized sources of wastewater contamination for
ground and surface waters, and the recognition of waterborne pathogen. that ore predicted either
less reliably or not at all by traditional bacterial indicators.
In addition to indexing the degree of treatment to source water quality, minimum treatment
requirements for all surface and groundwaters ore recommended. The minimum treatment
recommended for surface waters is filtration and disinfection. Disinfection alone cannot be relied
on to consistently produce drinking water of acceptable microbial quality from raw surface water.
There are few, if any, “pristine” surface waters remaining in this country. Even clear mountain
stream waters previously considered “pristine” have been associated with disease outbreaks due to
Glordia and possibly other agents having animal reservoirs. Virtually all surface waters contain
sufficient turbidity to interfere with disinfection processes and allow turbidity-associated bacteria
and perhaps other microbial agents to enter finished waters.
The minimum treatment recommended for groundwater is disinfection. The physical,
chemical and microbiological quality of many groundwater. meets current drinking water standards
without treatment. However, increased opportunities for groundwater contamination and the
continued occurrence of disease outbreaks due to untreated or inadequately treated groundwater
indicate the need for disinfection as a minimum treatment requirement.
Another aspect of drinking water production that requires regulation is evaluation and
monitoring of treatment process effectiveness. Here, too, the routine monitoring of water quality
by the use of indicators during treatment may be useful. For example, concerns about the
efficiency of enteric virus reduction in such processes as coagulation, filtration and disinfection
could possibly be dealt with by routine bacteriophage monitoring of infiuent and effluent water far
each of these unit processes. Similarly, concerns about Gigidia cyst reduction by coagulation and
filtration could be dealt with by turbidity monitoring of influent and effluent from these unit
processes. Standard plate counts could also be used to monitor unit process efficiency for overall
bacteria reductions. Drarncitic decreases in removal of such Indicators in a unit process would alert
plant personnel within a relatively short time period of process deficiencies requiring corrective
-------
50
action. Turbidity measurements can be made on an in-line, continuous basis or within a matter of
minutes, while bacteriophage and SPC analyses produce results within 8 and 24 hours, respectively.
Monitoring the efficiency and reliability of plant performance by measuring these indicators may
be more effective for controlling drinking water quality than just the routine coliform monitoring
of finished water.
Because of continued deterioration of raw source waters, the continued occurrence of
waterborne outbreaks, the recognition of waterborne disease agents that are either not predictable
or only marginally predictable by coliforms, and the resistance of these agents to natural self-
purification and water treatment processes, additional regulations for source water protection,
minimum treatment requirements for ground and surface water, and requirements for monitoring
treatment process effectiveness apear to be needed.
Finished Water ond Distribution Systems
During the period 197 1-77, 27% of all large waterborne disease outbreaks were due to
distribution system contamination. Therefore, it appears that additional effort must be made to
protect finished water quality once it leaves treatment plants and before it reaches the consumer.
Furthermore, there is a growing body of physical, chemical and microbial evidence that finished
water quality often degrades during transport through distribution systems. For example, a number
of studies have shown that SPC and turbidity levels of drinking water increase in distribution
systems. The reasons for such increases remain uncertain but could be due to the following:
(I) bacterial regrowth and colonization, (2) deterioration of distribution pipes, and (3) the entrance
of external contaminants by such means as broken or leaky pipes, pipe maintenance and repair
operations and cross connections.
RECOMMENDATIONS FOR REGULATIONS AND GUIDELINES
Regulations for Source Water Quality aid Treatment
On-site sanitary &uveys. Regulations for on-site sanitary surveys are recommended. For
source waters, such surveys will identify: (I) potential fecal and industrial waste sources that
could degrade raw water quality, (2) potential animal reservoirs of pathogens, and (3) other
-------
51
potential conditions that could lead to pathogen contamination, such as thermal enrichment causing
Legionella proliferation or toxic blue-green algal blooms,
Treatment. Because of continued deterioration of raw source waters, the continued
occurrence of waterborne outbreaks, the recognition of waterborne pathogens that are either not
predictable or only marginally predictable by coliforms, and the resistance of these pathogens to
natural self-purification and water treatment processes, regulations for treatment are
recommended.
Minimum treatment for all groundwater supplies should be chlorination with an effective free
residual or an adequate alternative disinfectant. In the absence of excessive turbidity this
treatment should effectively reduce such problematic pathogens as Yersinla and Compylobacter to
acceptably low levels. Minimum treatment for surface water supplies should be filtration and
chlorination with an effective free residual or an adequate alternative disinfectant. In the absence
of excessive turbidity this treatment should effectively reduce such pathogens as Yersinia,
Campylobacter and Giardia .
Re uIatkn. for FlnLdied Water ond Distribution SysLesns
Because of water quality problems related to distribution systems, Including the presence of
opportunistic pathogens, regulations are needed for monitoring finished water quality. Two
specifIc Issues must be addressed: (I) monitoring programs that are designed to accurately assess
finished water quality throughout the distribution system, and (2) better indicators of finished
water quality during distribution. The former requires the establishment of rational monitoring
programs in which the location of sampling points and the sampling frequency ore capable of
achieving this objective. The latter requires the use of additional water quality indicators other
than the coliform test. Two additional indicators for this purpose should be an appropriate
standard plate count and turbidity. Therefore, routine monitoring of total coliforms, standard
plate count organisms and turbidity in finished drinking water Is recommended at both the
treatment plant and at various locations In the distribution system that are truly representative of
the range of conditions the distribution system. This monitoring should be done within the
framework of a vigorous surveillance program for evaluation and protection of distribution system
integrity.
-------
52
1aflcns for Specific Disease Agents
Legionella prs.eumophila and Other Leglonella-Like Bacteria. When potable water is used f or
evaporative type cooling systems, adequate measures such as chlorine disinfection should be taken
to prevent the proliferation of LegioneHa to excessive levels. Excessive levels of Legionella in the
piping systems of new public use facilities such as hospitals, hotels and workplaces should be
controlled by disinfection according to the recommended AWWA procedure for new pipes or an
equally effective procedure. Cooling tower exhausts in such facilities should be directed In such a
way that they are not readily drawn into fresh air intakes. Institutions with populations at high
risk, such as hospitals, should be particularly made aware of the potential presence of Legionella in
finished water. For example, if they are located at the end of a water distribution system line or
have a self-contained water supply system, disinfection within the facility and monitoring of hot
water sources for la by appropriate methods should be considered.
Utilities having water supplies with large holding reservoirs or towers should be Informed that
jIa can potentially grow In these storage systems and that they should take necessary
actions to prevent its proliferation.
In public places where high population densities are likely to be near such “amplifiers” of
LeQionella as recirculating fountains, humidifiers (or dehumidlflers), cooling towers and air wash
systems, guidance should be given concerning the possibility that such sources may oct as
Leglonello habitats from which aerosol dissemination con occur.
Opportunistic Pat hogens. Hospitals, nursing homes and other Institutions with high risk
populations should be Informed of microbial water quality changes in distribution systems that may
lead to conditions of increased exposure to opportunistic pathogens within these facilities.
Water distribution systems in newly constructed hospital facilities should be disinfected
according to AWWA procedures for new distribution lines or an equally effective procedure. The
adequacy of disinfection should be verified by standard plate count measurement.
Enteric Viruses. There exists epidemlological evidence of waterborne outbreaks due to such
enter Ic viruses as hepatitis A, Norwalk-type agents and rotaviruses. Furthermore, enter Ic viruses
that are capable of causing human Infection have been Isolated from drinking waters that met
-------
53
standards for coliforms and turbidity and contained free chlorine residuals. Because there.appears
to be an association between virus occurrence and fecal coliform levels in raw water, it Is
recommended that where source waters contain >500 fecal collforms/ 100 ml, monitoring be
conducted to establish treatment process efficiency. Evaluation of treatment process efficiency
shall be based upon the measurement of an appropriate viral indicator In the following samples:
intake water, water just prior to final disinfection and finished water at the treatment plant.
Giardia lamblia . Although filtration and chlorination with a free residual are capable of
extensive Giardia reductions, it is advised that systems also employ appropriate continuous in-line
turbidity monitoring for filtration performance. This additional parameter is needed to detect
minor turbidity increases which may be indicative of cyst breakthrough. Furthermore, It is
suggested that filter bock wash water be wasted and that water also be wasted during filter
ripening periods to prevent cyst contamination of the filtered water.
Toxic Freshwater Cyanobacteria (blue-green algae). MunicipalitIes using surface waters in
which toxigenic algae are present In amounts exceeding IO 3 cells/mI by direct microscopic analysis
should monitor for the presence of a toxic strain via mouse bioassay and/or free toxin. The water
should be monitored both before and after treatment.
Use of finished waters containing toxins at an LD of <1,000 mg/kg body weight (i.p. by
mouse bioassay) should be discontinued until toxin bioassay indicates an LD 50 of>2,000 mg/kg.
Pon.l Conclusions for Other Agents
Acceptability of Existing Regulations. The panel concluded that adequate protection of
drinking water from public health risks due to Salmonella, Shigella , and enteropathogenic E. coil Is
provided by current regulations and technology.
Agents Lacking Public Health Significance. The following agents were judged by the panel to
be of inconsequential public health significance In drinking waters endotoxIns and antibiotic
resistance factors in bacteria.
Agents for Which There is No Current Evidence of Public Health Risk. The panel concluded
that *here was no evidence for current public health risks via US. drInking water for the following
-------
54
agents: Enfamoeba histolytico, Nae leria , Acanthamoeba sp. , and the helminths. However, the
potential public health significance of these agents in drinking water should not be ignored, and
should be periodically reevaluated. This is particularly important for Entamoeba histolytica ,
because of the documented historical evidence for its transmission via drinking water.
Agents for Which There is a Suggestion of Public Health Risk. Nontubercular mycobacteria
In drinking water do not appear to pose a risk to healthy consumers. However, they may pose a risk
to certain segments of the population that have existing health problems, especially if the
organisms colonize special environments. The panel established no regulatory recommendations for
these organisms, except those for opportunistic pothogens in general.
RESEARCH RECOMMENDATIONS FOR DISEASE AGENTS
tntr tfrmn
In order to promulgate regulations which adequately protect the public from exposure to
waterborne disease agents, additional Information Is needed concerning these agents in source
waters, water treatment processes and distribution systems. In addition there are research needs
specific to many of these disease agents.
Reaeorth Reconineidotions for Disease Agents in Source Water,
Treatment Processes cmd Distribution Systems
Source Waters. Quantitative Information is needed on the occurrence and survival of a
number of disease agents In source (natural) waters under a variety of environmental conditions.
These agents ore: the bacteria Yersinla enterocolitica, Campylobacter jejuni, Leglonella
pneumophlta , other Legionella-Ilke bacteria (LLB), and nontubercular mycobacterla; the enteric
viruses, especially rotavlruses, Norwalk-type agents and hepatitis A; and the protozoan Glardla
lomblia .
Water Treatment. information on the effectiveness of water treatment processes for many
disease agents is lacking. This Is particularly evident for the more recently recognized disease
agents, such as Yersinla enterocolitica, Campytobacter Jejuni, Leglonella pneumophlla and other
LLB, nontubercular mycobocteria, enteric viruses and algal toxins.
-------
55
s for Specific Dise e Agents
Salmonella and Shigella . Better quantification methods are needed for these two genera of
bacteria In natural and finished waters. The development of these methods should emphasize ease,
reliability and sensitivity.
Yersinta enterocoUtica . (I) Quantitative methods for defecting Yersinla In wafer are needed.
(2) QuantItative Information is needed on the survival of Yersinla in natural waters under a variety
of environmental conditions and on the reduction of Yersinla by such water treatment processes as
chemical coagulation, flocculation, filtration, and disinfection by free chlorine and alternative
dlsinfecfanfs. (3) There Is a need for Information on the pathogenicity and virulence of
environmentally isolated Yersinic organisms compared to clinical Isolates. Such studies should
include Information on the relationship between pathogenlclty/vfrulence and the presence of
endotoxin, heat stable enterotoxin and invasiveness factor.
Campylobacter Jejuni . (1) Improved methods are needed for Isolating and quantifying
Campylobacter in natural and finished waters. (2) Complete identification of the animal reservoirs
for this agent and their Impact on source water is needed. (3) The relationships between traditional
fecal Indicator bacterIa and Campylobacter in water should be quantitatively determined with
respect to survival In source waters and their reduction by conventional water treatment processes.
In order to determine the role of water In the transmission of Campylobacter better
information Ii needed on the endemic occurrence of this agent In the population and on its role in
gastrointestinal Illness.
Leglanella pneumoi h1la and Other Leqianefla-Like Bacteria.
Ecoloqy . Information Is needed on the physical, chemical and biological factors that
Influence Lealonella occurrence, survival and proliferation (growth) in both source and finished
waters, especially under storage conditions. Information Is also needed on gj j a occurrence In
groundwaters, Including deep aquifers. Research Is needed to determine the relationships, If any,
between and traditional Indicator bacteria Including fecal Indicators and standard plate
count organisms.
-------
56
Source Waters and Stored Finished Waters . Although quantitative information is available on
Legionel Ia occurrence in natural water in the eastern U.S., data are more limited for other parts of
the country. Furthermore, because most of these data have been based only on direct fluorescent
antibody analysis (DFA), there is a need to corroborate these findings using additional methodology
based upon viability counts.
Treatment . The reduction of Legionella in conventional water treatment processes must be
determined. Particular attention must be given to Legionel Pa destruction by free chlorine and a
variety of alternative disinfectants.
Distribution . Research information is especially needed on Leglonella occurrence and
proliferation in distribution and storage systems, with particular emphasis on the role of dead ends
and other areas of distribution systems having long residence times.
Transport Studies . Research is especially needed on the aerosolization of Legionella from
amplifier systems supplied with finished water, such as water fountains, shower heads, and
evaporative type cooling systems In environments where high risk exposure of the population may
occur.
Pathogenicity . Information is needed on the pathogenicity, virulence and human Infective
doses of Legionella from natural and finished waters. Such studies should Include the Identification
of pathogenicity/virulence factors.
Health Effects . The potential for Infection by ingestion and Inhalation of contaminated
drinking water needs to be determined. The relative significance of drinking water in the
transmission of Legionella should be evaluated.
Nontubercular Mycobacteria. Research is needed to determine the occurrence, persistence
and proliferation of mycobacterla in source waters, through conventional treatment processes and
within distribution systems. Such studies require the development of quantitative methods for the
recovery of these organisms. Furthermore, epiderfliologlca l studies on the waterborne transmission
of these disease agents should be conducted and surveillance should include skin testing of the
exposed population. Research is also needed on the pathogenicity and virulence of waterborne
nontubercular mycobocterla.
-------
57
Enteric Viruses.
Water Quality indicator(s) for E teric Viruses . Monitoring the effectiveness of specific water
treatment processes for enteric virus removal is recommended when source water quality is poor
( 5OO fecal coliforms/lOO ml). Research is needed to develop a suitable indicator parameter or
virus marker for enteric viruses that could be readily monitored in the field. The suitability of
collphages, turbidity, and possibly otI er potential enteric virus indicators should be determined.
Hepatitis A and Gastroenteritis \ Viruses . Hepatitis A and gastroenteritis due to Norwalk-type
agents and perhaps rotaviruses are the most frequent waterborne viral diseases in the U.S.
Research on these agents is needed’ In the following areas: (I) improved methods for their
detection, cultivation and assay; (2) studies on their removal and destruction in water and
wastewciter treatment processes and their survival in natural waters; (3) field studies on their
occurrence in raw waters used for water supply.
Enteric Virus Field Surveys . Although extensive field studies have been conducted on enterlc
virus occurrence In raw source waters and finished water supplies, new field studies are needed.
The reasons for this are: (I) the availability of improved virus detection methods and (2) recent
evidence from limited field studies showing that enteric virus removals in a full-scale, conven-
tional water treatment plant are not nearly as great as the predicted removals based on previously
reported laboratory and pilot plant studies. Such field studies should be done with the best
available methods and they should be done on a variety of different water supplies that represent a
wIde range of possible conditions for source water quality, type of treatment system and finished
water quality.
Enteric Virus Dose-Response Studies . Some dose-response information Is available on the
likelihood of infection (or illness) from ingesting different amounts of enteric viruses. However,
more studies using additional virus types, lower virus doses and dIfferent human host populations
are needed.
Epidemiological Studies . Epldemlologlcal studies are needed to determine if drinking water is
a significant source of virus infection In the population and If this route of transmission Is
Important for virus entry into and persistence within communities.
-------
58
Enteric Virus Detection Methods . Although methods to detect and quantify enteric viruses in
water have been greatly improved in recent years, it is likely that only a small and variable
proportion of the total viruses actually present in a field sample are being detected. Further
improvements in virus detection methods are needed in order to make them more efficient and
reliable. The best available methods should also be systematically evaluated in carefully designed
collaborative studies and quality assurance tests. In addition, methods are needed to monitor virus
recovery efficiency when processing field samples, possibly by using marker viruses or virus analogs
in field samples.
Opportunistic Patl-eogerts. The relationship between waterborne opportunistic organisms and
the occurrence of infection and disease in special risk populations such as infants, the elderly and
immunocompromised individuals should be delineated.
Glardia lamblia . (I) In order to be able to accurately determine the effectiveness of
disinfection, improved methods for indicating cyst viability are needed. (2) Research is needed on
the effectiveness of alternative disinfectants (to free chlorine) for their ability to destroy Giardia
cysts. (3) Improved methods are needed to detect and quantify Giardia cysts in natural and finished
waters. (4) Stool surveys of watershed animals in different regions of the country are needed in
order to determine the potential for contamination of water supplies by other than human wastes.
(5) Epidemiological studies are needed to determine if drinking water is a significant source of
Giardia infection in the population and if this route of transmission is important for Giardla entry
into and persistence within communities.
Cyanobacterla Toxins. (I) Research is needed to develop chemical methods for detecting
and quantifying algal toxins in water. (2) Quantitative information is needed on the effects of
water treatment processes on algal toxins. (3) There is a need for better epidemiological
surveillance information on the occurrence of waterborne algal toxin outbreaks. Such information
could be gathered through cooperative joint efforts by the EPA, CDC and appropriate state
agencies.
-------
59
LITERATURE CITED
Akin, E.W. 198 I. A review of infective dose data for enteroviruses and other enteric
microorganisms in human subjects. In: Microbial Health Considerations of Soil Disposal of
Domestic Wastewaters , National Center for Ground Water Research and U.S. EPA, Norman,
Oklahoma.
Allen, M.J. and EE. Geldreich, Jr. 1978. Evaluating the microbial quality of potable waters, pp.
3-10. In: Evaluation of the Microbiology Standards for Drinking Water , C.W. Hendricks (ed.),
U.S. El 5 7 , EPA---570/9-78-OOC, Washington, D.C.
Armstrong, J.L., D.S. Shigeno, J.J. Calomiris and R.J. Seidler. 1981. Antibiotic resistant bacteria
in drinking water. AppI. Environ. Microbiol., 42: 277-283.
Arozarena, M.M. 1977. Removal of Giardia muris cysts by granular media filtration. Masters
thesis, University of Cincinnati.
Bogdasaryan, G.A. and R.M. Abieva. 197 I. Survival of enteroviruses and adenoviruses in water.
Hyg. Sanit. 36:333-338.
Baine, W.B., C.A.Herron, K. Bridson, W.H. Barker, Jr., S. Lindell, G.F. Mall,son, J.G. Wells, W.T.
Martin, M.R. Kosuri, F. Carr and E. Voeker. 1975. Waterborne shigellosis at a public school.
Amer. J. Epid. 10 1:323—332.
Batik, 0., G.F. Craun and W.O. Pipes. 1983. Routine coliform monitoring and waterborne disease
outbreaks. J. Environ. Health 45(5):227-230.
Berendt, R.F. 1981. Influence of blue-green algae (Cyanobacteria) on survival of Legionella
pneumophila in aerosols. Infect. lmmun. 32:690.
Black, R.E., G.F. Craun and P.A. Blake. 1978. Epidemiology of common source outbreaks of
shigellosis in the United States. 196 1-1975, Amer. J. Epidemiology l08(l):47-52.
Black, H.J., E.J. Holt, K. Kitson, M.H. Maloney and D. Phillipps. 1979. Contaminated hospital
water supplies. Brit. Med. Jour. l(6177):l564-l565.
Blackman, J.A., F.W. Chandler, WB. Cherry, A.C. England, J.C Feeley, M.D. Hicklin, R.M.
McKinney and H.W. Wilkinson. 1979. Legionellosis. Amer. J Pathology 103(3):429-405.
Blaser, M.J. and L.B. Petler. 1981. Campylobacter enteritis. New England J. Med. 305:1444-1452.
Blaser, M.J., H.L. Hardesty, B. Powers, and W.L. Wang. 1980. Survival of Campylobacter fetus
subsp. jejuni in biological milieus. J. Clin. Microbiol. I l(4):309-3l3.
Blaser, M.J., J.G. Wells and R.A. Feldman. (In press) Epidemiology of endemic and epidemic
Campylobacter infections in the United States. In: Proceedings of the International Workshop
on Campylobacter Infections , A. Pearson (ed.)., Ptblic Health Laboratory Service, London.
Bopp, C.A,, J.W. Sumner, G.K. Morris and J.G. Wells. 1981. Isolation of Legionella spp. from
environmental water samples by low-pH treatment and use of a selective medium. J. Clin.
Microbiol. l3(4):7l4-7l9.
Boring, J.R., W.T. Martin and L.M. Elliott. 1971. Isolation of Salmonella typhimurium from
municipal water, Riverside, California, l965.Amer. J. Epid. 93:49-54.
Brady, F.J., M.F. Jones and W.L. Newton. 1943. Effect of chlorination of water on the viability of
cysts of En& iiioebo histolytico . War Med. 3:409-419
-------
60
Brickler, S.K. and B. Tunnicliff. 1980. Water Quality Analyses of the Colorado River Corridor of
Grand Canyon. University of Arizona, College of Agriculture, Paper 350, Tucson.
Burman, NP. (1962) Bacteriological control of slow sand filtration. Effluent and Water Treatment
J. 2:674:677.
Butterfield, CT. 1948. BacterIcidal properfies of chloramines and free chlorIne in water. Public
Health Reports 63:934-940.
Caprioll, T., A.J. Drapeau and S. Kasatiya. 1978. Yersinia enterocolitica : serotypes and blotypes
isolated from humans and the environment in Quebec, Canada. J. CUn. Microbiol. 8:7-Il.
Carmichael, W.W. (ed.) 1981. The Water Environment: Algal Toxins and Health . Plenum Press,
New York, 491 pp.
Carmichael, W.W. 1981 a. Freshwater blue-green algae (cyanobacteria), pp. 1-13. In: The Water
Environment: Algal Toxins and Health , W.W. Carmichael (ed.), Plenum Press, New York.
Carson, L.A., M.S. Favero, W.W. Bond and N.J. Peterson. 1973. Morphological, biochemical, and
growth charocteristics of Pseudomonas cepacia from distilled water. Appi. Microbiol.
25(3):476-483.
Center for Disease Control. 1972. Bolantidiasls-Truk District Micronesia. Morb. and Mort.
Weekly Rep. 2 1:59.
Center for Disease Control. 1980. Waterborne Disease Outbreaks In the United States -1978 .
Morb. and Mort. Weekly Rep. 29(46):58 1.
Centers for Disease Control. 1981c. Annual Summary 1980. Morb. and Mort. Weekly Rep.
29(54): 128.
Centers for Disease Control. 198 lb. Morb. and Mort. Weekly Rep. 30( 17) : 198- 199.
Centers for Disease Control. 198io. Shigellosis-United States-I 980. Morb. and Mort. Weekly Rep.
30(36):462-463.
Chang, S.L. 1943 Studies on Entamoebci histolyticci , II. Observations concerning encystatlon,
maturation and excystatlon ol E. histotytica and on the longevity of culture-Induced cysts in
various fluids and at different temperatures. J. Infect. Dis. 72:232-241.
Chang, S.L. 1958. The use of active iodine as a water disinfectant. J. Am. Pharm. Assn. Sci. Ed.
48:417-423.
Chcing, S.L. 1978. Pathogenesls of pathogenic Noegieria amoeba. in: Folio Parasitologica (Fourth
International Congress of Parasitology, August 19-26, 1978), Warsaw, Poland.
Chang, S.L. and G.M. Fair. 1941. Viability and destruction of the cysts of Entomoebo histolytica .
Jour. AWWA. 33:1705-1715.
Chcing, S.L., R.E. Stevenson, A.R. Bryant, R.L. Woodward and P.W. Kabler. 1958. Removal of
coxsackie and bacterial viruses in water of known chemical content by flocculation with
aluminum sulfate or ferric chloride under various testing conditions. Amer. J. Pub. Hith.
48:5 1-61.
Colwell, R.R., B. Austin and L. Wan. 978. Public health considerations of the microbiology of
potable water. pp. 65-75 in: Evaluation of the Microbiology Standards for Drinking Water ,
C.W. Hendricks (ed.), EPA 70I9-78-O0C, U.S. EPA, Washington, DC.
Committee on the ChoUenges of Modern Society, Pilot Study on Drinking Water Supply Problems:
Microbiology, North Atlantic Treaty Organization. In press.
-------
61
Cordes, L.G. et al. 1981. Isolation of Legionella pneumophila from hospital shower heads. Ann.
Intern, Med. 94:195-197.
Croun, G.F. 1978. Impact of the coliform standard on the transmission of disease. In: Evaluation
of the Microbiology Standards for Drinking Water , C.W. Hendricks, (ed.), U.S. EPA 570-9178-
OOC, U.S. EPA, Washington, D.C.
Craun, G.F. 1979a. Waterborne giardiasis in the United States: a review. Amer. J. Public Health
69:817-819.
Craun, G.F. l979b. Waterborne disease - a status report emphasizing outbreaks in groundwater
systems. Ground Water 17:183-196.
Craun, G.F. 198 Ia. Outbreaks of waterborne disease in the United States: 1971-1978.• J. Am.
Water Works Assoc. 6:360-369.
Craun, G.F. 1981 b. Personal communication.
Craun, G.F. and R.A. Gunn. 1979. Outbreaks of waterborne disease In the United States: 1975-
1976. J. Am. Water Works Assoc. 67:422-428.
Cronier, S., P.V. Scarpino and M.L. Zink. 1978. Chlorine dioxide destruction of viruses and
bacteria in water, p. 65 1-657. In: Water Chlorination; Environmental Impact and Health
Effects, Vol. 2, R.L. Jolly, H. rchey, and D.I-1. Hamilton, Jr. (eds.), Ann Arbor Science,
Ann Arbor, Michigan.
Cummins, B.B. and H.D. Nash. 1978. Microbiological implications of alternative treatment. j :
Proceedings of the American Water Works Association Water Quality Technology Conference,
paper 2B-l, pp. 1-14.
D’Aousf, J.Y. and H. Pivnick. 1976. Small infectious doses of salmonella. The Lancet. I(2):866.
Dailloux, M. et al. 1980. Study of the factors affecting presence of atypical mycobacteria in
water ora swimming pool. Rev. Epidemiol., Sante Publique (Fr.) 28:299-
Dailloux, M. et al. I980a. Study on the relationship between isolation of mycobacterla and
classicarml&obiology and chemical indicators of water quality in swimming pools. Zentrobl.
Bakteriol. Parasitenkd. lnfektionski. Hyg. Abt. (Ger.) 171:473-486
Deetz, T.R., E.M. Smith, S.M. Goyal, C.P. Gerba, J.J. Vollet, L. Tsai and H.L. Dupont. Rota-and
entero-viruses in potable and environmental water in Mexico, unpublished manuscript.
DiLuzio, N.R. and T.J. Friedmann. 1973. Bacterial endotoxins in the environment. Nature
244:49-51.
Dobberkau, H.J., R. Walter and St. Rudiger. 1981. Methods for virus concentration from water,
pp. 235-238. In: Viruses and Wastewater Treatment, M. Goddard and M. ButI r (eds.) Pergamon
Press, New Y k.
Dominowska, C. and R. Malottke. 1971. Survival of Yersinia in water samples originating from
various sources. Bull. Inst. Mar. Med Gdonsk. 22:173-182.
Dorn, J.M. 1974. A comparative study of disinfection of viruses and bacteria by mono hloramine.
M.S. Thesis, University of Cincinnati, Ohio.
duMoulin, G.C. 1979. Airway colonization by Flovobacterlum In an Intensive care unit. J. Clin.
Microblol. 10:155-160.
duMoulin, G.C. and K.D. Stottmeier. 1978. Use of cetylpyrldinium chloride in the decontamination
of water for culture of mycobacterla. Appi. and Environ. Mlcroblol. 36:771-773.
-------
62
duMoulin, G.C., I. Sherman and K.D. Stottmeier. 1981. Mycobocferium intrucellulare : an
emerging pathogen. Abstr. 81st Ann. Meeting, Amer. Soc. Microbiol., WashFngton, D.C.
Dutka, B.J. and J.B. Bell, 1973. Isolation of Salmonella from moderately polluted water. Journ.
Water Poll. Control Fed., 45(2):3 16-324.
Dutka, B.J. and K.K. Kwan. 1980. Bacterial die-off and stream transport studies. Water Res.
14:909—95.
Echeverria, P., P.J. Mejia and C. Duangmoni. 1981. Effect of antibiotics on the prevalence of
enterotoxigenic Escherichia coil in two populations in the Philippines. Antimicrobial Agents
and Chemotherapy I 9:290-29
Eden, K.V., M.L. Rosenberg, M. Stoopter, B.T. Wood, A.K. Highsmight, P. Skaliy, J.G. Wells and
J.C. Feeley. I 977. Woterborne gastrointestinal illness at a ski-resort -isolation of Yersinia
enterocolitica from drinking water. Public Health Reports 92:245-250.
Edwards, E. and A.P. Wyn-Jones. 1981. Incidence of human enteroviruses in water and wastewater
systems associated with the River Wear, p. 109-i 12. In: Viruses and Wastewater Treatment ,
M. Goddard and M. Butler (eds.), Pergamon Press, New u’ork.
Eichhorn, J.H. 977. Contamination of medical gas and water pipelines in a new hospital building,
Anesthesiology 46(1 4):286-289
Englebrecht, R.S., M.J. Weber, C.A. Schmidt, and D.L. Sulter. 1978. Virus sensitivity to chlorine
disinfection of water supplies. Final Report, EPA 600/2-78-123, Washington, D.C.
Esposito, M.P. 1974. The inactivation of viruses in water by dichlorarnine. M.S. Thesis, Univ.
Cincinnait, Ohio.
Evinson, L. 1977. Disinfection of water with ozone: Comparative studies with enteroviruses,
phage and bacteria. Unpublished paper presented at the Third World Ozone Congress, Paris,
1977. Sponsored by the International Ozone Institute (now Association), Cleveland, Ohio.
Fair, G.M., S.L. Chang and J.C. Morris. 1945. Final Report on Disinfection of Water and Related
Substances to the Committee on Medical Research of the Office of Scientific Research arid
Development, Washington, D.C.
Farrah, S.R., S.M. Gopel, C.P. Gerba, R.H. Conklin and E.M. Smith. 1978. Comparison between
adsorption of poliovirus and rotavirus by aluminum hydroxide and activated sludge floes.
AppI. Environ. Microbiol. 35:360-363.
Favero, M.S., L.A. Carson, W.W. Bond and N.J. Peterson. 1971. Pseudomonas aeruginosa : growth
in distilled water from hospitals. Science 173:836-838.
Favero, M.S., N.J. Petersen, L.A. Carson, W.W. Bond and S.I-l. Hindmari. 1975. Gram-negatIve
water bacteria in hemo-dialysis systems. Health Laboratory Science 12(4):321-334.
Feachem, R.G. (977. Infectious disease related to water and excreta disposal facilities. Amblo.
6:55-58.
Feochem, R.G., D.J. Bradley, I-I. Garelick and D.D. Mara. 1978. Health Aspects of Excreto and
Wastewater Management , The World Bank, Washington, D.C. ( also to be u6flshed In modified
form as: Son itation and Disease: Health As ts of Excreta and Wastewater Management ,
World Bank Studies in Water Supply and Sanitation, No. 3, Johns Hopkins University Press ,
Baltimore, in press.)
Fliermcns, C.B., W.B. Cherry, U-i. Orrison anc L. Thacker. 1979. Isolation of Legionella
pneumophlla from non-epidemic-related aquatic habitats. AppI. Environ. Microblol.
37(6): 1239-1242.
Fliermans, C.B., W.B. Cherry, L.H. Orrlson, S.J. Smith, D.L. Tison and D.H. Pope. 1981.
Ecological distribution of Leglonelici pneumophila . AppI. Environ. Microblol. 4l(I):9- 16.
-------
63
Foy, ftM., C.V. Broome, D.S. 1-byes, I. Allan, M.K. Cooney and R. Tube. 1979. Legionnaires’
Disease in a prepaid medical-care group in Seattle, 1963-1975. Loncet i:767-770.
Geidreich, E.E. 1972. Waterborne pothogens, p. 207-24 I. In: Water Pollution Microbiology , R.
Mitchel (ed.), John Wiley and Sons, New York.
Geldreich, E.E., MJ. Allen and R.H. Taylor. 1978. Interferences to coliform detection in potable
water supplies, p. 13-19. In: Evaluation of the Microbiology Standards for Drinking Water ,
CW. Hendricks (ed.), U.S. ‘A 570/9-78-OOC, Washington, D.C.
Gilcreas, F.W. and S.M. Kelly. 1955. Relation of coliform organism test to enteric virus pollution.
J. Amer. Water Works Assoc. 47:683-693.
George, K.L. 1980. Epidemiology of infection by nonfubercular mycobacterba. ii Growth and
Survival in Natural Waters. Amer. Rev. Resp. Dis. 122:89-94
Goldfield, M. 1976. Epidemlological Indicators for transmission of viruses by water. In: Viruses In
Water , G. Berg et al. (eds.), Amer. Pubi Helth Assoc., Washington, D.C., pp.70-85
Good, R.C. 1980. IsolatIon of nontuberculous mycobocterla In the United States 1979. J. lnf. Dis.
142:779-783.
Gorden R.W. and C.B. Fllermans. 1978. Survival and viability of Eacherlchla coil In a thermally
altered reservoir. Water Research 12:343-352.
Grabow, W.O.K., O.W. Prozesky and L.S. SmIth. 1974. Drug resistant coliforms call for review of
water quality standards. Wat. Res. 8:1-9.
Greenberg, HB, J.R. Valdesuso, A.R. Kallca, R.G. Wyat, V.J. McAu liffe, A.Z. Kapikion and R.M.
Chanock. 1981. Proteins of Norwalk virus. J. VIrol. 37:994-999.
Grinstein, 5,, J.L. Melnick and C. WallIs. 1970. VIrus Isolations from sewage arid from a stream
receiving effluents of sewage treatment plants. Bull. Wld. Hith. Org. 42:291-296.
Guy, M.D., J.D. Mclver and M.J. Lewis. 1977. The removal of virus by a pilot treatment plant.
Water Research I 1:421-428.
Harvey, S., J.R. Greenwood, M.J. Plckett and R.A. Mah. 1976. Recovery of Yerslnla enter-
ocoUtica from streams and lakes of California. Appi. Environ. Mlcrobioi. 32(3):352-354.
Hazen, T.L., C.B. Fliermons, R.P. Hirsch and G.W. Esch. 1978. Prevalence and distribution of
Aeromonas hydrophilo in the United States. Appi. Environ. Mlcrobiol. 36:732-738.
HeJkaI, 1. W,, B. Keswlck, R. L. LaBelIe, C. P. Gerba, Y. Sanchez, C. Dreesman, B. l-lafk ln and J.
L. Melnlck 1982, VIruses in a community water supply associated with an outbreak of
gastroenteritis and infectious hepatitis. J. Amer. Water Works Assoc. 74:318-321.
Herman, L. S. and C. K. Himmelsbach 965. DetectIon and control of hospital sources of
Flovobacteria . Hospitals 39:72-76.
Herrmann, JE., K. 0. Kostenboder arid D. 0. Cliver. $974. Persistence of enteroviruses In lake
water. Appi. Mlcrobiol. 28:895-896.
I-bindman, S.H., M.S. Favero and A. Peterson. 1975. Pyrogenlc reactions during hemodlalysls
caused by extramural endotoxins. The Lancet 18:732-737.
-------
64
Hinshaw, V., J. Punch, M.J. Allison and H.P. Dalton. 969. Frequency of R factor-mediated
multiple drug resistance in Klebsiella and Aerobacter . Appi. Microbiol. 17:214-218.
Hoehn, R.C., C.W. Randall, F.A. Bull, Jr., and P.T.B. Shaffer. 977. Trihalomethanes and viruses
in water supply. J. Envr. Engr. Div. ASCE, 103:803-814.
Hoekstra, A.C. 1978 Desinfectie van Haags drinkwater en zuiverende werking van de lang zame
zondfilter, periode 1966-1978. Report of DWL’s Gravenhage, 5 pages, Drinkwaterleiding
van’s-Gravenhage, ‘s-Gravenhage, The Nether lands
Hoff, J.C. 1978. The relationship of turbidity to disinfection of potable water, p. 103-Ill. In:
Evaluation of the Microbiology Standards for Drinking Water, C.W. Hendricks (ed.), U.S. EP ,
EPA 570/9- 78-OOC, Washington, D.C.
Holmes, l.H. 1979. Viral gastroenteritis. Prog. Med. Virol. 25:1-36.
Huekelekian, H. and H.B. Schulhoff. 1935. Studies on the survival of B. typhosus in sureface waters
and sewage. NJ. Agricultural Experiment Station Bulletin 589, New Brunswick, NJ.
Hugh, R. and G.L. Gilardi. 1974. Pseudomonas , p. 250-269. In: Manual of Clinical Microbiology ,
E.H. Lennetfe, E.H. Spaulding, J.P. Truant (eds.), 2n ed., Am. Soc. for Microbiology,
Washington, D.C.
Hurst, C.J. and C.P. Gerba. 1980. Stability of simian rotavirus in fresh and estuarine water. AppI.
Environ Microbiol. 39:1-5.
Jakubowski, W. Detection of Giardia cysts in drinking water: state-of-the-art. In: Giardia and
Giardiasis, Plenum Press O jress).
Jakubowski, W. and T.H. Erickson. 1979. Methods for detection of Giardia cysts in water supplies.
pp. 193-210 In: Waterborne Transmission of Giardiasis , EPA-600/9-79-a U.S. EPA
Cincinnati, OFC
Jakubowski, W. and T.l-1. Ericksen 1980, Health significance of bacterial endotoxins in drinking
water, In: Proceedings of the American Water Works Association Water Quality Technology
Conference VIII, Amer. Water Works Assoc., Denver, Colorado
Jakubowski, W. and J.C. Hoff (eds.) 1979. Waterborne Transmission of Giardiasis ,
EPA-600/9-79-OOl, 306 p., Cincinnati, OH.
Jarroll, E.L., A.K. Bingham, and E.A. Meyer. 980. Giardiasis cysts destruction: Effectiveness of
6 small quantity water disinfection methods. Am. J. Trop. Med. Hyg. 29:8- I I.
Jarrol, E.L., A.K. Blngham and E.A. Meyer. 1981. Effect of chiroine on Giardla cysfviobility. AppI.
Environ. Microbiol. 43:483-487.
Jordon, E.O., H.L. Russell and F.R. Zeit. 1904. The longevity of the typhoid bacillus in water. Jour.
Infec. Dis. 1:641-689.
Keet, E. 1974. Yersinia enterocolitica septicemia. New York State J. Med. 74:2226.
Kelch, W.J. and J.S. Lee. 1978. Antibiotic resistance patterns of gram-negative bacteria isolated
from envrionmental sources. Appi. Environ. Microbiol. 36:450-456
Kokina, A.G., N.A. Lukashevich and N.V. Novitskaya. 1977. Effect of chemical and microbiological
quality of groundwater on the survival of enteroviruses and enterobacteria. Sanit .
0(5): I I I - 113 (Russian).
Knill, M., W. G. Suckling and A. D. Pearson. 1978. Environmental Isolation of heat-tolerant
Campylobacter in the Southampton area. Lancet 11:1002-1003.
-------
65
Kruse, C. W., V. Hsu, A.C. Griffith and R. Stringer. 1970. Halogen action of bacteria viruses and
protozoa. p. 113-137. In: Noti. Speciality Conf. on Disinfection, Amer. Soc. Civil Eng., New
York
Lahiri, M.N., P.C. Dos and K.S. Malik. 1939. The viability of Vibrio chioerae in natural waters.
Indian Medical Gazette 74:742-744.
Lassen, J. 1972. Yersinia enterocolitica in drinking water. Scand. J. Infect. Dis. 4:125-127.
Lennetfe, E.H. 1976 Problems posed to man by viruses in municipal wastes. In: Virus Aspects of
plying Municipal Waste to Land , L.B. Baldwin !i’ . (eds.) Univ. o(Florldo, Gainesville,
Lippy, E.C. 1978. Tracing a giardiasis outbreak at Berlin, New Hampshire. J. Am. Water Works
Assoc. 70:512-520.
Lippy, E.C. 1981. Waterborne disease: occurrence is on the upswing. J. Am. Water Works Assoc.,
73:57-62.
Lippy, E.C. and J. Erb. 1976. Gastrointestinal illness at Sewickley, PA. J. Am. Water Works
Assoc. 68:606-610.
Logsdon, G.S., J.M. Symons, R.L. Hoye and M.M. Arozareno. I 98 1. Alternative filtration methods
for removal of Giordia cysts and cyst models. J. Am. Water Works Assoc. 73:111-118.
Mahnel, H. K. Ott and M. Herlyn. 1977. Stabilltat von neun Viruarten unt rschiedlichen Genera in
Trink-und Oberflachenwasser, zbl. Bakt Nyg, I Abt, Or 1g. B, 164:64-84.
MaIlman, W.L. and D. Kahler. 1948. Bocterial removal by lime-alum and alum-clay flocs. J. Am.
Water Works Assoc. 40:6(5-624
Marzouk, Y., S.M. Goyal and C,P. Gerba. 1977. Prevalence of enteroviruses in ground water of
Israel. Ground Water 17:487-491.
Mausner, J.S. and A.K. Bahn, 1974. Sources of data on community health, p. 160-183. In:
EpidemloloQy , W.B. Saunders Co., PhIladelphia, PA. —
McFetars, G.A., C.K. Bissonnette, J.J. Jezeski, C.A. Tomson and D.C. Stuart. 1974. Comparative
survival of Indicator bacteria and enterle pathogens in well water. AppI. Microbiol.
27*823-829.
Melnlck, J.L. 976. Enterovlruses, pp. 63-201, j s Viral Infections of Humans: Epidemloloay and
Control , A.S. Evans (ed.) Plenum Press, New York
Melnick, J.L., GA. Bogdasorjon, C. Berg, S.E. Hul, E. Lund, B.C. Roo, K. Ren, 1-1.1. Shoval, J.S.
Siade, I-I. Yunte. 1979. Human Viruses In Water, Wastewater, and Soil , World Health
Organization, Technical Report Series No. 639, Geneva.
Mendls, N.M.P., P.U. De Ia Motte, P.D.P. Gunnatilloka and W. Nagoratnam 1976 Protracted
infection with Salmonella bareflty In a maternity hospital, J. Trap. Med. 1-tyg. 79 : 142- 150.
Mentzing, L.O. 1981. Waterborne outbreaks of Campylobocter enteritis in central Sweden. The
Lancet 2:352-354.
Meyer, P.W. and S.J. Lerman. 1980 . Rise and fall of Shigella antibiotic resistance Antimicrob.
Agents and Chem. 17:101-102.
Mohodjer, S. and S. Mebrabian. 1975. StudIes on the survival of j [ a flexnerl In river and tap
water. Arch. Roum, Exp. Microblol. 34:307-3(2.
Morris, R. and W. M. Waite. 981. Indicators, Viruses and Water Quality p. 257-276. In: Viruses
and Wastewafer Treatment , M. Godard and M. Butler (eds.), Pergamon Press, New ?&k.
-------
66
Neefe, J.R., J. Stokes, Jr., J.B. Bat>’ and J.G. Reinhold. 1945. Disinfection of water containing
causative agent of infectious (epidemic) hepatitis. J.A.M.A. 128:1076-1080.
Neefe, J.R. and J. Stokes, Jr. 1945b. An epidemic of infectious hepatitis apparently due to a
waterborne agent. J.A.M.A. I 28:1063-1075.
Neefe, J.R., J.B. Bat > ’, J.G. Reinhold et at. 1947. Inactivation of the virus of infectious hepatitis in
drinking wafer. Am. J. Pub. HltF 37:365-372.
Neogy, K.N. 1965. Viability of V. cholerae and V. El br in food and water. Bull. Calcutta School
Trop. Med. 13: 10- Il.
Newton, W.L. and M.F. Jones. 1949. Effect of ozone in water on cysts of Enfamoeba hisfolytica .
Amer. J. of Trop. Med. 29:669-681.
O’Brien, R.T. and J.S. Newman. 1977. Inactivation of polioviruses and coxsackieviruses in surface
water. Appi. Environ. Microbiol. 33:334-340.
Olson, B.H. and L. Hanami. 1980. Seasonal variation of bacterial populations in water distribution
systems. Proceedings of the American Water Works Association 8th Water Quality
Technology Conference, pp. 137-151, AWWA, Denver, Colorado.
Orrison, LH.,, W.B. Cherry, C.B. Fliermans, S.B. Dees, L.K. McDougal and D.J. Doss. 1981.
Characteristics of environmental isolates of Legionella pneumophila . Appl. Environ.
Microbiol. 42(I): 109-l 15.
Osterhoim, M.T. et al. 1981. An outbreak of foodborne Giardiasis. New EngI. J. Med. 304:24-28
Payment, P. 1981. Isolation of viruses from drinking water at the Pont-Vicu water treatment
plant. Canada J. Microbial. 27(4):417-420.
Peterson, D.A., T.R. Hur ley, J.C.Hoff and L.G. Wolfe. 1982. Effect of chlorine treatment on
infectivity of hepatitis A virus. Abstracts Ann. Mtg., Am. Soc. Microbiol., Washington, D.C.
Poynter, S.F.B and J.S. Slade. 1977. The removal of viruses by slow sand filtratkk Prog. Water
Tech. 9:75-88.
Powell, B.L., Jr and JE. Steadham. 1981. Improved technique for isolation of Mycobacteriurn
kansasii from water. J. Clin. Microbiol. 13:969.
Pozhor, V.N. 1973. Viability of enteroviruses in the alkaline and acid effluents of the Gaysk ore
dressing combine. Gig. Sanit. 38:101-102.
Prior, BW.H. 1976 Yersinia enterocolitica in surface waters. Masters Thesis, University of
Washington, Seattle.
Reilly, J. K. and J. S. Kippin. 1981. Interrelationship of bacterial counts with other finished water
quality parameters within distribution systems. Project Summary, EPA 600/52-81-035, u.s.
EPA Municipal Enviornmental Research Laboratory, Cincinnati, Ohio.
Reinhard, D.J. 1980. Isolation, identification and quantitation of Gram-Negative nonfermentative
bacilli from aqueous and aquatic sources. Dev. md. Microbiol. 20:705-721.
Rice, E.W. and J.C. Hoff. 1981. Inactivation of Giardia tamblia cysts by U.V. irradiation. Appl.
Environ. Microbiol. 42:546-547.
Rice, E.W., J.C. Hoff and F.W. Schaefer II. 1982. Inactivation of Giardia cysts by chlorine. Appl.
Environ. Microbiol. 43:250-251.
Robeck, G.G., N.A. Clarke, and K.A. Dostat. 1962. Effectiveness of water treatment processes in
virus removal. J.Am. Water Works Assoc. 54:1275-1292.
-------
67
Rollond, D. and J.C. Block. 1980. Simultaneous concentration of Salmonella and enterovirus from
surface water by using micro-fiber glass filters. AppI. Environ. Microbiol. 39:659-661.
Rosenberg, M.L., J.P. Koplan, l.K. Wachsmith, J.G. Wells, E.J. Gangarosa, R.C. Guerrant and D.A.
Sack. 1977. Epidemic diarrhea at Crater lake from enterotoxigenic E. coli. Ann. Internal
Medicine 86(6):7 14-7 18.
Saari, T.N. and G.P. Jansen. 1979. Waterborne Yersinia enterocolitica in the Midwest United
States, p. 185- 196. In: Contributions to Microbiology and Immunology, Vol. 5, P.B. Cater, L.
Lafleur and S. TomaTeds.), S. Karger, Basel.
Scarpino, P.V., M. Lucas, D.R. Dahling, G. Berg and S.L. Chang. 1974. Effectiveness of
hypochlorous acid and hypochloride ion in the destruction of viruses in bacteria, p. 359-368.
In: Chemistry of Water Supply Treatment and Distribution , A.J. Rubin (ed.), Ann Arbor
rcience Publishers, Ann Arbor, MI.
Seidler, R.J. et al. 1980. Isolation, enumeration and characterization of Aeromonas from polluted
waters encountered in diving operations. AppI. Environ. Microbiol. 39:1010-10 18.
Seligman, R. and R. Reitler. 1965. Enteropathogens in water with low Esch . coli:titer. Jour. Am.
Wafer Works Assoc. 1572- 1574.
Show, P.K., R.E. Brodsky, D.O. Lyman, B.T. Wood, C.P. Hibler, G.R. Healy, K.I.E. MacLeod, W.
Stahl and M.G. Schultz. 1977. A community-wide outbreak of giardiasis with evidence of
transmission by municipal water supply. Ann. Internal Med. 87:426-432.
Shayegani, M., I. DeForge, D.M. McGlynn and 1. Root. 1981. Characteristics of Yersinia
enterocolitica and related species isolated from human, animal, and environmental sources.
J. Clin. Microbiol. 14:
Skiers, D.L., P.V. Scarpino, M. Lucas, G. Berg and S.L. Chang. 1973. Destruction of viruses and
bacteria in water by monochloramine. Abstracts of the Annual Meeting of Am. Soc. for
Microbiol. E27.
Siegl, G. 1982. Structure and biology of hepatitis A virus. pp. 13-20 In: Viral Hepatitis, 1981
international Symposium , W. Szmuness, H.J. Alter and J. E. Maynard (eds.) The Franklin
Institute Press, Philadelphia.
Simkova A. and Z. Wallnerova. 1973. Survival of small amounts of coxsackievirus Ak in Danube
River water under laboratory conditions. Acfa Virol. 17:505-506
Skally, P. and H.V. McEachern. 1979. Survival of the Legionnaires’ disease bacterium in water.
Ann. intern. Med. 90:662-663.
Skaliy, P., l.A. Thompson, G.W. Gorman, G.K. Morris, H.V. McEachern and D.C. Mackel. 1980.
Laboratory studies of disinfectants against Legionello pneumophila . AppI. Environ. Microbial.
40:697-700.
Smith, H.W. 1970. Incidence in river water of Escherichia coIl containing R-factor. Nature.
288:1286-1289.
Smith, E.M. and C.P. Gerba. 1982. Development of a method for detection of human rotavirus in
water and sewage. Appl. Environ. Microbiol. 43:1440-1450.
Sobsey, M.D., C.H. Dean, M.E. Knuckles and R.A. Wagner. 1980. Interactions and survival of enteric
viruses In soil materials. AppI. Environ. Microbiol. 40:92-101.
Spector, B.K., and J.R. Baylis, 0. Gullans. 1934. Effectiveness of filtration In removing from
water and chlorine in killing the causative organism of amoebic dysentery. Pub. Health
Reports 49:786-800.
-------
68
Stagg, C.H., C. WaDis and C.H. Ward. 1977. Inactivation of clay-associated bacteriophage MS-2
by chlorine. Appi. Environ. Microbiol. 33:385-391.
Stottmeier, KD., L. lose and C. Jones. 1973. Clinical and bacteriologic evaluations of pulmonary
mycobacterioses in the greater Boston area. Am. Rev. Resp. Dis. I 08:1 I 27-I 130.
Streeter, H.W. 1927. Studies of the effect of wafer purification processes. Pub. Health Bull 172.
Streeter, H.W. 1930. The efficiency of water purification processes. Pub. Health Bull 193.
Streeter, H.W. 1935. Limiting standards of bacteria quality for sources of purified water supplies.
J. Am. Water Works Assoc., 27(9): 11 10-I I 19.
Stringer, R. and C.W. Kruse. 1971 • Amoebic cycticidal properties of halogens In water.
p. 319-33 I. In: Natal. Speciality Conf. in Disinfection, Am. Soc. Civil Eng., New York.
Stringer, R.P., W.N. Cramer and C.W. Kruse. 1975. Comparison of bromine, chlorine, and iodine
as disinfectants for amoebic cysts. pp. 193-209. In: Disinfection: Water and Wastewater,
J.D. Johnson (ed.), Ann Arbor Science Publishers, inE Ann Arbor, MI.
Sykora, J.L., G. Keleti, R. Roche, D.R. Volk, G.P. Kay, R.A. Burges and M.A. Shapiro. 1980.
Endotoxins, algae and Limulus Amoebocyte lysate test in drinking water. Water Research
14 :829-839.
Taylor, D.N., K.J. McDermltt, J.R. Little, and M.J. Blaser. 1981. Woterborne Campylobocter
eriteritis. In: Proc. Ann. Meeting, Amer. Soc. Microblol., Am. Soc. MicrobloL, Washington ,
D.C.
lichen, W. and R.L. Vogt. 1978. Waterborne Campylobacter gastroenteritis - Vermont. Morb.
and Mort. Weekly Report 27:207.
Tison, D.L., D.H. Pope, W.B. Cherry, and C.B. Fllermons. 1980. Growth of Legloneila pneumophlla
in association wiTh blue-green algae (cyanobocteric). App). Environ. Micrabiol.
39(2) :456-459.
Tobin, J. O’H., R.A. Swann, and CL.R. Bartlett. 1981. Isolation of Leglonella pneumophila from
water systems: Methods and preliminary results. British Medical Journal 282s5l5-5 17 .
Turnberg, W.L. 1980. Impact of Renton treatment plant effluent upon the Gre.n-Duwamlsh River,
Masters Thesis, University of Washington, Seattle.
Wang, W.L.L., B.W. Powers, M.J. Blaser and N.W. Leuchtfeld, 1982. Laboratory studies of
disinfectants against Compylobacter leluni . Proc. Ann. Mtg. Am. Soc. Microblol., Am. Soc.
Microbiol, Washington, D.C.
Worrlner, T. R. 1967. Inaktivering cv polio virus med kiordloxide. Vatten. 67:284-290.
Welbel, S.R., F.R. Dixon R.B. Weidner and L.J. McCabe. 1964. Waterborne dIsease outbreaks. J.
Am. Water Works Assoc. 56(8);947-958.
Wetdenkopf, S.J. 1958. InactivatIon of typel poflomyefltis virus with chlorine. VIrology 5:56-67
Welssmon, J.B., G.F. Craun, D.N. Lawrence, R.A. Pollard, M.S. Saslow and EJ. Gangarosa. 1976.
An epidemic of gostroenterItis traced to a contaminated public water supply. Am. J.
Epidemiol. 103:391-398.
Welke, C., V. Friedrich and K. Mal. 1973. InactivatIon time determination far echoviruses 6, Ii,
30 and 33 In water. Z. Gesamte Hyg. 5:353-355.
Wetzier, T.F. 1981 Personal communication.
-------
69
Wefzler, T.F. and J. Allard. 1977. Yersinia enterocolitica from trapped animals in Washington State.
Paper presented at: International Conf. Dis. in Nature Communicable to Man. Yellow Bay,
Montana
Wetzler, T.F., J.T. Rea, G. Vuen and W. Turnberg. l978a. Yersinia enterocolitic in wafers and
wastewoters. 106th Ann. Meeting, Am. Public Health Assoc., Los Angeles, CA.
Wetzler, T.F., J. Allard and T. J. Quan. I 978b. Yersinia enterocolitica from trapped animals in
Washington State, Paper presented at: International Conf. Dis. of Nature Communicable to
Man. Vancouver, B.C., Canada.
Wet zler, T.F., J.R. Rea, G.J. Ma and M. Glass. 1979. Non-associaticn of Yersinia with traditional
coliform indicators. In: Proc. Ann. Mtg. of the Am. Water Works Assoc., Am. Water Works
Assoc., Denver, CO
Wheater, D.W.F., D.D. Mara, L. Jawad and J. Oragui. 1980. Pseudomonas aeruginosa and
Escherichia coIl in sewage and freshwater. Water Res. 14:713-721.
Wilson, R., C.E. Haley, D. Relman, E. Lippy, G.F. Craun, J.G. Morris and J.M. Hughes. 1981.
Waterborne outbreaks related to contaminated groundwater reported to the Center for
Disease Control, 197 1-1979, In: Microbial Health Consideral Ions of Soil Disposal of Domestic
Wastewaters . U.S. EPA, Was1 ington D.C .
Wolinsky, E. 979. Nontuberculous mycobacteria and associated diseases. Am. Rev. Resp. Dis.
119:107-159.
Yin, E.T. et at. 1972. Picogram-sensitive assay for endotoxin: gelation of Limulus polyphemus
bIoocV fl lysate induced by purified lipopoly saccharides and lipid A from gram-negative
bacteria, Biochem. Biophys. Acta 261:284-289.
Yolken, R.H., F. Lawrence, F. Leister, H.E. Takiff and S.E. Strauss. 1982. Gastroenteritis
associated with enteric type adenovirus in hospitalized infants. J. Pediatrics 101:21-26.
-------
MEASUREMENT OF MICROBIAL QUALITY
V. P. Olivieri
Panel Members: Jeanne Baliesfero, Victor J. Cabelli, Charles Chamberlin, Dean Cliver, Al DuFour,
Walter Ginsburg, George Heal ’, Anita Highsmlth, Raulston Read, Don Reasoner, Marjorie Shovlin,
Richard Tobin
ABSTRACT
INTRODUCTION 2
MEASUREMENT OF MICROBIAL QUALITY 3
Criteria for Surrogate Measurements 4
Indicators of Microbial Quality 1
Microbiological Indicators 4
Chemical Indicators 12
Evaluation and Application of Indicators of Microbial Quality 14
Interpretation of Indicators of Microbial QualIty 18
REGULATORY ASPECTS 20
Historical and Global PerspectIve 20
Urban vs. Rural 22
Experience 22
RESEARCH RECOMMENDATIONS 24
SUMMARY AND CONCLUSIONS 25
LITERATURE CITED 26
APPENDIX A Chamberlin 9!. 30
APPENDIX B Minority Report 38
ABSTRACT
The existing collform and turbidity regulations are appropriate and provide substantial but
not absolute safeguard from disease. Minor modifications in procedures for coliform determination
will not alter the utility of Its use. At present, In drinking water, total coliforms are still the best
indicators available. However, the relationship of coliforms to the risk of disease transmission
through contaminated water is not well established. The panel felt strongly that the importance of
the sanitary survey be re-emphasized by specific incorporation Into the primary drinking water
regulations. This will permit better interpretation of the microbial quality of water. The sanitary
survey should be conducted in all water systems and updated annually. Complete surveys should be
conducted at least every 5 years. Sanitary surveys should be conducted before construction in all
new, modified or expanded water systems and prior to use. Since the sanitary survey has been
neglected, the committee recommends that a set of guidelines be developed in the form of a
handbook or similar document that can be used to conduct the survey.
Human pathogens should be absent from finished drinking water. However, it should be noted
that a negative test for the presence of any such agent in a specified volume of finished water
-------
2
cannot prove the absence of the agent from the water from which the sample was taken. On the
basis of what Is now known, any standard requiring testing to demonstrate that a given pathogen Is
not detectable In a specified volume of finished water seems unlikely to add to the protection of
public health so as to justify the added cost of the required sampling and testing. Therefore, the
measurement panel took the conservative position of recommending surrogate determinations.
Three distinct aspects of the use of surrogates to measure microbial quality were recognized.
The surrogates may serve as (I) Indicators of feces (the classical role of indicators); (2) indIcators
of water treatment efficiency; and (3) indIcators of deterioration, degradation and/or
recontamination.
No single measurement adequately fulfills all the requirements for these aspects of indicators
throughout the water system from source to tap. Total coliform, plate count and free chlorine
residual methods were the best indicators for (2) treatment efficiency and (3) degradation,
deterioration and recontamination.
INTRODUCTION
The records and remains of ancient civilizations suggested an appreciation of the need to
protect the quality of wafer to be used for human consumption. However, It was not until 1855
that John Snow (1855) clearly demonstrated that cholera was transmitted by water contaminated
by feces. Shortly thereafter, William Budd (1857) demonstrated a waterborne route of transmission
for typhoid fever. Even as the germ theory of disease was being debated by Pasteur and Liebig, by
1870 It was generally agreed upon that drinking water contaminated by feces was a serious danger
to health. By the end of the 19th century, the microbial etiology of many of the diseases that
followed an anal-oral route of transmission had been established and the health hazards associated
with water contaminated by feces were confirmed beyond doubt. The major thrust, then, was
directed at the recognition of fecal contamination and was reflected in the U.S. Treasury
Department, Bacteriological Standard for Drinking Water in I9l1 .
Early workers relied heavily on the sanitary survey to evaluate the fecal contamination of
water supplies (ChadwIck, 1842 and Shattuck et ., 1917). These surveys included information on
the source of water, the topography and geology of the area, the soil characteristics, and the
sources of possible contamination, As the analytical technology developed, fecal contamination
was evaluated by the laboratory analysis for Inorganic materials and Included measurements of
-------
3
chloride, free and albuminold ammonia, nitrate and nitrite, solids, and oxygen consumption. By the
turn of the century, Information on the presence and levels of microorganisms in water had
accumulated and was beginning to be employed in conjunction with field studies and chemical
analysis to evaluate water supplies. At that time, it was quickly recognized that the bacterial
determinations were much more sensitive than the chemical measurements, and significant effort
was directed toward the detection and enumeration of microorganisms from feces.
An excellent review of the early water microbiology con be found in Elements of Water
Bocterloloqy with Special Reference to Sanitary Water Analysis by Prescott and WInslow (1904). It
should be noted that the above volume was publIshed 10 years before the first regulations were
enacted and should be reviewed to provide some historical perspective to the development of the
initial bacteriological water standards.
MEASUREMENT OF MICROBIAL QUALITY
Early efforts to isolate pathogenic microorganisms from water were unproductive. Prescott
and Winslow (1904) reviewed the isolation of specific pathogens and summarized the reports to
1904:
“On the whole it seems that since a positive result is always open to serious doubt, and
a negative result signifies nothing, the search for the typhoid bacillus itself, however
desirable theoretically, cannot be regarded at present as generally profitable.”
While the methods for the isolation of the spirillum of cholera from water were considered less
difficult than those for the typhoid bacillus, similar conclusions as to the practical utilization of
these methods were drawn. Despite the dramatic improvement In culture media, isolation,
IdentIfication, and enumeration methods for bacteria, viruses and protozoa since 1904, the Isolation
of specific pothogens still “cannot be regarded at present as generally profitable.” The low
numbers of pathogenic microorganisms relative to the natural microbial populations and the wide
variety of pothogeni that may be present are severe limitations to the usefulness of procedures for
the enumeration of pathogeni. Human pathogeris should be absent from finished drinking water.
However, it should be noted that a negative test for the presence of any such agent in a specified
volume of finished water cannot prove the absence of the agent from the water from which the
sample is token. On the basis of what is now known, any standard requiring testing to demonstrate
that a given pathogen Is not detectable in a specified volume of finished water seems unlikely to
odd to protection of the public health, so as to Justify the added cost of the required sampling and
-------
4
testing. As a result, microbiologists suggested a surrogate microbial determination to indicate the
presence of fecal contamination rather than rely on the isolation of pothogens.
Criteria for Surrogate Measurements
The general criteria for the surrogate measurement of microbial quality are listed below.
The primary function of the surrogate measurement was to indicate the occurrence of recent fecal
contamination.
I. The indicator should always be present when the source of the pathogenic micro-
organisms of concern is present and absent in clean uncontaminated water.
2. The indicator should be present in large numbers.
3. The Indicator should respond to natural environmental conditions and to water and
wastewoter treatment processes in a manner similar to the pathogens of interest.
4. The indicator should be easy to isolate, Identify, and enumerate.
While the criteria for indicators of microbial quality appear to be simple and straightforward,
no one microorganism or group of microorganisms, or chemical or biological test adequately
satisfies all of the above criteria.
lndlcatoa of Microbial Quality
Numerous indicators for the evaluation of microbial quality of drinking water have been
proposed. They may be divided into three broad categories: fecal indicators, treatment indicators
and degradation and/or recontamination Indicators. The surrogates Include a vórlety of micro-
biological and chemical determinations.
Microbtologicczl Indicators. The classical bacterial Indicator evolved as an Indicator of human
feces. Von Frltsch (1882) reported Kiebsiella pneumonia and K. rhlnoscleromatls as being
characteristic of human feces and Escherich (1885) reported the Isolation of Bacillus coil
(subsequently designated Escherlchla coil) from feces of a cholera patient in 1885. Since the work
of the early bacteriologists, the microbial inhabitants of the Intestinal tract of man and animals
have received considerable attention. Review of the early recoveries and 1solatIon can be found in
the six editions of Prescott and Winslow, 1904 to 1946. Recent estimates of the microbial flora in
human feces have been reported by Geldreich (1977) and Leclerc (Lec lerc, et c ii., 1977). The most
consistent microorganisms and groups of microorganisms found In human feces were Bacteroldes
-------
5
fragilis (an anaerobic, gram-negative bacilli), fecal coliforms, total coliforms, E. coil, fecal
streptococci, and enterococci.
Col iform Group . The early workers recognized the presence in large numbers In feces and
sewage of a group of bacteria whose most significant member was Bacillus coIl. Although the
name of the indicator was changed several times over the years, the current coliform group is
essentially the same group of microorganisms that has served since the late 19th century as an
indicator of the presence of fecal contamination of water. The coliform group is currently defined
in the 15th edition of Standard Methods for the Examination of Water and Wastewater (1981) as
“aerobic and facultative anaerobic, gram-negative, nonsporeforming, rod-shaped bacteria which
ferment lactose with gas formation within 48 hrs at 35°C”.
Aside from the difficulties associated with determining the presence of groups of micro-
organisms rather than a single species and the not uncommon false positive, false negative, and
atypical reactions, the major deficiencies of the coliform group are listed below;
I. The natural die-away of members of the coliform group differs markedly from the die-
away of nonbacterial pathogens. The enteric viruses in particular and probably cysts of
pathogenic protozoa survive for longer periods of time in the natural aquatic
environment.
2. The response of the coliform group to conventional water treatment processes,
particularly disinfection, also differs from that of the nonbacterial pathogens. The
coliform group appears more sensitive to disinfection than virus and probably the
protozoan cysts. This aspect is of little Importance when free residual chlorination Is
practiced, but becomes a major deficiency when combined chlorine Is depended upon for
disinfection.
3. The altergrowth of some members of the coliform group differs from that of some
bacterial enteric pathogens; viruses and protozoan cysts cannot multiply In the
environment.
4. The suppression, in laboratory testing, of the growth of coliforms by high populations of
other microorganisms influences recovery and enumeration.
-------
6
Fecal Coliform Group . Fecal coliforms, a subgroup of the coliform group, have been
employed to evaluate wastewater plant effluents and surface waters and are believed to be more
indicative of contamination with the feces of warm-blooded animals than are the total coliforms
(Geldreich, et al., 1962). It should be recognized that the fecal coliform determination does not
distinguish between human and animal fecal contamination. The numbers of fecal coliforms
observed In feces, sewage, and various water samples are considerably lower than the numbers of
total coiiforms. Several workers suggest that E. coIl Is a more useful indicator because It Is more
certainly of fecal origin (Barrow, 1977 and Dufour, 1977).
cal Streptococci . The fecal streptococcus group has been used to assess the quality of
streams and lakes. Although present In large numbers In feces, they are considerably less numerous
than the coliform group In human feces and thus are a less sensitive Indicator. In addition, the
problems In evaluating the levels observed by different methods and the limited sanitary
significance of several members of the fecal streptococci (Geldrelch and Kenner, 1969) have
resulted in their limited usefulness In drinking water. MaIlman (1961), however, points out that the
fecal streptococci may be functional in situations where the coliform test is of limited value. The
fecal streptococci may be useful on relatively clean streams and lakes used for recreation. Similar
to the situation with coilforms and fecal coliforms, the enterococci, a subgroup of the fecal
streptococci restricted to the strains of Stre tococous faecalls and foeclum have been suggested
as useful Indicators (Clouseii, sE .. 977). Fecal streptococci have had limited application In
drink!ng water.
Pseudomonas aerualnosa. Pseudomonas aeruglnosa , an opportunistic pathogen noted for its
resistance to antibiotics, its proclivity for infecting Individuals that are In a debilitated state, and
Its association with eye and ear Infections has been proposed as an indicator for swimming pool and
recreational waters. This microorganism can be found in feces and appears to be associated with
human rather than animal feces (Hoadly, 1968). Levels In the aquatic environment sometimes
approach those of the fecal coliforms. However, P. aeruainosa appears to be ubiquitous in the
environment and apparently does multiply under natural conditions (Cabelli, [ ., 1976).
-------
7
Clostridium perfringens. Clostridlum perfrlngens , on anaerobic sporeformer, has been
advocated as an indicator particularly since the spores persist long after coliforms have died out
naturally in the aquatic environment or have been killed by disinfection (Cabelli, 1977 and Hoather,
1952). The spores appear to be too resistant to chlorination to be of use as indicators In drinking
water. The cultural methods are adequate for research and special investigations. In Britain, C.
perfrlngens determinations are frequently employed to provide supplemental information on good
quality waters.
Bifidobacterium . Organisms of the genus Bifidobacterium are anaerobic, nonsporeforming,
nonmotile, gram-positive, pleomorphic rods that have been suggested as fecal indicators (Evison
and James, 1974; and Levin, 1977). These microorganisms appear to satisfy many of the indicator
criteria and may be useful In differentiating between human and lower animal contamination.
However, only a few studies have been conducted to obtain quantlt9tive Information on these
organisms in water, and methods and procedures require development.
Yeasts . Buck (1979) reviewed the usefulness of Candida albicans as an indicator. Despite the
lock of quantitative data on the distribution in natural waters, he suggested this yeast appeared to
be associated with humans and animals. Engelbrecht and Hoos (1977) have observed several
chlorine-resistant yeasts and have proposed these microorganisms as disinfection indicators. Media
and methods for a reliable enumeration procedure require further study.
Bacteriophage . Bacteriophages have been evaluated as Indicators, particularly for enteric
viruses (Grabow, 1968; Kott, 1977; and Scarpino, 1978). The RNA coliphages have been employed
successfully as model viruses In disinfection studies (Cramer ., 1976; Olivlerl, 1974; and Show
and McCamish, 1972) but appeared to have a very low occurrence in human feces. Poppel 1 (1979)
reported that only 2 of 124 human fecal samples were positive for RNA coliphage. The DNA
coliphages (T-phages) occurred more frequently In human feces, were more sensitive than enteric
viruses to free chlorine and appeared to be able to multiply In the aquatic environment. Although
both DNA and RNA coliphoges were readily recovered from sewage In sufficient numbers, their
levels in surface waters were not high enough to permit their use as Indicators without
concentration.
-------
8
Plate Count . The bacterial plate count has long been employed to evaluate the quality of
water. Miquel in 189 I (in Prescott and Winslow, 1904) suggested that the sanitary quality of water
be determined by plate count. Table I shows Miquel’s quality designations expressed as
bacteria/cc. Excessively pure water contained fewer than 10 bacteria/cc while very impure water
contained greater than 100,000 bacteria/cc. Clearly, Miquel’s classification was somewhat
arbitrary.
A major difficulty with the standard plate count is that no one plate count method has been
universally adopted. The bacterial plate count has generally been performed with a relatively rich
complex medium and aerobic Incubation and thus has yielded on estimate of the number of aerobic
heterotrophic bacteria. Most anaerobic and autotrophic mIcroorganisms will not grow and,
therefore, will not be enumerated. The bacterial plate count will be Influenced by the medium
employed and the time and temperature of incubation. A diluted, weaker medium may yield higher
numbers than a richer medium. Victoreen (1977) reported consistently higher plate counts using an
8-fold diluted Standard Methods plate count medium supplemented with Iron. Conditions of
incubation often differ among laboratories, and this makes comparIsons and evaluations of
bacterial plate count data difficult. Table 2 shows the Incubation conditions recommended by
various agencies and laboratories in different parts of the world for the bacterial plate count test.
Temperatures of 20, 22, 28, 35 and 37°C and Incubation times of as long as 9 days and as short as I
day have been employed.
Alien, el al., (1976) and Geldreich, ., (1978) demonstrated that the presence of high
levels of bacteria interferes with the determination of coliforms with the membrane filter
procedure. They have suggested that the Incubation temperature and time for the standard plate
count does not necessarily reflect the number of microorganisms capable of growth on the medium
at optimum condition. Significant alteration of incubation temperature and time does dramatically
alter the plate count.
Figure I shows the number of colonies, as a function of time, on standard plate count agar
incubated at 20°C and 35°C. At 35°C the plate count increased with time till a plateau was
observed at 6 days, while at 20°C the plate count appeared to reach a plateau at 12 to 14 days.
Whether the time or temperature of incubation has any particular significance and whether the
plateaus are consistent for other samples and conditions remain to be determined.
-------
9
Table I. Sanitary Quality of Wafer
(in Prescott & Winslow 1904)
Quality No. Bacteria/cc.
Excessively pure < 10
Very pure 10-100
Pure 100-1,000
Mediocre 1,00010,000
Impure 10,000-100,000
Very impure > 100,000
-------
I0
Table 2. Recommended or Reported Incubation Temperature and Time for
The Bacterial Plate Count for Water Samples (Snead, et c ii., 1980)
Source
Temperature °C
Time, days
European Economic CommunIty
(1975)
37
2
USEPA
(1975)
35
2
German Water Regulations
(1975)
20
2
United Kingdom Report #71
37
I
Water Research Center,
United Kingdom 1976
22
22
3
7
Vlctoreen
(1977)
28
.
Standard Methods
(1971)
35
20
I
2
Snead,etal
(1980)
20
35
9
i
-------
I I
T ;—•
IL”
, I 35’C
0 I 2 3 4 5
INCUBATION TIME
Tv ’
A
6 7 a
d cy $
120
100
80
60
z
Z 40
20
120
100
0
o 90
60
40
20
0 ___________
Effect of Incubation tim 1 on the nignber of bacterial colonies obtained on standard
plate count ogar at 20 C and 35 C. Each point presents the mean of 35 to
37 samples, while the bars represent I standard deviation around the mean (Srtead,
!t9i• 1980)
0 2 4 6 8 JO 2
INCIJ8ATION T$M , do
14 16
20 C
Figure I.
-------
12
In the present state of knowledge, the bacterial plate count is believed to have only limited
sanitary significance. No relationship between it and enteric diseases transmitted by water has
been reported. However, members of genera containing some species pathogenic to man
( Pseudomonas, Flavobocterium and Aeromonas ) have regularly been isolated from plates for the
standard plate count test (Allen, et al ., 19Th and Geldreich, et 01., 1978). These pathogenic
microorganisms are opportunistic in nature and present a health hazard in unique environments or
circumstances. The bacterial plate count does provide an exceptionally sensitive measure of water
treatment operation and changes in water quality in the distribution system. The bacterial plate
count may be used as a method to check microbial growth in systems. it may become an
increasingly important tool to monitor the condition of activated carbon filters in the treatment of
water (i-lass, 1982).
Chemical Indicators.
Free Residual Chlorine . The free residual chlorine (HOd and/or OCI under normal pH
values found in drinking water) has been a useful indicator of water quality in municipal water
systems. The National Interim Primary Drinking Water Regulations in 1975 allowed for the
substitution of chlorine residual data for up to 75% of the monthly microbiological samples. The
free residual chlorine determination has the definite advantage of providing: (I) immediate
Information to evaluate water quality, (2) continuous monitoring data and (3) a red flag for
potentially unsafe conditions (loss of a free chlorine residual).
The species of chlorine present in the distribution system decidedly affects the residual
performance and usefulness as an indicator. Snead, et al. (1980) reported that a free chlorine
residual affords more protection than an equivalent level of combined chlorine under the same
environmental conditions, Free chlorine was a more potent and faster acting bactericide and
viricide for each of the conditions of pH, temperature and sewage challenge tested. In addition, a
free chlorine residual can serve as a “marker” for contamination. Since free chlorine is normally
maintained, the absence of a free residual is evidence that chlorine demanding substances may
have entered the system. Chemical results obtained in these experiments Indicate that a total
chlorine residual is present even after the addition of sizeabie amounts of contaminant so that the
detection of a combined chlorine residual does not assure water potabilIty. The relative
-------
13
Inefficiency of a combined chlorine residual in protecting against contamination materials and its
failure to act as a “marker” indicate that the addition of ammonia to create chloramfnes is
contrary to good public health practice.
It is important to emphasize that the use of free chlorine as a “marker or flag” is heavily
dependent upon the ability to measure free chlorine in the field and on the chlorine residual history
in the distribution system. Despite the existence of numerous color tests and field procedures for
the determination of chlorine, a reliable, sensitive and specific test for free available chlorine is
not yet available. Cooper, j. (1974) compared chlorine determinations by the modified
orthotoildine-arsenite, stabilized neutral orthotolidine (SNORT), leuco crystal violet, N,
N-diethyl -p-phenylene diamine (DPD) and syrlngaldazine (FACTS) tests. In the hands of operators,
all except syringatdazine yielded false positive readings for free chlorine. The FACTS procedure
was, however, the least sensitive of the methods tested. DPD has since been modified to the DPD
steadiFAC by the addition of thloacetamide to enhance free chlorine specificity (Palm, 1978), and
FACTS has been modified to increase sensItivity (Cooper, et cii., 1975). The presence or absence of
free chlorine is difficult to interpret unless the chlorine residual history at a given sample station
is known. Several stations in the Baltimore City distribution system consistently have zero free
chlorine In late summer months. The absence of free chlorine at these stations at this time of year
does not suggest any post-treatment contamination. The conspicuous absence of free chlorine at a
station where free chlorine was continually observed deserves serious attention.
The free chlorine residual Is more difficult to maintain than a combined chlorine residual.
Converting a distribution system to free chlorine requires considerable effort and patience.
Several months or years at elevated chlorine levels are often required to “push” a free chlorine
residual Into the distribution system (Buelow and Walton, 1971 and Umbenhaver, 1959). A free
chlorine residual Is sometimes Impossible to maintain in systems where excessive corrosion,
turberculatlon and scaling are found.
Fecal Sterols . Coprostanol, coprosterol, cholesterol and coprostanofle have been suggested as
chemical indicators of feces of man and higher animals and several investigators have shown a
close correlation between levels of fecol sterols and fecal contamination (Dutka, et al., 1974;
Klrchmer, 1971; Murtaugh and Bunch, 1967; SmIth and Gouron, 1969; Tabak, et c i., 1972; and Wun,
cii., 1976). However, cholesterol was also found in foods and was found to be widely distributed
-------
14
in sea water (Smith and Gouron, 1969) and coprostanone was found in low levels in feces (Dougan
and Tan, 1973). Quantitative determinations of fecal sterols may be useful since less preparation
and processing time are required compared to bacterial Indicators.
Considerably more information is necessary before fecal sterois can serve as indicators in
drinking water. Biodegradation of fecal sterols and their response to water/wastewater treatment
processes require further elaboration. The fecal sterois appear to be unaffected by chlorination
and heat (NATO, 1980).
Other Determinations . Limulus amebocyte lysate (LAL) assays for the levels of endotoxins
(Levin and Bang, 1964), adenoslne triphosphate (AlP) measurements (Picciolo etal., 1977),
epifluorescence counts (Hobbie j al., 1977), carbon- 14 release (Levin, 1963), and optical brightner
determinations have been suggested as indicators. However, the significance in drinking water of
these methods remains to be determined.
Evaluatien aid Application of Indk utars of Mic obiaI Quality
The panel’s evaluations of selected surrogate measurements in drinking water are summarized
in Table 3. Each surrogate was rated on a scale of - to ++++ by their utility for each category of
Indicator application. E. coIl was rated as the most useful Indicator of fecal contamination. Total
cotiform, heterotrophic plate count and free chlorine residual were felt to provide the best
indicators for treatment efficiency and post treatment degradation, deterioration and/or con-
tamination.
The panel’s recommended guidelines for surrogate measurements for drinking water systems
are shown In Table 4. Routine testing was defined as collection and assaying samples at regular
intervals Periodic testing was defined as collecting and assaying samples at less frequent, but
regular intervals. Diagnostic sampling is used as required and is designed to solve or define
problems occurring in the water network. Chlorine demand, turbidity and any fecal indicator were
recommended for testing of source waters. Chemical and physical measurements were proposed
because of their ability to provide rapid Information on the quality of source water, and the fecal
indicator was recommended because it can provide information on long term water quality changes.
Free chlorine residual, turbidity, total coliform and aerobic heterotrophic bacteria were recom-
mended for testing treatment efficiency, and any deterioration, degradation or recontamination of
-------
Table 3. Candidate Indicator Rating for Each Aspect of Surrogate Measurements in Drinking Water
Feccil
Degradation
Deterioration
Contamination
Treatment
Recontamination
Comments
Indicator (ci)
(b)
(c)
Total Coliform ++ ++++ (b) best available-has deficien-
cies relative to viruses and
Glardia .
Fecal Coliform ++ +4
E. coil ++ ++ (c) use for recontamination
problems rather than deterior-
ation.
Fecol streptococci ++ ++ ++
Enterococci ++ (b) use when concerned about
viral Inactivation.
Coliphage + (b) may be useful as viral
surrogate data ore limited.
Mycobocteria-Veast + (b) potentially useful for this
application-data limited.
Plate Count ++++ (b) some as (b) for total coliforms
p rfrinqens ++ +4 (a) excessively persistent
(c) Not useful for degradation
problems.
idobacteri i ++ (a) useful only for extremely
recent contamination.
(b) may be useful as indicator
for viruses and Glardia
(C) growth nutrient stimulated
Aeromonas ++ (a) has many alternate sources
eudomonas + + (b) has potential for growth on
activated charcoal.
Fecol Sterols 4+ (a) not amenable to continuous
assay.
Optical Brighteners ++ (c) recontamination indicator
AlP (c) need more research
-------
16
Table 3. CandIdate Indicator Rating for Each Aspect of Surrogate Measurements in Drinking Water
(Cont’d)
Degradation
Fecal
Contamination
Indicator (a)
Treatment
(b)
Deterioration
Recontamination
(c)
Comments
LAL -
-
-
(c) data limited for this use.
Epifluorescence -
+
-
(b) may be useful-limited data
available.
(c) need more research
Disinfectant Residual -
++++
(Free
chlorine)
++++
(Free
chlorine)
(c) use with total coliforms and
plate counts, good for “real
time”, free chlorine best.
C 14 Sugar -
+
+
(a) not to be employed
for coliform determinations.
(c) need more research.
Turbidity
+÷+
+++
(b) should be used in
conjunction with chlorine
residual measurement.
-------
17
Table 4. Recommended Surrogate Measurements for Water Treatment Systems
Class of
Test
Source
Water
Treatment
Train
Distribution
System
Routine
Periodic
(I) Chlorine Demand
(2) TurbIdity
(3) Fecal Indicator
(I) Free Chlorine Resdual
(2) TurbIdity
(3) Total Coliform
(4) Plate Count
(I) Enterococci
(2) C. perfringens
(l)FreeChlorineResidual
(2) Turbidity
(3) Total Coliform
(4) Plate Count
(I) Enterococci
(2) C. perfringens
Diagnostic
Sanitary Survey
(I) Sanitary Survey
(2) Microbial IdentifIcation
(I) Sanitary Survey
(2) Mcroblal Identification
-------
18
the distribution system. Enterococci and Clostridium perfringens spores were suggested for
periodic testing in the treatment train in order to provide a record of inactivation and removal of
indicators known to be more resistant to chlorination than the more commonly used total coliform
group. These two organisms were also suggested as indicators of water quality testing of
distribution systems.
The sanitary survey was recommended as the most useful approach to diagnosing water
quality problems in the water system. Ancillary microbial identification tests should be used in
conjunction with the sanitary survey for problems occurring in treatment trains and the distribution
system.
Each of surrogate measurements proposed above should be employed in the same manner that
a physician uses disease symptoms, laboratory analysis and patient history to diagnose illness and to
prescribe a remedy. Each determination represents an important piece of information that can be
used to interpret the water quality. The more the Information, the better the interpretation.
The heterotrophic plate count was considered carefully by the measurement panel. While it
was agreed that the recommendation of a regulation with inherent legal requirements and
enforcement provisions would not be useful, the application of the plate count for water treatment
quality control and distribution system monitoring remains desirable. Every water network should
be encouraged to apply a consistent heterotrophic plate count method. The plate count results
should be recorded on a quality control chart, and observed values exceeding established limits
should evoke changes in the water system operation.
Interpretatian of indicutori of Microbial Q’ aflty
The qualitative and quantitative determinations of bacterial indicators were never intended
to be the sole information on which to judge the health hazard associated with a particular water
or to evaluate water quality. The importance of a field or sanitary survey cannot be overly
stressed.
Fox (1886) recognized the importance of “the history of a water.” Rafter (1889) emphasized
“personnel inspection” and Drown (1892) required a “knowledge of locality and surroundings” in
interpreting bacteriological and chemical analysis. Prescott and Winslow (1904) point out:
“The first attempt of the expert called in to pronounce on the character of a
potable water should be to make a thorough sanitary inspection of the pond,
stream, or well from which it is derived. Study of the possible sources of
-------
pollution on a watershed, of the direction and velocity of currents above and
below ground, of the character of soil and the liability to contamination by
surface-wash ore conceded to yield evidence of the greatest value.”
In 19 14, the Commission appointed by the Treasury Department to recommend standards of purity
for drinking water supplied to the public by Common Carriers engaged in Interstate Traffic stated
(Prescott and Winslow 1904):
“It Is a fact so well established as to need no further discussion that the results of
bacteriological and chemical examination of a sample of water ought always to be
correlated with the knowledge of the source, treatment, and storage of the supply
in order to enable a Just estimate of the sanitary quality of such supply.”
Subsequent advisory committees in 1925 and 1943 reiterated the importance of the sanitary
survey. The United States Public Health Service continued to stress the sanitary survey In the
Manual of Recommended Water Sanitation Practice in 1944. While the importance of the sanitary
survey has been stressed for potable waters, similar importance of the sanitary survey should be
noted for recreational, shellfish, and other waters. Unfortunately, the sanitary survey and field
observation were often omitted and evaluation of water quality and health effects were based
solely on analytical determinations.
To date, the coliform group Is still the most reliable indicator for potable water. No other
organism or group of organisms has been found to fulfill the above criteria for potable water. The
recent National Academy of Sciences (1977) report, Drinking Water and Health , stated:
“It would be undesirable and extremely risky to substitute any organism for the
coliform group now, although research studies that compare other indicator
organisms with coliforms are warranted.”
The plate count is an Important and useful measurement of water quality and should be
included as one of the microbial observations In drinking water. However, the plate count has
limited sanitary significance and a questionable relat1onsh p-tO health. The conditions for
incubation for the plate count differ from laboratojylo laboratory. The standard plate count
( Standard Methods , 1981) is useful for the eyaI otion of interference in the coliform determination
but may not be the best indicator of mt robIaI quality.
The substitution of chlorine residual determinations for a portion of the required microbial
measurements circumvents some Of the disadvantages of the bacterial assays. However, the
species of chlorine present is critical. Free chlorine functions as a flog or marker while combined
chlorine does not. The measurement of free chlorine in the field is difficult, and most methods will
yield false positive readings in the presence of combined chlorine.
-------
20
REGULATORY ASPECTS
HstoslcaI aid Global Perspective
Standards and regulations have evolved from the information gathered by early micro-
biologists. A review of the significance of the microbiological measurements can be seen in the
previously mentioned book by Prescott and Winslow (1904). In 1897, the American Public Health
Association (APHA) appointed a committee to review and recommend procedures and methods for
the determination of microorganisms in water. The procedures recommended by the APHA
committee evolved in 1904 to the first edition of Standard Methods and provided a methodological
underpinning for the initial bacteriological standard for drinking water developed by U.S. Public
Health Service and applied by the U.S. Treasury Department (USTD) to wafer supplies used by
interstate carriers under the authority provided by the Interstate Quarantine Acts of 1893 and 1897
(in Wolf, 1972 and Chamberlin, et L, 1979). The USTD standard established the “limits of
permissible impurity” for drinking water in terms of £91! and a 37°C plate count after 24 hour
incubation. The limits for B. call were not more than one out of five $0 cc portions of any sample
should show the presence of B.coli. The plate count should not exceed 100 per cc. It should be
noted that physical and chemical limits and the sanitary survey were not included in the 1904 USTD
standard but the importance of the sanitary survey was recognized by the committee that
formulated the standard (Wolf, 1972). Chamberlin, et al. (1979) noted that administrative
expediency was the compelling argument for the bacteriological standard. The sanitary survey was
too difficult to administer, and thus the cornerstone of interpretation of water quality was
sacrif iced. Subsequent revisions of the drinking water standards again restored the sanitary survey,
but the bacteriological standard had evolved as a “magic number” for water potability.
The Royal Institute of Public Health in Great Britain appointed a committee of bacteri-
ologists and sanifarians to study drinking water quality. In contrast to the USPHS committee, the
British group refused to “lay down any fixed standards.” They did, however, suggest that 0.3
B,colI/lOO ml indicated waters of good quality in (Chamberlin, etal., 1979).
The drinking water standards in the United States have been subsequently revised by the U.S.
Pvbllc Service in 1925, 1942, 1946 and 1962 and by the U.S. Environmental Protection Agency In
1975. The World Health Organization (WHO) recommended standards in 1958, 1963 and 1971, and
-------
21
European standards were recommended in 1961 and 1970. The highlights of these bacteriological
regulations were given by Chamberlin, et al. (1979) and are included in Appendix A. Major changes
from the original standards and guidelines include: (I) recognition of the statistical nature of
bacteria determinations, (2) requirements for samples to be collected in the distribution system,
(3) acceptance of the membrane filter method for determination of coliforms, and (4) recognition
of differences between community and individual water supplies.
Despite all the discussions, commissions, committees, workgroups, regulatory agencies and
reports throughout the world, there is an unparalleled agreement on the measurement of drinking
water quality. The current coliform group has evolved from B. coli originally isolated by Escherich
(1885) as the single most utilized determination for wafer quality. The effectiveness of the
coliform group as a fecal indicator and the usefulness of coliform determinations for the control of
transmission of disease by the water route are well established. With only a few exceptions over
the last 80 years, the epidemiological evidence strongly supports the conclusion that waters
containing less than I coliform/l00 ml and having a good sanitary survey are unlikely to transmit
detectable levels of infectious disease.
The reporting of some waterborne diseases is not so rigorously practiced today as to rule out
the unnoticed occurrence of large numbers of cases of waterborne disease. These diseases,
however, may also be transmitted by nonwaterborne routes, and only a few cases may be
attributable to wafer. If so, the complete prevention of waterborne transmission of the disease in
question is unlikely to have a perceptible effect upon the total annual incidence of the disease.
Even if waterborne disease is not occurring at a rate that is significant on a national scale,
there may well be local incidents in which contaminated water produces a high level of infection
with a given agent or causes the introduction of a “new” agent into a community. Gastroenteritis
of unknown or diverse etiology has been shown to occur, in incidents undetected by health
authorities, in small communities served by water supplies that are of inferior microbiological
quality as determined by already established criteria. Infectious agents may be introduced into a
community in which they have not previously occurred through drinking water, whose source is a
river that receives waste discharges from an upstream community. The proposition is sufficiently
plausible, and the means to test it are well enough established, that a “worst-case” study of
communities located along a multiple-reuse waterway seems warranted. Such a study would
-------
22
probably represent a better use of available resources for disease prevention than new regulations
that would require testing for pathogens, or more rigorous water or wastewater treatment.
Urbai vs. Rral
The data base developed over the last 80 years supports the use of the coliform index In
municipal systems where the drinking water receives considerable treatment. Such systems are
characteristic of urban areas and serve the overwhelming majority of the population in the United
States. The very small municipal systems or individual water supplies derived from ground water
or surface waters that receive little or no treatment have frequently been implicated in disease
transmission. Ground water surveys suggest that total coliform may be of limited value since the
levels of coliforms may be masked by excessive bacterial populations (Pipes, 1978), and total
coliforms may come from a variety of sources other than feces. Recent outbreaks of giardiasis in
unfiltered supplies suggest that coliforms may not be useful indicators under these conditions.
The addition of supplemental tests for other indicators has been suggested. The plate count,
C. perfrngens, Bacteroides and Bifidobocterium bifidus may prove to be useful determinations to
help Interpret the coliform data. A reemphasis of the sanitary survey certainly would provide a
considerable amount of Information to interpret water quality in rural areas.
The impact of the USTD regulation on the transmission of disease by water provides an
Interesting insight into the use of regulation and standards. Figure 2 shows the typhoid death rate
in the United States per 100,000 populatIon for the years 1880 to 1980 (Kruse, et ol., 1981). In 1880
almost 100 persons died of typhoid fever in the United States per 100,000 peopIe. In 1914 when the
USTD regulation was placed in effect, the typhoid death rate per 100,000 population had already
decreased to just more than 10 typhoid deaths per 100,000. The typhoid death rate had been
reduced by almost 90% before the coliform standard was in effect. It should be noted that this
dramatic decrease in disease was not due solely to the improvement in urban water supplies.
Better sewage collection and disposal, milk and food sanitation, insect control and nutrition were
also important factors. Nevertheless, significant reductions in typhoid death rates were observed
In the absence of regulations.
Early workers generally opposed the implementation of regulations and standards. Sedgewick
(in Chamberlin, et al., 1979) commented
-------
23
0
0
100
8O
0 6O
;40
-a0
w
8
6
I - .
4
a
a
0
= I
1880 $900 19Z0 $940 $960 $900
YEARS
FIgure 2.
Typhoid death rates in the United States - 1880 to 1980 (Kruse, et cii., 198 I).
-------
21&
“Standards are. . . devices to save lazy minds the trouble of thinking.”
“Standards are often the guess of one worker, easily seized upon, quoted and requoted, until
they assume the semblance of authority.”
Wolman (1960) noted that standards tend to discourage the application of common sense and good
judgement and tend to be treated as “magic” numbers. These magic numbers were often chosen
without a firm scientific base, and were difficult to revise thereafter (Wolman, 1960; Chamberlin,
et cii., 1979). The standards do not provide absolute assurance and in some cases restrict the
utilization of the tools at hand to evaluate water quality. Analytical methodology to determine the
coliform itself is a poor, Imprecise “magic number”. Conventional multiple tube fermentation
procedures for 5-10.0 ml, 5-1.0 ml and 5-0.1 ml sample volumes have a most probable number
(MPN) less than 2 coliforms/I00 ml when no tubes are positive. The coliform test lacks specificity
and represents a large group of microorganisms. Despite all the difficulties with the coliform
measurement, the coliform index has had a significant impact on the transmission of disease by
water. However, use of the coliform measurement before it was required by regulation had
promoted the evolution of effective treatment processes, sound treatment plant design, reliable
operation and maintenance procedures, and good engineering practices.
In a contemporary Industrialized society, standards and regulations have evolved as adminis-
trative realities. They are a necessary evil and provide a mechanism for even handed administra-
tive practices and judicial decisions. For the standards to be effective, they should reflect
scientific thinking and good judgment. The standards should be realistic, straight-forward and
simple and above all have a demonstrable Impact on the problem to be solved. Unnecessary
regulations cloud the real problems, confuse the priorities and divert the effort. Regulations for
the sake of regulations will create more problems than they solve.
RESEARCH RECOMMENDATIONS
Currently there are several unresolved issues that requIre attention. Unless these issues are
addressed, they will continue to trouble future workshops, committees and panels. The issues are
unresolved due to a lack of data. The following research work is needed:
I. Studies to determine the level of d’sease in the population transmitted by drinking
water.
2. Correlation of the Incidence of enteric disease to observed indicator levels in drinking
water.
-------
25
3. Development and in-system testing of rapid and preferably real-time bacterial mea-
surement systems, including fluorescence methods to detect laundry whiteners!
brighteners; epifluorescence, ATP, LAL, and 14 C-release methods to defect and
quantify bacterial levels; and multi-test, C-retease methods to provide rapid
detection of specific genera.
L&. Determination of the conditions under which turbidity Interferes with effective
disinfection of viruses and other pathogens.
5. Correlation of enterococci and Clostridlum perfringens levels in actual distribution
systems to observed total coliform and free chlorine residual levels.
6. Measurement of Campylobacter, Yersinla , and Legionella death rates in water distri-
bution systems and assessment of the availability and utility of surrogate
measurements.
7. Assessment of the levels and significance of bacteria known or suspected to grow within
distribution systems, Including genera detected by total coliform methods ( Enter-
obocter, Kiebsiella , and Citrobacter ) and other genera including Flavobacterium,
Acinetobacter, Aeromonas , and non-fluorescent pseudomonads.
8. ComparIson and evaluations of alternative heterotrophic aerobic plate count methods
based on application In diverse source waters and regions.
9. Broadening the data on occurrence of coliphage (both DNA and RNA) in human and
animal fecal material and response of coliphage to water treatment processes.
SUMMARY AND CONCLUSIONS
I. The existing total coliform and turbidity regulations are appropriate; their proper use
can ensure, essentially If not absolutely, that diseases will not be transmitted by
drinking water.
2. The sanitary survey should be reemphasized by specific Incorporation IntO the primary
regulations. This will permit better Interpretation of the microblolobical, chemical and
physical measurements.
3. E. coil was rated by the panel as the most useful indicator of fecol contamination.
Total coliforms, heteroirophic plate counts, and free chlorine residuals were felt to
provide the best indicators for treatment efficiency and post treatment deterioration.
-------
26
LITERATURE CITED
Allen, M. J., R. H. Taylor, and E. E. Geldreich 1976. The impact of Excessive Bacterial
Populations on Coliform Methodology. In: Proceedings of the Am Water Works Assoc. Water
Quality Technology Conference, San Diego, California.
Barrow, G. I. 1977. Bacterial Indicators and Standards of Water Quality in Britain. In: Bacterial
Indicators/Health Hazards Associated with Water, A. W. Hoadi>’ and B. J. Dutka, (eds.),
American Soc. for Testing and Materials, Philadelphia, Pennsylvania.
Buck, J. D. 1977. Caridida albicans . In: Bacterial Indicators/Health Hazards Associated with
Water, A. W. Hoadly and B. J. Dutka, (eds.), American Soc. for Testing and Materials,
Philadelphia, Pennsylvania.
Budd, W. 1857, 1931. Typhoid Fever, Its Nature, Mode of Spreading and Prevention, The
Commonwealth Fund, New York.
Buelow, R. W., and Walton, C. 1971. Bacteriological Quality vs. Residual Chlorine J. Am. Water
Works Assoc., 63:28.
Cabelli, V. J., H. Kennedy, and M. A. Levin. 1976. Pseudomonas aeruginosa - Fecal Coliform
Relationships in Estuarine and Fresh Recreational Waters. J. of Water Pollution Control
Fed., 48:367-376.
Cabelli, V. J. 1977. Clostridium perfringens as a Water Quality Indicator. In: Bacterial
Indicators/Health Hazards Associated with Water, A. W. Hoadly and B. J. utka, (eds.),
American Soc. for Testing and Materials, Philadelphia, Pennsylvania.
Chadwick, E. 1842. Report from the Poor Law Commissioners on an Inquiry into the Sanitary
Conditions of the Laboring Populations of Great Britain, Her Majesty’s Stationery Office,
London.
Chamberlin, C. E., J. Boland, A. Malik, and I-I. Shipmon 1979. Wholesome and Potable Drinking
Water: A Background Paper on Water Quality Aspects of Water Supply. Agency for
International Development, Washington, DC.
Clausen, E. M., B. L. Green, and W. Litsky 1977. Fecal Streptococci: indicators of Pollution. In:
Bacterial Indicators/Health Hazards Associated with Water, A.W. 1-badly and B.J. Dutka,
(eds), American Society for Testing and Materials. Philadelphia, Pennsylvania.
Cooper, W. J., E. P. Meler, J. W. Highfill and C. A. Sorber 1974. The Evaluation of Existing Field
Test Kits for Determining Free Chlorine Residuals in Aqueous Solutions, Final Report. U.S.
Army Medical Bloengineering Research and Development Laboratory Technical Report 7402.
Cooper, W. J., C. A. Sorber, and E. P. Meler 1975. A Rapid Specific Free Available Chlorine Test
with Syringeldazine (FACTS). J. Am. Water Works Assoc., 67(1 ) :34-39 .
Cramer, W. N., K. Kawata, and C. W. Kruse 1976. ChlorInation and lodinat ion of Pol lovirus and f2.
J. of Water Pollution Control Fed., 48:61.
Drown, T. M. 1892. The Interpretation of Water Analysis, Annual Report of the Massachusetts
State Board of Health. Ins State Sanitation Vol. Il, G. C. Whipple, (ed.), Harvard University
Press, Chicago, IilInols.
Dougan, J., and L. Tan 1973. Detection and Quantitative Measurement of Fecal Water Pollution
Using a Solid-Injection Gas Chromatographic Technique and Fecal Steroids as a Chemical
Index. J. of Chromatography, : 107-I 16.
-------
27
Dufour, A.P. (977. Eschedchia coIl: The Fecal Coliform. In: Bacterial Indicators/Health Hazards
Associated with Water, A.W Hoadly and B.J. Dutkaieds.), American Soc. for Testing and
Materials, Philadelphia, Pennsylvania.
Dutka, B.J., A.S.Y. Chau, and J. Coburn (974. RelationshIp Between Bacterial Indicators of Water
Pollution and Fecal Sterols. Water Res. 8:1047-1055.
Erigelbrecht, RS. and C.N. Hoas (977. Acid-fast and Yeast Organisms as Disinfection indicators:
Enumeration Methodology. j : Proceedings of the Water Quality Technology Conference,
Am. Water Works Assoc.
Escherich, T. (885. DIe Dormbacteren des Neugeborenen und Sauglings, Fortschritte der
Medicine, Ill, 5 5, 547.
Evison, L.M., and A. James 1974. Blfidobacteriurn as an indicator of Fecoi Pollution In Water. in:
Proceedings of the 7th International Conference on Water Pollution Research, Pergam i
Press, Ltd, New York, New York.
European Economic CommunIty (975. Proposals for a Council Directive Relating to the Quality of
Water for Human Consumption. Journal of European Community C2 (4/2, Brussels.
Fox, C. B. (886. Sanitary Examinations of Water, Air, and Food . J. A. Churchill, London.
Geidreich, E. E. (977. Bacterial Populations and Indicator Concepts in Feces, Sewage, Stormwater
and Solid Wastes. in: indicators of Virus in Water and Food , G. Berg, (ed.), Ann Arbor
Science, Ann Arbor, lch [ gan.
Geidreich, E. E., M. J. Allen, and R. H. Taylor 978. interferences in Coliform Detection In
Potable Water Supplies. In: Evaluation of the MicrobioIo y Standards for Drinking Waters C.
W. Hendricks, (ed.), PA-57O/9-78-0OG. U.S. Environmental Protection Agency ,
Washington, DC.
Geidreich, E. E., R. H. Bordner, C. B. Huff, H. C. Clark, and P. W. Kabier 1962. Type Distribution
of Coliform Bacteria in the Feces of Warm-blooded Animals. J. of Water Pollution Control
Fed. 3Li(3)295.
Geldreich, E. E., and B. A. Kenner 1969. Concepts of Fecal Streptococci in Stream Pollution. J. of
Water Pollution Control Fed., 41:336-352.
German Water Regulations. Verordnung Uber Tr lnkwasser and Brauchwasser fur Lebensmittel
betriebe.
Garbow, W. 0. K. (968. The Virology of Waste Water Treatment. Water Res. 2:675.
Hoadly, A.W. (968. The Significance of Pseudomonas aeruginosa in Surface Water. J. of New
England Water Works Assoc. 83:99.
I-loather, R. C. 1952. Bacteriological Examination of Water, 8011, FluorIdes In Water Supplies,
1161 Oe, J. of Inst. Water Engineers 6:426.
Hobble, J.E, R.J. Daley, and S. Jasper. (977. Use of Nucleopore Filters for Counting Bacteria by
Fluorescence Microscopy. Appli. Environ. Microblol . 33: (225-1228.
Klrchmer,.C. J. 1971. 5 -choleston-3 -at: An Indicator of Pollution. Ph.D. Thesis, University of
Florida.
Kott, V. (977. Current Concepts of Indicator Bacteria. In: Bacterial Indicators/Health Hazards
Associated with Water, A.W. i-badly and BJ. Dutka, (eds.), American Soc. for Testing and
Materials, Philadelphia, Pennsylvania.
-------
28
Kruse, C.W., M.C. Snead, and V.P. Olivieri 1981. Design of Alternative Disinfecting Systems. In:
Proceedings of the 23rd Annual Public Water Supply Engineers’ Conference, Chicago, Illinois.
Leclerc, H., D.A. Massol, P.A. Trinel, and F. Gavini 1977. Microbiological Monitoring - A New
Test for Fecal Contamination. In: Bacterial Indicators/Health Hazards Associated with
Water, A.W. Hoodly and B.J. l5utka, (eds.), American Soc. for Testing and Materials,
Philadelphia, Pennsylvania.
Levin, G.V. 1963. Rapid Microbiological Determinations with Radiolsotopes. Advances in Applied
Microbiology (Edited by Umbreit, W.W.) 5:95- 133, Academic Press, New York, USA.
Levin, J. and F.B. Bang, 964. A Description of Cellular Coagulation in Limulus . Bull., Johns
Hopkins Hospital 115:337-345.
Levin, M.A., 1977. Bifidobacteria as Water Quality Indicators. In: Bacterial Indicators/Health
Hazards Assoc. with Water, A.W. Hoadly and B.J. Dutka, (eds. American Society for Testing
and Materials, Philadelphia, Pennsylvania.
MaIlman, W.L. 1961. The Enterococci. In: Proc. Rudolf Conference: Public Health Hazards of
Microbial Pollution of Water, Rutgers University, New Brunswick, New Jersey.
Murtaugh, J.J. and R.L. Bunch 1967. Sterols as a Measure of Fecal Pollution. J. of Water Pollution
Control Fed., 39:404-409.
National Academy of Science 1977. Drinking Water and Health . Washington, DC.
National Interim Primary Drinking Water Regulations, 1975. U.S. EPA.
NATO 1980. Pilot Study on Drinking Water supply Problems.
Olivieri, V.P. 1974. The Mode of Action of Chlorine on f2 Bacterial Virus. ScD Thesis, The Johns
Hopkins University, Baltimore, Maryland.
Palm, A.T. 1978. A New DPD-SteadiFAC Method for Specific Determination of Free Available
Chlorine. Presented before the Division of Environmental Chemistry, American Chemical
Society, Anaheim, California.
Picciolo, G.L., Thomas, R.R., Deriiing, J.W., Chappelle, E.W. 1977. Environmental Applications of
the Firefly Luciferase ATP Assay Flow Techniques for Monitoring the Wastewater Effluent
and of Drinking Water Supplies. j. : Proceedinqs of the Second Bi-Annual ATP Methodology
Symposium , pp. 547-564, SA I Technology, 4060 Sorrento Valley Blvd., San Diego, california.
Pipes, W.O., 1978. Water Quality and Health Significance of Bacterial Indicators of Pollution,
Workshop Proceedings, Drexel University and National Science Foundation.
Poppell, C.F. 1979. Enumeration and Occurrence of RNA Coliphciges in Wastewater. Master of
Science Degree Thesis, The Johns Hopkins University, Baltimore, Maryland.
Prescott, S.C., and C.E.A. Winslow 1904. Elements of Water Bacteriology . John Wiley & Sons,
New York, New York.
Rafter, B.W. 1889. On the Fresh Water Algae and Their Relation to the Purity of Public Water
Supplies. Trans. American Society of Civil Engineers 21:483.
Scarpino, P.V. 1978. Bacteriophage Indicators. In: Indicators of Viruses in Water and Food , G.
Berg, (ed.), Ann Arbor Science, Ann Arbor, Michigan.
Shah, P.C., and J. McCamish 1972. Relative Resistance of Poliovirus I and Coliphage f2 and 12 in
Water. Applied Microbiology, 29:658.
-------
29
Shattuck, L., N.P. Banks Jr., and J. Abbot 1917. Report of the Massachusetts Sanitary
Commission. In: State Sanitation Vol. 1, G.C. Whipple, (ed.), Harvard University Press,
Cambridge, Massachusetts.
Smith, L.L. and RE. Gouron 1969. Sterol Metabolism. VI. Detection of 5-cholestan-3-ol in
Polluted Waters. Water Res. 3:1k 1-148.
Snead, M.C., V.P. Olivieri, C.W. Kruse and K. Kawata I 980. Benefits of Maintaining a Chlorine
Residual in Water Supply Systems. EPA-600/2-80-OlO.
Snow, J. 1855, 1936. On the Mode of Communication of Cholera. In: Snow on Chlorea , 2nd ed.,
The Commonwealth Fund, New York.
Standard Methods for the Examination of Water and Wastewater, 15th ed., 1981. American Public
Health Association, New York, New York.
Tabak, H.H., RN. Bloomhuff, and R.L. Bunch 1972. Coprostanol: A Positive Tracer of Fecal
Pollution. In: Development in Jndustrial Microbiology, V. 13:296-307. Publication of the
Society for T dustrial Microbiology, Am. Inst. of Biological ScT Washington, DC.
Umbenhaver, E.J. 1959. Chlorine Residual. J. Am. Water Works Assoc., 51:224.
United Kingdom Report #71 1969. - Department of Health and Social Security The Bacteriological
Examination of Water Supplies. Reports on Public Health and Medical Subjects. H.M.
Stationery Office, London, UK.
U.S. Public Health Service 944. Manual of Recommended Water Sanitation Practice . U.S.
Department of Health, Education, and Welfare, Washington D.C.
U.S. Treasury Department 1914. Bacteriological Standard for Drinking Water. Public Health Rep.
29:2959-2966.
Victoreen, H.T. 1977. Water Quality Deterioration in Pipelines. In: Proceedings of Water Quality
Technology Conf. Am. Water Works Assoc., Kansas City, Missouri.
von Fritch, A., 1882. Zur. Aetiologie des Rhinosklorems. Viener Medizinische Wochenschrift, 32,
970.
Water Research Center 1976. United Kingdom. Deterioration of Bacteriological Quality of Water
During Distribution. Notes on Water Research, No. 6:1.
Wolf, H. 1972. The Coliform Count as a Measure of Water Quality In: Water Pollution
Microbiology , R. Mitchell, (ed.), Wiley-Interscience, New York, New Yorl , p. 333-345.
Wolman, A. 1960. Concepts of Policy in the Formation of So-Called Standards of Health and
Safety. J. Am. Water Works Assoc., Vol. 52.
Wun, C. K., R. W. Walker, and W. Litsky 1976. The Use of XAD-2 Resin for the Analysis of
Coprostanol In Water. Water Res. 10:955-959.
-------
30
APPENDIX A
WHOLESOME AND PALATABLE DRINKING WATER:
A BACKGROUND PAPER ON WATER QUALITY ASPECTS OF WATER SUPPLY
By Charles E. Chamberlin, John Boland, Arunpal MaHk, & Harold Shlpman
July 31, 1979
Agency for International Development, Washington, D.C.
-------
31
Bacteriological Standards
U.S. Treasury Department Standard — 1914
Maximum limits of permissible bacteriological impurity:
(1) The total number of bact ria developing on standard agar plates,
incubated 24 hours at 37 C shall not exceed 100 per cubic centimeter.
The estimate shall be made from not less than two plates and should
be reliable and accurate..
(2) Not more than one out of five lOc.c. portions of any sample examined
shall show the presence of B. coli. (A testing procedure which
demonstrates the presence of aerobic, gas forming, lactose—fermenting
organisms is outlined; this procedure is essentially equivalent to
a completed test.)
U.S. Public Health Service Standards
1925
(1) Not more than lO%ofba i. the lOc.c. portionsa examined shall show the
presence of B. coli . ‘
(2) Occasionally, three or more of the five lOc.c. portions ofa sample
may show the presence of B. coiL This shall not be allowed if it
occurs in more than—
(a) 57. of the samples when 20 or more samples are examined;
(b) One èample when less than 20 samples are examined.
The series of samples must conform to both requirements (1) and (2) above.
The completed test is considered evidence of the presence of B. coli .
1943
These standards allow for the use of either 10 ml. portionsa or
100 ml. portions
a 5 standard portions of lOc.c. (ml.) each constitute a standard sample.
The term “standard” (portion or sample) has been omitted in the text
for brevity.
bB. coli group as defined in Standard Methods of Water Analysis , 1923.
cSee note at end of this section.
standard portions of 100 ml. each constitute a standard sample. The
term “standard” (sample or portion) has been omitted in the text for
brevity.
-------
32
If 10 ml. portions used:
(1) Not more than 10% of all the 10 ml. por ions examined per month shall
show the presence of the colifortn group
(2) Occasionally 3 or more of the 5 10 ml. portions of a sample may show
the presence of the coliform group. This shall not be allowable if
it occurs in consecutive samples or in more than
(a) 5% of the samples when 20 or more samples are examined per month.
(b) One sample when less than 20 samnies are examined per month.
If 100 ml. portions used:
(3) Not more than 60% of all the 100 ml. portions examined per month
shall show the presence of the colif arm group.
(4) Occasionally all of the five 100 ml. portions of a sample may show
the presence of the colif arm group. This shall not be allowed if it
occurs in consecutive samples or in more than—
(a) 20% of the samples when 5 or more samples are examined per month.
(b) One sample when less than 5 samples are examined per month.
The series of samples must conform to either requir nents (1) and (2) or
requirements (3) and (4) above.
(5) When three or more of the five 10 ml. portions, or all 5 of the 100 ml.
portions, constituting a standard sample show the presence of the
coliform.group, daily samples shall be col1ec ed promptly and examined
until the results be of satisfactory qua1ity.
The completed test or the confirmed test under certain conditions
specified in the Standards, is considered evidence of the presence of the
colif arm group.
1946
Standards are identified to the 1943 USPHS Standards except that it
is now stated that samples collected following an unsatisfactory sample
(e.g. as in item (5) above) shall not be included in the determination
of the number of samples examined monthly. Neither shall subsequent
unsatisfactory samples in this daily series be used as a basis f or
prohibiting the supply, provided that (1) immediate, active efforts are
made. to locate the cause of contamination, (2) immediate action is taken
eColiform group of bacteria includes all organisms of the coli—aerogenis
group ‘as set forth in Standard Methods for Examination of Water and
Sewage , 1936. See note at end of this section.
When this occurs in waters of unknown quality, simultaneous tests should
be made on multiple portions of a geometric series.
-------
33
to eliminate such cause and (3) samples taken following such remedial
action are satisfactory.
1962
These standards allow for use of either the fermentation tube method
or the membrane filter technique.
If fermentation tube method used with either 10 ml. or 100 ml.. portions:
Standards are essentially identical to the 1946 TJSPHS Standards, except
for a few minor changes in wording.
If membrane filter technique used:
(1) The arithmetic mean coliform 8 density of.all standard samples examined
per month shall not exceed one per 100 ml.
(2) Coliform colonies per standard sample shall not exceed 3/50 ml. ,
4/100 nil., 7/200 ml. or 13/500 ml. in:
(a) Two consecutive samples;
(b) More than one sample when less than 20 are examined per month;
(c) More than 5% of the samples when 20 or more are examined per
month.
(3) When coliform colonies in a single standard sample ex eed the above
values, daily samples shall be collected promptly and examined until
the results obtained from two consecutive samples show the water to
be of satisfactory quality . These unsatisfactory samples are
regarded in the same manner as those used for the fermentation tube
method — see 1946 USPHS Standards.
U.S. Environmental Protection Agency Regulationé — 1975
Maximum contaminant levels.
These standards also allow for the use of either the fermentation
tube method or the membrane filter technique.
If membrane filter technique used:
(1) See corresponding item (1) in 1962 USP S Standards above.
(2) The number of coliform bacteria shall not exceed 4/100 ml. in:
(a) More than one sample when less than 20 are examined per month; or
(b) More than 5% of the samples when more than 20 are examined per
month.
group as defined in Standard Methods for the Examination of.
Water and Wastewater , current edition. See note at end of this
section.
-------
34
(3) When the coliform bacteria in a single sample exceed 4/100 ml., at
least two consecutive daily check samples shall be collected and
examined until the results from two consecutive samples show less
than one coliform bacterium per 100 ml.
If fermentation tube method used, with either 10 ml. or 100 ml. portions:
Standards are essentially identical to those specified in items
(1), (2), (3), and (4) of the 1943 USPHS Standards except for the elimin-
ation of the requirement regarding the number of portions showing the
presence of coliforms in consecutive samples.
The requirement in these Standards corresponding to item (5) in the
1943 TJSPHS Standards is more stringent in that twice daily resampling
(check samples) must be continued until the results from two consecutive
samples show no positive tubes .
Check samples, both for fermentation tube and membrane filter
techniques, shall not be included in calculating the number of samples
taken each month for compliance with sampling frequency requirements.
Neither check samples nor special purpose samples shall be used to
determine compliance with maximum contaminant levels f or co] .iform bacteria.
In addition there are requirements for record maintenance, routine
reports to the State and, in the event that a maximum contaminant level
is exceeded, for notification of the State and the public.
Note: The definitions of 3. coli , the “coli—aerogenes g çoup” and the
“coliform group” are considered equivalent (AP}IA 1.965).
-------
35
Bacteriol &ical Standards
World Health Organization Standards
tnt ernat ional
1938
I. Reco ended standards for treated water.
a. Coliforin bacteria shall not be detected in 90% of the samples
examined in any year, or the MPH shall be ‘ .1.0.
b No sampl. shall have an MPH index 10.
c. An MPN index of 8—3.0 should not occur in consecutive samples.
d. See item d below.
II. Recommended standards for untreated water.
a. The MPN index of 90% of the samples examined in any year should
b.’l0.
b. No sample should show an MPH under 20.
c. An MPH index 15 should not be permitted in consecutive samples.
d. When two consecutive samples show an MPH iitd.x 8, in the case
of created watir, or 10 in th. case of untreated water
additional sample. from the sampting point should be examined
immediately. Further investigation may also be desirable.
1963
These standards also allow for the us. of the membrane filter technique.
I. Recommended standards for treated water.
Requirements are identical to thos. of 3.958 International Standards.
Colif arm group includes ai .1a.robic and facultative anaerobic Gram—
negative non—spore—forming rods ca ble of fermenting lactose with the
produc’tion of acid and gas at 35—37 C in ‘ 48 hours.
MPN index for coliform bacteria (in mU cases here),.
-------
3’
If membrane filter technique used:
The arithmetic mean of numbers of coliform group organisms shall be
‘ 1 per 100 ml., and shall not exceed 4 per 100 ml. in two consecutive
samples or in more than 10% of the samples examined.
II. Recommended standards for untreated water.
Requirements include all those of 1958 International Standards and
an additional requirement that no more than 40% of the number of cólifo m
microorganisms shown by the MPN index shall be faecal coliform bacteria
If membrane filter technique used:
The arithmetic mean of the numbers of coliform group bacteria
determined shall be ‘ 10 per 100 m l .. , and shall not be 20 per 100 ml.
in two consecutive samples or in more than 10% of the samples examined.
1971
I. Standards recommended for piped supplies.
1. Water entering the distribution system.
(a) Chlorinated or otherwise disinfected supplies.
Coliform organisms should be absent in any sample of 100 ml.
If this standard is not met, an immediate investi a ion
into the efficacy of the purification process and the method
of sampling and testing.
(b) Non —disinfected supplies.
E. co1i should be absent in 100 ml. Occasionally, if
E. coli is absent, the presence of 3 co].iform organisms
per 100 ml. may be tolerated. If this is exceeded the
supply should be considered unsuitable for use without
disinfection.
2. Water in the distribution system.
(1) 95% of the samples in any year should not contain any
coliform organisms in 100 ml.
hpaecai coliform group is defined as a Cram—negative non—spore—forming,
rod which is capab e of fermenting lactose with the production of
acid and gas at 44 C in ‘ 24 hours.
E. coli is regarded as a Gram—negative, non—spore—forming rod capable
of 0 fetmenting lactose with the production of acid and gas at both
37 C and 47 0 C in 48 hours; it produces indole in peptone water
containing tryptophane and is incapable of utilizing sodium citrate
as its sole source of carbon.
-------
37
(2) No sample should contain E. coli in 100 ml.
(3) No sample should contain more than 10 coliform organisms
per 100 ml.
(4) Coliform organisms should not be detectable in 100 ml. of
any two consecutive samples.
If any coliform organisms are found the xniniin .un action required is
ixrm ediate re—sampling. Additional investigation may be advisable,
depending on local conditions.
LI. Standards recommended for individual or small, community supplies.
The standard f or piped supplies should be aimed at and everything
possible should be done to prevent pollution of the water.
It should be possible to achieve a coliform count of ‘ 10 per 100 ml.
Persistent failure to achieve this, especially if E. coli is. repeatedly.
found, should lead to condemnation of the supply.
European
1961
Standards recommended for piped supplies.
1. Water entering the distribution system.
Coliforin organisms must be absent, whether the water is disinfected
or naturally pure. In either case, the presence of colifortn
organisms calls for immediate investigation.
2. Water in the distribution system.
The presence of one or more coliform organisms in a 100 ml. sample
can be permitted in 5% of the samples examined if a positive result
is not obtaine4 in two or more consecutive samples and at least 100
samples of 100 ml. each, regularly distributed over the year, are
examined.
When one 100 ml. sample shows the presence of coliform organisms, a
further sample from the sampling point should be examined immediately.
Additional investigation may be advisable, depending on local conditions.
1970
Standards are essentially identical to the 1961 European Standards.
-------
38
APPENDIX B
MINORITY REPORT - MEASUREMENT OF MICROBIAL QUALITY
-------
3,
APPENDIX B
Minority Report - Measurement of Microbial Quality
The minority report is at variance with that of the majority on two major issues and the
implications thereof, They are (I) the manner in which the objective of water quality is stated
with reference to the potential for infectious disease and (2) the choice of the water quality
indicators to be used in monitoring against the objective.
Water-Quality Objectives. An alternative to the objective as stated in the majority report is,
that human pathogens should not be present in drinking water at levels which will produce
unacceptable levels of infectious disease. This restatement of the objective In the context of the
acceptability of risk rather than its detectability has two implications.
First , the modified statement does not guarantee “safe” water in the sense of absolute
protection from all possible waterborne, Infectious diseases. Reported outbreaks of drinking
waterborne viral gastroenteritis continue to occur, although the coliform standard was reported to
have been met in most of those outbreaks associated with ground water supplies (Wilson, et of.,
1981). However, in their review of Norwalk virus gastroenteritis outbreaks, including several
associated with drinking water, Kaplan, et al. (1982) found that the attack rate exceeded
20 percent in every case. Undoubtedly, outbreaks with lower attack rates are occurring but go
undetected because of the passive nature of the surveillance system, especially for this rather
benign, “nonreportable” Illness. This raises the possibility that such outbreaks and, more likely,
“sporadic cases” of viral gastroenteritis occur from drinking water which meets (if barely so)
existing standards. This possibility also is suggested from the acute gastroenteritis-indicator
relationships obtained in the bathing beach epldemlological studies (Cabelli, al., 1982).
Furthermore, the degree of treatment and disinfection needed to reduce the level of the etlological
agents beyond some acceptable incidence of drinking waterborne, viral gastroenteritis, much less
to guarantee the absence of the agents, may be prohibitive. At least with regard to chlorination,
this Is suggested by some recent results of a studyof the chlorination of the human rotavlrus (M.
Butler, personal communication). The findings from an active disease surveillance system,
specifically a prospective epidemiological study, are needed. Until the results from such a study
-------
40
show that sporadic cases do not occur from the consumption of finished waters which meet current
regulations, it would seem prudent to use the more realistic definition.
Second , the objective as stated in the majority report is an open invitation to examine ever
increasing quantities of water for the presence of all potentially waterborne human pathogens
without regard to the available epidemiological evidence concerning their importance or their
infective doses. Both the expenditure of funds and the concern caused the public by the
announcement of the finding of a single cell or virion in some unrealistically large quantity of
water are hard to justify. Furthermore, the objective as stated places an unnecessary requirement
on the sensitivity of the indicator systems used to monitor water quality or the quantity of water
to be examined. Thus, one of the criteria for a surrogate measurement could be restated as
follows: ideally, the indicator should be present in some “reasonable” quantity of water (100-
1000 ml) In numbers such that a reasonably precise density estimate can be made (CV 25 percent)
in finished waters with an unacceptably high risk of infectious disease.
Water Quality Indicators. The following considerations were used in arriving at the
recommended surrogate measurements for water treatment systems. First, the trends In reported
drinking wafer-associated disease outbreaks over the past two decades indicate that the potential
for giardioses and viral gastroenteritis will most limit the choice of microbial indicators and
standards used in preventing infectious disease via this transmission route. Second, the majority of
the available information on the survival of coliforms relative to viruses in aquatic environments
(VasI, 198 I), in sludge (Berg and Berman, 1980) and, during chlorination (Scarpino, 19Th) is that the
former generally do not persist as well as the latter. Moreover, it has been shown by Vasl (1981)
and Berman and Berg (1980) that fecal streptococci or enterococci better simulate the survival of
viruses than do the coliforms.
Third, as implied in the majority report, there is a need for “real time”, automated, surrogate
measurements, at least for public water supplies. Ideally the systems would have alarms, and even
automatic feed-bock loops for corrective action. Where disinfection Is practiced, the measurement
of an active disinfectant residual and turbidity approaches, if not reaches, this objective.
Therefore, we subscribe to the “comments” in Table 3 about the deficiencies in the total and fecal
coliform measurements, In the use of enterococci for the concern about viral inactivation and the
value of a free residual chlorine determination for a real time measurement, where applicable.
-------
41
Moreover, we feel that these comments should be acted upon as per the modification of Table 4
shown in Table B-I.
Since enterococcus and C. perfringens measurements are recommended as surrogates, it
would seem appropriate also to recommend some limits which can be used as guidelines or, at least,
“yellow flags.” There is no epidemiological-microbiologicol data base from which to make
judgments. However, the same can be said for the total coliform standard. This information along
with some notion as to the “best” indicators could be obtained from the prospective epidemiological
studies noted earlier and constitutes another reason for their conduct.
LITERATURE CITED
Berg, G. and D. Berman 1980. Destruction by Anaerobic Mesophillic and Thermophillic Digestion
of Viruses and Indicator Bacteria md igenous to Domestic Sludge. Appl. and Env. Microbiol.
39: 361-368.
Cabelli, V.J., A.P. Dufour, L.J. McCabe, and MA. Levin 1982. Swimming-Associated Castro-
enteritis and Water Quality. Am. J. Epidem. 115: 606-616.
Lupo, L. 1979. Bacteriophage as Indicators of Fecal Pollution. M. S. Thesis, University of Rhode
Island, Kingston, Rhode Island.
Kaplan, J.E., CW. Gary, R.C. Baron, N. Singh, L.B. Schonberger, R. Feldman, and H.E. Greenberg
1982. Norwalk Gostroenteritis and the role of Norwalk Virus in Outbreaks of Acute Bacterial
Gastroenteritis. Ann. of Int. Med. 96: 756-761.
Scarpino, P.V., M. Lucas, D.R. Dahling, G. Berg, and S.L. Chang 1974. Effectiveness of
Hypochlorous Acid and Hypochlorite Ion in Destruction of viruses and Bacteria. In: A. J.
Rubin. Chemistry of Water Supply, Treatment and Distribution (ed.). Ann Arbor ci., Ann
Arbor, Mich.
Vasl, R., B. Fattal, E. Kotzenelson, and H. Shuval 1981. Survival of Enteroviruses and Bacterial
Indicator Organisms in the Sea. In: Viruses and Wastewater Treatment , M. Goddard and M.
Butler (eds.), Pergamon Press, Lt2 Oxford, England.
Wilson R., C.E. Haley, D. Relman, E. Llppy, G.F. Craun, G.C. Morris, Jr., and J.M. Hughes 1981.
Waterborne Outbreaks Related to Contaminated Ground Water Reported to CDC 1971-1979.
Conference on Microbiological Health Considerations of Soil DIsposal of Domestic Waste-
water, No!man, Oklahoma.
-------
Table B-i.
Surrogate Measurements for Water Treatment Systems
Class of Test
*Confj ijous or a
Mean-continuous
Routine
Source Water
(I) Chlorine Demand
(2) Turbidity
(I) Chlorine Demand
(2) Turbidity
*(3) Fecal Indicator X
*(3) Fecal Indicator
Treatment Train
(I) Free Chlorine Residual
(2) Turbidity
(I) Free Chlorine Residual
(2) Turbidity
*(3) Total Coliform X
(4) Plate Co& f
(I) Enterococci
(2) Cl perfringens
Periodic
Distribution System
Diagnostic
Sanitary Survey
(I) Free Chlorine Residual
(2) Turbidity
(I) Enterococci
(2) Cl perfringens
*(3) T tal and Differential
Coliform
*(4) Plate Count b
*(5) A. hydrophila
(I) Sanitary Survey
(2) Microbial Identification
(I) Sanitary Survey
(2) Microbial Identification
*Differences between Minority and Majority recommended measurements.
a re applicable, routine where not.
b an indicator of “nutrient loading or accumulation”; instructive use.
-------
MONITORING OF MICROBIAL WATER QUALITY
Wesley 0. Pipes
Panel Members: R. Bordner, R. R. Christian, A. H. El Shaarawi, C. W. Fuhs, H. Kennedy, E. Means,
R. Moser, H. Victoreen
ABSTRACT 2
INTRODUCTION 2
OBJECTIVES OF MONITORING 2
Safety 3
Reliability 3
Contamination 5
MEASURES OF MICROBIOLOGICAL WATER QUALITY 5
Coliform Bacteria 6
Residual Chlorine 6
Turbidity 7
Plate Count
Other Measures 8
MECHANICS OF MONITORING 8
Dispersion of Coliform Bacteria in a Water Distribution System 9
Random Dispersion-Poisson Distribution I I
Fraction Positive-Binomial Distribution 12
Models of Aggregation-Lognormal Distribution 13
Chlorine and Turbidity 18
Residual Chlorine 18
Turbidity 18
Selection of Sampling Locations 20
Distribution System Elements 20
Representative Sampling LocatIons 22
Special Sampling Problems 23
MAXIMUM CONTAMINANT LEVEL 24
Coliform Bacteria 24
Statist ical Parameter 26
Examination Method 27
Sample Volume 27
Reporting Period 27
Water Quality Evaluation Period 28
MCL 28
Number of Samples 31
Remedial Action 32
Chlorine Substitution 32
Turbidity 33
MONITORING GUIDELINES
Microbiological Monitoring During Treatment 33
Finished Water Monitoring 33
Water Source Considerations 35
Monitoring of Treatment Processes 35
Monitoring During Treatment Modifications 36
Monitoring Stored Water 36
Distribution System Monitoring 36
Seasonal Usage 36
Emergency Conditions 37
COST OF MICROBIOLOGICAL MONITORING 37
Collection and Transportation 37
Examination Costs 38
Violation of the MCL 39
LITERATURE CITED
APPENDIX A Minority Report 41
-------
2
ABSTRACT
The microbiological maximum contaminant levels (MCL’s) of the National Interim Primary
Drinking Water Regulations (NIPDWR) are extremely complex. They consist of nine different rules
using three different statistical parameters. This issues paper presents information and a line of
reasoning oriented toward a single microbiological MCL which is more effective in limiting the
extent of microbiological contamination in a water distribution system and which will be easier to
understand than the present rules. The proposals presented include: I) elimination of mean
coliform density as an MCL parameter; 2) reliance on the fraction of samples with coliforms
present as the sole parameter for the MCL; and 3) an increase in the minimum number of samples
examined each month for the smallest systems from I to 5. The need to distribute sample
collection sites around the entire distribution system and vary the sites from month to month is
discussed. Also discussed are monitoring of turbidity, substitution of free chlorine measurements
for coliform measurements, guidelines for monitoring and the cost of microbiological monitoring.
INTRODUCTION
The rationale behind microbiological monitoring of water distribution systems is to provide
evidence of protection against the transmission of waterborne infectious diseases. Historically,
there have been many major epidemics of diseases transmitted by water contaminated by human
wastes. Although the incidence of waterborne disease in the U.S. has been relatively low since
1950 compared to the early part of the twentieth century, the potential for waterborne spread of
infectious diseases still exists (Craun, 1978). Protection against waterborne disease depends upon
multiple barriers; i.e., selection of the waters of the highest quality available for public supply,
protection of the water source, use of various treatment processes to remove particulate material,
disinfection, and maintaining the physical integrity of the distribution system to prevent cross
connections and leaks into the system. The most difficult barrier to maintain, and the one whose
penetration poses the greatest threat of waterborne disease, is the physical Integrity of the
distribution system.
OBJECTIVES OF MONITORING
Outbreaks of waterborne disease are very rare in any one system but twenty to thirty
outbreaks are reported in the United States every year. Thus, It. appears that the multiple barrier
-------
3
approach as presently practiced is effective but its effectiveness is not absolute. Microbiological
monitoring of the product delivered to the consumer is one method of directing attention to the
necessity of maintenance of the barriers. A distinction can be made between the safety of an
individual portion of the water and the reliability of the water system as is explained below. The
Monitoring Panel takes the position that routine microbiological monitoring of water distribution
systems is intended to demonstrate the reliability of the overall system and not necessarily the
safety of Individual portions of the water.
5—
Although there are no infallible Indicators, the demonstration of the absence of bacteria
which could be of fecal origin from water in a distribution system does provide some reasonable
assurance of the safety of that water In terms of a low probability of transmission of waterborne
infectious disease. Unfortunately, It is impractical to test all of the water which will be consumed
by humans to achieve this reasonable assurance of safety. It is quicker and more effective to boil
and cool water before consumption than to test for microorganisms. However, the purpose of the
water system, which includes the water source and the collection, treatment, storage and
distribution facilities, is to provide water which does not have to be boiled before it Is consumed.
The potential number of samples which could be collected from a water distribution system
and examined Is, for all practical purposes, infinite in relation to the actual number of samples
tested and the safety of any Inalvidual portion of the water is never established. The major
question about the objectives of monitoring to be considered is what Inferences can be drawn about
the reliability of the system from the samples which are collected and tested.
Relithitity
It must be admitted that for any water supply there will always be some conditions under
which the barriers against transmission of waterborne disease may be penetrated. RelIabilIty Is a
characteristic which Is related to persistence of protection in time and over the area served by the
distribution system. A distribution system which is free of microbiological contamination one day
but not the next would not be considered to be reliably protected. Likewise, a system free of
microbiological contamination in some areas but not in other areas would not be considered to be
reliably protected. Thus, in monitoring to demonstrate reliability of protection the sampling must
extend over time and over the entire area covered by the distribution system.
-------
k
In any water distribution system, there are a large but finite number of sites where samples
may be collected and, at any time, not all of the water is equally accessible. it would not be
practical to attempt to sample every potential sampling site in a month’s time. The best that can
be done to Include all of the water In the system is to make sure that none of the potential
sampling sites are excluded from being sampled. Contamination may occur undetected for a short
period of time in a limited port of the distribution system but, if it persists over a long period of
time, it should be detected by the monitoring program.
The accumulation of negative results over a period of time develops confidence that the
barriers to microbial contamination of the water are reliable. The extent of microbiolglcai
contamination of a water distribution system may change from one day to the next and almost
certainly will change over the seasons of a year. These changes can be considered to be normal
variability resulting from the characteristics of the water system. The characteristics of the
system may change when there are alterations in the water source, treatment processes, or
distribution system. Examples of these alterations would Include changes in water treatment
processes, placing new water mains in service, replacement or repair of existing water mains, and
abnormal variations in water pressure. These are identifiable occurrences and can be used to
separate groups of data used to demonstrate reliability of the system.
Even in the absence of identifiable changes in the water system, it may be expected that
there wilt be slow deterioration ? the distribution system which would change the reliability of
protection against microbial contamination. It seems likely that the characteristics of a water
system would remain consistent for more than one month but probably not for as long as five years.
It also seems reasonable to evaluate the reliability of protection of a water supply ogainst
microbiological contamination on the basis of data collected over one or two-year periods in the
absence of identifiable occurrences which could change the characteristics of the system.
There Is a precedent for evaluating the reliability of protection over a one-year period under
the previous drinking water standards of the Public Health Service (1962). Water supplies had
“approval” if they met the mi robloiogicai standards all 12 months of a year, “provisional approval”
if the standards were violated only once in 12 months, and were “use prohibited” if the standards
were violated any two months in a year’s times.
-------
5
Clearly, evaluating Indicator bacteria results over a month’s time Is a matter of convenience
for reporting and has no particular scientific basis. However, there is no evidence that hourly,
daily, weekly, quarterly, or yearly evaluation would give a superior measure of indicator bacteria
occurrence in the system. Since the monitoring and reporting system is run by people for the
protection of people, it is appropriate to Select the reporting period as a matter of convenience in
the absence of scientific evidence that one reporting period would be superior to any other.
CcntamInaTton
The most serious degradation of drinking water quality by microorganisms occurs when
pathogenic organisms are present which, clearly, makes the water not suitable for drinking. Since
pathogens are very difficult to detect in water and most of the etlological agents of waterborne
disease are fecal organisms, contamination of water with fecal matter (or sewage) Is considered to
make the water unsafe to drink; although, in fact not all fecal matter contains pathogens. Some
group of feccil (or sewage) bacteria Is used as an indicator of contamination of the drinking water.
The density of the Indicator bacteria In sewage Is much, much higher than the densities of the
pathogens and there often are non-fecal and non-sewage sources of some members of the indIcator
group. There are many Instances when Indicator bacteria are present In drinking water and the
pathogens are not. Because of the very high ratio of Indicator bacteria to pathogeris in sewage, the
non-sewage sources of some members of the indicator groups, and the impossibIlity of keeping all
of the water passing through a distribution system free of indicator bacteria at all times, an
arbitrarily selected density level Is designated as the line of demarcation between “contaminated”
and “uncontaminated” water. The density level is selected low in order to be very conservative in
protecting health but high enough to be attainable in terms of current practice. “Contaminated”
water might have some pathogens present but, on the other hand, it might be safe to drink.
MEASURES OF MICROBIOLOGICAL WATER QUALITY
It is assumed that there will be some monitoring of treated water entering a distribution
system and of water in the distribution system itself to provide some assurance of the rellablility
of the water system in protecting against transmission of woterborn. disease. The measures of
microbiological water quality which have been used in the past are described herein, It is
recognized that other measures of microbiological quality might be adopted for future drinking
-------
6
water regulations and one such new measure is described briefly because it has some features
advantageous for monitoring.
Colifcrm Bacteria
The coliform group of bacteria has been used for over seventy years as a measure of the
sanitary quality of water. The group is operationally defined at present by the laboratory
procedures used for detection and Includes Escherichia which is the most numerous focultative
bacterium in the feces of warm-blooded animals, plus species belonging to the genera Entero-
bacter, Klebsielia , and Citrobocter which are present in sewage but can be derived from other
environmental sources such as soil and plant materials. Experience has demonstrated that a
properly treated and disinfected water supply can be essentially free of coliform bacteria at the
point of entry to the distribution system. The occurrence of coliform bacteria in a distribution
system can result from lapses in treatment, leakage of contaminated water into the distribution
system, Introduction of bacteria from the air at points where there is an air-water Interface (such
as in storage tanks), or growth of coliform bacteria in the pipes which make up the system.
Under the present Regulations (EPA, 1976) there are two accepted laboratory techniques for
estimation of coliform density of a sample; namely, the membrane filter (MF) procedure and the
fermentation tube (FT) technique. The MF procedure results in a count of colony-forming units
(CFU) per 100 ml and the results of the FT technique are used to make a most probable number
(MPN) estimate of the coliform density of the sample. Both methods are subject to interferences
from large numbers of noncoliform bacteria In the samples (Seidler, etal., 1981).
Residual Chlorine
Chlorine is by far the most widely used disinfectant for potable water supplies. Since
chlorine reacts with some substances In the water, a dosage adequate to leave a residual chlorine
concentration after several hours should be used. Distribution systems contain materials which
reduce chlorine and the chlorine residual may disappear In ports of some systems.
There is no current national requirement that chlorine be used for disinfection or that a
chlorine residual be maintained in the distribution system. However, If a free chlorine residual of
at least 0.2 mg/I Is maintained in all parts of the distribution system and samples are collected
every day, chlorine residual measurements may be substituted for up to 75 percent of the samples
-------
7
to be examined for coliform bacteria on the basis of at least four chlorine residual tests for each
microblologial sample replaced (EPA, 1976). Chlorine residual in samples from The distribution
system is usually measured colorimetrically with a visual comparator at the time of sample
collection. Chlorine residual can also be measured electrometrically and some water treatment
plants monitor and record chlorine residual in the finished water continuously.
Turbidity
Turbidity of water is caused by any particles which scatter light. This includes micro-
organisms and many different types of inorganic and organic materials. The effects of turbidity in
relation to the use of the water depends to a very large extent on the nature of the material which
causes the turbidity.
Particulate matter may have a chlorine demand or may surround viruses or pathogenic
bacteria and protect them from the disinfectant effects of chlorine and may interfere with the
maintenance of a chlorine residual throughout a distribution system. If a water sample collected
for bacteriologlc l examination by the MF technique has a very high turbidity, the particles may
coot the surface of the filter and interfere with the development of coliform colonies so that
coliform bacteria are not detected in samples where they are present. Turbidity in drinking water
also is an aesthetic problem because most people prefer to drink clear water rather than turbid
water. The turbidity particles may also Interfere with the use of the water for laundry and
cleaning.
The present MCL for turbidity is one turbidity unit (TU) for a monthly average and 5 TU for
the average of two consecutive days (EPA, 1976). Samples are collected daily at representative
points of entry of treated water into the distribution system. There is no limit for the turbidity of
samples collected from other parts of the distribution system.
Plale Casmt
There are many heferotrophic bacteria which can grow in the water In a distribution system.
Many of these are considered to be Innocuous but some are opportunistic pathogens and some
Interfere with the detection of coliforms In water samples when they are present. Although there
Is no MCL for overall bacterial density at the present time, water with a low bacterial density is
considered to beof better quality than water of a high bacterial density (Vietoreen, 1969).
-------
8
The heterotrophic bacterial population of a sample of water is measured by pour-plating or
spread-plating a specified volume (usually I ml) and incubating the plates for a set period of time
at an established temperature. There is also an MF plate count procedure for larger volume
samples. Several different media, incubation temperatures, and incubation times have been tried
in the past. Each variation in technique gives a somewhat different plate count; however, any of
the several standardized procedures will give results which can show relative changes.
Aerobic bacteria grow slowly in drinking water and several days are required for the water to
reach the maximum density which may be l0 to l0 colony forming units (CFU) per ml or higher.
In most water distribution systems the average residence time of the water is less than one day and
bacterial densities are usually low. However, there may be areas of quasi-stagnant water where
bacterial densities can reach high levels particularly during the warmer months of the year.
OTher M
It is possible to estimate the densities of other groups of indicator organisms in samples from
water distribution systems. Each of the other groups of indicator organisms provides some
information about water quality which is somewhat different from the information gained from
detection of the presence of members of the coliform group.
There is one relatively new procedure which deserves special mention as another measure of
microbiological contamination of drinking water because it shows promise of providing more
information abotit the microbiological quality of drinking water at a lower cost. This is the
presence-absence (P-A) test (Clark, 1969 and Clark, 1980). It is a fermentation tube method which
gives evidence of the occurrence of several indicators. The indicators which can be differentiated
by subsequent confirmatory tests include coliforms, fecal coliforms, fecal streptococci,
Staphylococcus, Pseudomonas aeruginosa, Aeromonas , and Clostridium perfringens .
MECHANICS OF MONITORING
The mechanics of monitoring include: how many samples to collect; when to collect them;
where to collect them; and what inferences can be drawn from the data, If is Impossible to collect
and examine a large number of samples from a water distribution system simultaneously. Thus, the
“population” being sampled Is all possible standard volume samples of the water which flows
through the distribution system during the sampling period. The goals are to detect lnd cator
-------
9
bacteria when and where they are present, to obtain reliable estimates of the indicator bacteria
densities, and to obtain reliable estimates of the variation in concentration of continuously variable
parameters (such as residual chlorine and turbidity) related to, microbiological quality.
If a water supply is properly treated and the distribution system is protected, very few of the
samples collected for microbiological monitoring will have coliforms present. The standard sample
for the ME is 100 ml and the standard sample for the FT technique is five 10 ml portions from a
sample of about 100 ml or five 100 ml portions from a sample of about one liter. In either case the
result is an estimate of the coliform bacteria density per 100 ml. In the vast majority of cases, the
result is negative and the density of the samples is recorded <1/100 ml when the MF procedure is
used and either <2.2/100 ml for five 10 ml portions or <0.22/100 ml for five 100 ml portions when
the FT technique is used. When no cotiforms are found in a sample, this does not mean that there
are no coliforms In the water In the distribution system, only that the coliform density is below
what can be detected in a sample of limited volume. This lack of ability to detect low coliform
densities is called “low range data truncation.”
There is also an upper limit to the coliform bacteria density which can be measured using the
accepted sampling and analytical procedures. When the number of colonies on a membrane filter is
greater than 80 It Is almost certain that some colonies merge and cannot be counted separately.
Thus, ME counts greater than 80 are recorded as too numerous to count (TNTC) or >80/100 ml. The
MPN estimate of the upper limit of detection using the FT technique with five 10 ml portions is
>16/100 ml and>l.6/lOO ml with five tOO ml portions. This is “high range data truncation.”
Coliform data as obtained, using presently accepted sampling and analytical methods, are
double truncated data. This double truncation Introduces a large measure of uncertainty into the
statistical analyses of these data.
Dispersion of Collforrn Bacteria In a Water Distribution System
Table I presents some cal iform monitoring data obtained using the MF technique from four
different water distribution systems in Pennsylvania and New Jersey (Pipes and ChrIstian, 1982).
All the systems are relatively small. The WH and BL systems are groundwater systems and the
other two have surface supplies.
In all systems except WH the water is filtered. The WH monitoring data for 1979 and 1980
are typical. Normally four samples per month are collected from that system to meet the present
-------
l0
Table I. Examples of coliform monitoring data collected from various
water distribution systems.
System
WH
CV
D I
BL
Sampling
Period
4/27
1979 1980 1981 1981
March
1979
January
1980
April June
1980 1981
Number of
Samples
Number of
Samples with
noColiforms
Counfsin
Samples with
Coliforms
(per lOOmI)
48 48 41
118 59 4 32
I I I
I I
5 12 I
15 I
I
2
3
15
42
138
133
I
I
2
2
8
105
101
I
I
4
>80
107 l 9
104 157
I
I I
>80
I
2
2
2
3
3
4
5
78
mean
variance
0 0.35 0.27 1.63
0 3.94 3.01 47.44
0.10
0.53
>0.82
>61.03
>0.9 0.61
>59.8 36.29
fraction with
counts >4
0 0.034 0.022 0.049
0.007
0.019
0.009 0.0(2
-------
II
monitoring requirements. In 1979 monitoring samples, no coliforms were found. In 1980, positive
samples were found two months and extra samples were taken in an attempt to “average down.”
Even so, for one of the months, the arithmetic mean of the sample counts was above I per 100 ml.
Two sets of samples collected from the WH system on individual days in May and June 1981 are
included for comparison purposes. The data for the other three communities were special sets
collected over one-month periods.
Data sets showing no coliforms in any samples or showing a greater fraction of samples with
coliforms and higher coliform counts could have been selected. However, these sets are not
atypical and they provide the data needed to illustrate the pertinent features for analysis. In all
cases the great majority of sample counts are zero. When coliforms are found, the count Is most
likely to be I or 2/100 ml but higher counts are not rare.
Random Dispersion - Poisson Distribution. One part of the present microbiological MCL Is a
mean MF count of no more than 1/100 ml for all samples examined during a month (EPA, 1976). If
the FT technique is used, the equivalent part of the MCL is positive results in no more than
10 percent of all 10 ml portions examined or in no more than 60 percent of all 100 ml portions
examined. The 10 percent positive results in all 10 ml portions is an overall MPN of 1.051100 ml
and the 60 percent positive results in all 100 ml portions is an overall MPN of 0.92/100 ml. A
statistical interpretation of this part of the MCL is that we want to use the sample mean, x = Xj/fl
where the x values are the individual sample counts and n Is the number of samples, to estimate
the actual mean density of the coliform bacteria in the water flowing through the distribution
system during the water quality evaluation period.
The most parsimonious model of the dispersion of coliform bacteria in water flowing through
a distribution system Is based on the assumption that the Indicator bacteria are randomly dispersed.
This is a naive assumption but the model gives the smallest possible number of samples needed to
achieve some preselected precision for parameter estimation and thus provides a lower limit for
sampling frequency.
A random dispersion of coliform bacteria in the water passing through a water distribution
system implies that counts in unit volume samples will fit a Poisson distribution. An essential
characteristic of the Poisson distribution is that the variance Is equal to the mean. From the
values near the bottom of Table I, It Is clear that the variance Is considerably larger than the
mean.
-------
12
If the coliform bacteria were randomly dispersed In the water passing through a distribution
system and the mean density were greater than 1/100 ml, then monitoring to demonstrate this
would require only a few samples. With random dispersion, all sampling locations are equal; the
coliform bacteria are just as likely to be found one place as another. Suppose that we adopt the
criterion that when the arithmetic mean of the sample counts Is 1/100 ml we want 95 percent
confidence that the actual mean density in the water passing through the distribution system Is less
than 2/IOU ml. This criterion can be stated mathematically as
<2
where is the mean sample count, t is from Student’s t distribution, S 2 is the variance of the
sample counts, and n Is the number of samples collected during the water quality evaluation period.
We expect S 2 to be close to c. The value of t varies with n, but for a one-sIded 95 percent
confidence interval, It is approximately two for thre to ten samples. When these substitutions are
mode, the criterion becomes 2 117 < I which has a solution of n >5. In other words, In the simplest
case (random dispersion), which Is not realistic, at least five samples are required to provide
95 percent confidenc. that the actual mean density of coliform bacteria in the system Is less than
2/100 ml when the mean sample count Is 1/100 ml.
Fraction Positive - Binomial Distribution. The binomial distribution can be used to model The
frequency of occurrence of coliform bacteria above son e selected density level. Part of the
present microbiological MCL (EPA, 1976) is that no more than 5 percent of the samples, examined
by the MF method, have a coliform count of more than 4 per 100 ml. When the FT technique Is
used, the MCL Is no more than 5 percent of the samples with three or more of five 10 ml portions
with coliforms present or no more than 20 percent of the samples with all five of the 100 ml
portions positive. Two of five 10 ml portions positive is an MPN of 5.1 coliforms per 100 ml and
four of five 100 ml portions is an MPN of 1.6 per IOU ml.
A statement of the statistical problem is that some fraction, P, of the water passing through a
distribution system In some period of time has coliform bacterIa present at a density of greater
than I/V where V Is the volume of wóter select d for examination. Some number, ii, of samples Is
collected and p samples are found to hove coliform bacteria present. The ratio r p/n is used as an
estimate of . The precision of this estimation procedure is considered further In the section on
Maximum Contaminant Level.
-------
13
Models of Aggregation - Lognormal Distribution. From the data in Table I it is clear that
coliform monitoring data are extremely skewed. There are several frequency distributions,
including the negative binomial, the gamma, and some exponential distributions, which can be used
for describing these types of data. The lognormol distribution has been used in water supply
practice for many decades, Is familiar to some water works personnel, and is amenable to graphical
interpretation. For those reasons, we have decided to use it in preference to any of the others. It
is well to remember, however, that this is an arbitrary selection and extrapolations based on the
lognormal distribution should be considered with some reservations.
Pipes and Christian (1982) fitted 14 different sets of coliform monitoring data to the
lognormal distribution and found geometric means ranging from .4 x lO_6 per 100 ml to 2.1 x 10 I
per 100 ml and geometric standard deviations ranging from 10 to 441.
If coliform counts fit a lognormal distribution, the value of the arithmetic mean density ( )
of Indicator bacteria In the water can be calculated from
In i= In Pg+Y2( Yg) 2
where hg is the geometric mean and Is the geometric standard devision. Also the variance (a 2 )
of the density can be calculated from
lnci 2 = 1n2 1 +(lnag 2 +fl [ xP(lflag) 2 _O)
Some calculated examples to show these relationships numerically are presented in Table 2.
Several observations may be made about the lognormal distribution. First, it is the
arithmetic mean which Is related to the number of coliforms present in some large volume of water
passing through the system not the geometric mean. The geometric mean and geometric standard
deviation are measures of how the coilform bacteria are dispersed in the water but not measures of
their total numbers. Thus, there Is no intent to suggest using the geometric mean in any
regulation.
In all cases in Table 2, when the arithmetic mean Is close to I per 100 ml the variance Is very
large, 1O 3 or greater. To assure 95 percent reliability, the sampling criterion used for the Poisson
case was that enough samples should be collected so that when the sample mean was about I, the
true mean density would be <2. If this same criterion is applied in this case, the number of samples
needed would run Into the thousands. Even with thousands of samples, the problem with truncation
of the sample counts at 80/100 ml for the MF technique and lower values for the FT method would
-------
(4
Table 2. Arithmetic means and variances for lognormal distributions for different
values of the geometric mean and geometric standard deviotion*
Geometric
Mean
Geometric
Standard
Deviation
Arithmetic
mean
(per 100 ml)
Variance
(per 100 ml)
2
Expected Sampling Results
Fraction of
I < x
Samples
80 < x
l0
(0
100
1,000
IA2xlO 5
4.03 x 10
2.3x $0
0.081
5.3 x 10
5.5x (0
<0.0001
0.00(5
0.033
—
<0.0001
0.004
io
(0
tOO
I ,000
1.42 x
4.03 x I0
2.3 x
0.81
5.3 x ,o
5.5 x ,o36
<0.0001
0.0056
0.045
—
0.0003
0 .009
l0
$0
(00
1,000
I.42x I0
4.036
2.3x 10
9.1
5.3 x IO
5.5x 10
<0.0001
0.0202
0.089
—
0.0023
0.023
l0
10
100
1,000
(.42x _2
4.037
2.3x $0
8.1
53 x lO
5.5x tO
0.00(2
0.0612
0.167
<0.0001
0.0062
0.049
io_2
(0
$00
1,000
1.42 x I0
4.03 x I0
2.3 x 10
8$ x io2
5,3 lot’
5.5 x I0
0.0231
0.167
0.223
0.0001
0.033
0.092
I0
10
(00
1,000
1.423
4.03 x (09
2.3x (0
8.1 x
5.3 x oI7
5.5x IO 39
0. 167
0.304
0.363
0.002
0.073
0.145
*ln examining data from nine small distribution systems, Pipes and Christian ($982) found
geometric means ranging from I c r 6 to I 0 and geometric standard deviations ranging from $0 to
440.
-------
15
prevent precise estimation of the true mean. Thus, the attempt to estimate the mean density of
coliform bacteria In the water flowing through the distribution system is subject to such large
errors that the result has very little meaning.
The probability of obtaining a zero count for any 100 ml sample Is always high as can be seen
by subtracting from I the values in the fifth column of Table 2. The probability of obtaining a
count greater than 80 per 100 ml is very small, except for systems In whlch’the arithmetic mean
density is much greater than I.
Consider further two water systems, one In which the coilform bacteria density has Pg
and = 100 (p 14.03) and the other with a col Iform bacteria density having = 10- and
= 15 (p = 3.91). Tables 3 and 4 give expected sampling results for the two systems. There are,
of course, many other sets of sampling results which would fit these lognormal distributions but the
ones given are likely results. The arithmetic sample means were calculated on the assumption that
the counts In each sample were the highest possible counts for the range to which the samples were
assigned; e.g., counts in the range 1-2 were assumed to be 2, counts In the range 9-16 were
assumed to be 16, and so forth,
The number of co llform bacteria in a mIllion gallons of water from the first system Is 1.8 x
I0 and In a mIllion gallons of water from the second system Is 1.5 x io8. From the data given In
Tables I and 2, it Is extremely unlikely that a violation of the MCL would be detected from
sampling data of the first system, while there Is a good chance of detection in the second system
with some reasonable numbór of samples. The largest difference in the two lies In the samples
which have 80 indicator bacteria In 100 ml. In the first system these counts in some samples
would have to be much, much greater than 80 whereas In the second system they could be only
somewhat greater than 80 In most cases.
The lognormal distribution gives the following insights Into the problems of sampling a water
distribution system; they are as follows:
I. The probability of a sampl. mean of >1/100 ml Is !2Q very strongly related to the total
numbers of Indicator bacteria in a water passing through the system.
2. It Is expected that only a small fraction of the 100 ml samples examIned will have
IndIcator bacteria present unless the water in the system is very grossly contaminated.
3. The IndIcator bacteria counts in some of the samples are expected to be much greater
than con be measured by the presently used technique.
-------
Tthle 3. Expected sampling results which fit a lognormal distribution with p = I0 and a 100
(mean coliform density = k.03 per 100 ml). g g
Number
Range of Counts in 100 ml Samples
of
Maximum
Samples
x< I
I 0.328
-------
Tthle4. Expecteds ling results which fit olognormal distribution with p = I0 and a = 15
(mecxi coliform density = 3.91 per 100 ml) g g
Niznber
.
R ge of Coixits in 100 ml Somples
of
Maximum
Samples
x c i
lcxc2
22.96
1,000
808
64
48 30 21 14 7
2
6
>2.04
-J
-------
18
4. The arithmetic sample mean will, in most cases, be much less than the actual
arithmetic mean density in the water.
Chlorine crid Turbidily
Chlorine residual and turbidity are continuously variable parameters which are related to the
microbiological quality of drinking water. Frequency distributions which would describe the
variability of these parameters in water distribution systems would probably be different from
those suitable for describing the dispersion of bacteria in the system. However, this question is not
pursued further herein because of the reasons given below.
Resithal Chlorine. If measurements of residual chlorine are substituted for some of the
coliform bacteria samples, there is little to be gained by sampling for chlorine at intermediate
points around the distribution system. It is well known that the concentration of residual chlorine
in water in a distribution system decreases as a function of time. The lowest chlorine residuals are
to be found in the water which has been in the distribution system the longest. Thus, sampling for
residual chlorine concentration should be related to the configuration and hydraulics of the
distribution system rather than to a frequency distribution.
One point which should be made about substitution of free chlorine measurements for total
coliform counts (which is permitted in the present regulations) is that the presence of free chlorine
in the water at the time of sampling is noteqoivalent to the absence of coliform bacteria. Table 5
gives an example of this using the WH data of 614/81 (Pipes and Christian, 1982). The data of
Table 5 differ from those of Table I because coliform count from two 100 ml replicates collected
at each site sampled are included in Table 5. These data show no relationship between free
chlorine and the occurrence of coliforms. Many other examples could be given (Goshko, et al.,
1980).
Turbidity. The control of turbidity in a water distribution system is a problem which is
different from the control of microbiological contamination of the water. An increase in the
turbidity of water in a distribution system may be related to microbiological contamination or it
may be unrelated. Thus, turbidity is not a direct measure of the microbiological quality of drinking
water. However, a high turbidity can interfere with microbiological examination techniques. It
would be quite reasonable to make turbidity measurements on the samples collected for
microbiological examination.
-------
19
Table 5. Free chlorine measurements and confirmed
collform counts WH data of 6/ 4/8 I.
Free Residual
Number
Number
Fraction
Confirmed
Chlorine
of
of
Positive
Coliform
Concentration
(mg/I)
Samples
Samples
with
Coliforms
.
Counts
0
18
2
0.22
I, I
0.1
20
I
0.1
42
0.15
2
0
0.2
8
1
0.25
I
0.25
2
0
0.3
4
0
0.35
4
0
0.40
16
4
0.5
I, I, 3, 15
0.45
6
I
0.33
2
0.55
2
0
-------
20
Selection of Sampling Locations
The present regulations (EPA, 1976) state that samples should be taken at points which are
representative of conditions within the distribution system. We assume that this means throughout
the distribution system and that sampling locations should be selected so that all parts of the
system are sampled at some time or another.
Distribution System Elements. A distribution system is composed of pipes of various sizes,
finished water storage reservoirs (which usually are elevated), pumping stations, service connec-
tions, and fire hydrants. The pipes can be classified as transmission mains, distribution mains, and
street laterals.
Transmission mains are usually the largest pipes in the system and serve to convey water
from the source (treatment plant or well) to various sections of the distribution system. If
elevated storage Is provided, it usually is connected to a transmission main on the side of the
community away from the source of water. Reversal of flow direction is common In a transmission
main unless all of the finished water storage Is adjacent to the water source.
Distribution mains serve to convey water from a transmission main to one or more Isolated
sections of the distribution system. The direcfion of flow in a distribution main is usually not
reversed unless there Is some unusual condition such as a high water demand for fighting a fire.
Street laterals convey water from distribution mains to individual service connections and to
other street laterals. Street laterals which serve other street laterals are called central street
laterals. Street laterals which serve only service connections In a single city block are called
peripheral street laterals. Peripheral street 0 laterals which stop at the end of a block are called
dead end street laterals. Peripheral street laterals may occur in any port of the distribution
system, not Just around th. edges.
A water distribution system is often divided into sections whIch are more or less Isolated
from each other by the flow of water from the transmission main to the services where the water
Is used. Figure I gives, in diagrammatic form, four different simple confIgurations or geometries
whIch are used for isolated sections of distribution systems; namely, linear, loop, dendritic, and
grid. The dendritic configuration is characterized by a great many dead end street laterals while a
grid configuration has all of the street laterals connected to ether. Isolated sections of a water
-------
21
SECTION A GRID
31 SAMPLING LOCATIONS
DISTRIBUTION MANIFOLD
STREET
LATERAL
F 1GURE 1.
SECTION D DENDRITIC
9 SAMPLING LOCATIONS
Diagram of simple configuration for isolated secflons of a water distdbutLon
system.
WATER SOURCE
SSION MAIN
SECTION B LINEAR
I SAMPLING LOCATION
III II I I I I III
Li 1 1 1L
II I II I Fl I
SECTION C LOOP
4SAMPLING LOCATIONS
DISTRIBUTION MAIN
-------
22
distribution system may have a compound configuration such as loop-dendritic. in Figure I, the
term “sampling location” is used to refer to a street lateral between intersections with other street
laterals.
The significance of isolated sections of a distribution system for monitoring Is that microbial
contamination existing in one section is not detectable in samples collected in other sections. For
instance, if microbial contamination occurred in Section A of Figure I, it could only be detected by
collecting samples from Section A and not by sampling from Sections B, C, or D.
Representative Sampling Locations. It is clear that it is impossible to obtain representation
of the microbiological water quality conditions throughout even a small water distribution system
with only a few samples. Consider the idealized distribution system diagrammed in Figure I and
assume that the water entering the transmission main at the source is treated and disinfected so
that no coliform bacteria are introduced into the system. Coliform bacteria then will be found
only when they are introduced at some point in the distribution system. Each service connection,
each joint, and any breaks in the pipes will be a point for possible Introduction of contamination.
So will any distribution system storage where there Is an air-water Interface. It would be Imprac-
tical to attempt to sample intensively enough to assure that microbiological contamination never
enters the system.
Under normal pressure and flow conditions in a well-designed water distribution system,
leakage will be Out of the system. However, when low pressure occurs, there may be back flow
into the system and contaminated groundwater and/or sewage may enter and mix with the treated
water. The factors which may contribute to low pressure conditions Include high water demand,
reduction in pipe diameter due to scale formation, increase In pipe friction due to corrosion or
deposits of sediment in the pipes, and variations In the topography of the area served.
If low presiure occurs, the number of possible points of entry of contaminated water Into the
system will be a function of the total length of pipe in the ground because the number of joints,
connections, etc. is more or less proportional to the length of pipe. It Is not known how far
contamination will persist from the point of entry Into the system. The distance of persistence wilt
be a function of the volume of contaminated water entering the system, the flow In the pipe, the
bacterial density in the contaminated water, and the disinfectant residual In the water flowing In
the pipe. The microbiological contaminants may be inactivated by the disinfectant residual or they
-------
23
may settle out and be incorporated in the sediment in the bottom of the pipe. The disinfectant
residual may be expected to be the greatest in areas of high flow and settling will occur most
rapidly in areas of low flow. in either case the contamination is not expected to persist for a very
long distance in the system. The contamination would be detected only in the general vicinity of
the point of entry.
Special Sampling Problems. In addition to monitoring based on samples collected at service
connections, there are locations and phenomena which are especially likely to be sources of
microbiological contamination. These situations require sampling in addition to the routine
monitoring of the distribution system. Such sampling should be undertaken when coliforms have
been detected in the distribution system and on Investigation to determine the source of the
contamination is undertaken.
Finished water stored In reservoirs, standpipes, and elevated tanks has a free water surface in
contact with the atmosphere. Open reservoirs which were once widely used for finished water
storage have been largely eliminated, except for the very large water systems, because of the large
number of water quality problems which they Introduced. Even covered distribution system storage
devices are subject to contamination by microorganisms in the dust in the air, if the screen
covering the air vent is rusted or loose (which is frequently the case), birds and/or rodents may gain
access to the stored water and introduce microorganisms. It may be necessary to chlorinate the
water from the storage device for disinfection and to restore the chlorine residual if it has
dissipated.
During installation of a new main, contamination may be introduced. New mains should, of
course, be disinfected before being placed Into service. It is necessary to flush the new main to
remove the disinfectant. After flushing, the new main should be sampled and the samples
examined for coliform bacteria. Also, placing a new main In service will change flow patterns in
some part of the distribution system which may resuspend sediment from the bottom of pipes or
shear microbial growths from the walls of the pipes, resulting in an unusually high number of
positive coilform results.
Periods of high water usage ore times when there are high flow velocities In the distribution
system and possibly reversal of direction of flow in some mains and street laterals. These velocity
changes may resuspend sediment from the bottom of pipes or shear microbial growths from the
-------
24
walls of the pipes. Thus, extra samples should be collected during periods of high water usage to
determine if there is continuing contamination of the system which Is not detected by routine
monitoring.
MAXIMUM CONTAMINANT LEVEL
The MCL’s for microbiological contamination of a water distribution system In the present
regulations (EPA, I 976) are very complex and therefore subject to misinterpretation. The
Monitoring Panel recommends that it be simplified. In addition, specific recommendations are
made in this secflon for a new MCL which can be Interpreted In terms of the extent of
microbiological contamination of the water flowing through a distribution system.
Coilform Bacteria
The MCL’s for coliform bacteria vary depending on the number of samples collected and the
examination method used. The required number of samples per month Is specified In the
regulations (EPA, 1976) in tabular form related to the total population served by the water system.
It appears that the MCL’s are intended to limit the mean coliform density In the water flowing
through the system and the fraction of the samples with a coliform density greater than
approximately 5 per 100 ml. In this section the microbiological MCL’s are broken down Into their
component parts and each part is discussed separately.
Table 6 gives some rules abstrcicted from the present regulations (EPA, 1976). The
presentation of the rules in this form Is not intended to explain fully how they are used but as a
basis for discussing certain features. It is Impossible to understand what these rules mean without
considering the examination technique, the density limit, the statistical parameter calculated, the
reporting period, and the number of samples collected. It Is misleading to dIscuss only one element
of the rules. Thus, no part of the following discussion Is independent of the rest.
Some number, n, of samples is collected over a water quality evaluation period, Q, which may
be divided into r shorter reporting periods, R; thus, rR = Q. For most of the present regulations R
Q = one month, but other values for Q could be selected. Q should be some period of time over
which the reliability of protection of the drinking water does not change. R Is some period of time
shorter than 0 selected for convenience of reporting and so that if the MCL Is violated, remedial
action can be taken In a reasonable period of time.
-------
Tthle 6. Micmbioiogical MCL’s of the 1976 National Interim Prirncry Drinking Water Regulations
[ Abstract aid Interpretations of Certain Secticns]*
(I)
Section
of
Regulation
a) (3)
Method Coliform
Density
Limit
( d v)
(4)
Statistical
Paometer
(x,x or r)
(5)
Water
Quality
Evaluation
Period (Q)
(6)
Reporting
Period
Sec. 141.31
(R)
(1)
Minimum Number
of Samples per
Month
Sec. 141 .21
(n)
141.14
(aX I)
1/100 ml
aithmetjc
meori
one
month
one month
(40 days)
4-500
141.14
(0X2)
MF 5 ,100 ml
single
soniple
one
month
one month
(40 days)
4-19
141.14
(aX3)
5/lOOmI
5%of
s si les
one
month
one month
(40 days)
20-500
141.14
(bX lXi)
FT all
(10 ml MPN =
portions) 1.05/ 100 ml
overall
meor i
one
month
one month
(40 days)
4-500
141.14
OXIX1 I)
141.14
(bXlXiii)
FT swnple
(lOrni M’N=
portions) 5 . 1/lOOmI
FT sample
(10 ml M’N =
portions) 5. 1/lOOrni
single
sa iple
5%of
samples
one
month
one
month
one month
(40 days)
onemonth
(40 days)
4-19
20-500
141.14
(bX2X I)
FT overall
(100 ml MPN =
portions) 0.92/100 ml
overall
meon
one
month
one month
(40 days)
4-500
141.14
OX2X1 I)
FT sample
(‘00 ml MPN>
portions) 2.3/00 ml
single
sample
one
month
one month
(40 days)
4
141.14
O X2Xiii)
FT suiip le
(100 ml MPN>
pcrtiais) 141100 ml
20% of
samples
one
month
one month
(40 days)
5-500
141.14
(c)
all limitsaidpcraneters
iI st ed thove
•
three
months
40duys
1-3
*One re$ Jai is that if c/V is e,cceeded using i or if r is eicceeded using c/V then the MCL is violated but the National Interim Primory
Driu*Iig Water Regulations (Envirurwnental Protection Agency, 1976) thould be consulted for f Il expknation.
I.J
U’
-------
26
Statistical Parameter. At the end of an evaluation period or a reporting period, some
statistical parameter of the sample examination results is calculated. In theory, this parameter
could be almost anything but in practice if should be easy to calculate and should have some
intuitive significance to the operator of the water system. At the present time the parameters
used include the coliform count of a single sample, x, the arithmetic mean of the sample counts, c,
and the fraction, r, of the samples with a count greater than some preselected value, For the MF
method the parameters are I) the arithmetic mean of the sample counts, and 2) the fraction of the
counts exceeding a count of 4. For the FT technique the parameters are I) the overall fraction of
the portions which give positive reactions and 2) the fraction of the samples which have three or
more of five 10 ml portions positive or all five 100 ml portions positive.
The Monitoring Panel finds no reason to continue using the arithmetic mean of the MF counts
as a statistical parameter for the MCL (Section l4l. 14(aXl)). To do so would Infer that there Is
some health significance associated with the mean density of Indicator bacteria In the water and
that It is possible to estimate the mean density within reasonable limits with a reasonable number
of samples. The estimate of the mean density of indicator bacteria in the water passing through
the system is subject to very large errors because of the great variability of Indicator bacteria
densities and data truncation at the densities either too large or too small to be measured by the
techniques now in use.
Similarly, the Monitoring F 5 aneI finds no reason to continue the use of the fraction of all
portions positive by the FT technique as a statistical parameter for the MCL (Sections
ll&l.14(b)(IXi) and (b)(2X1)). The fraction of the portions from a single sample which give positIve
results can be used to make an MPN estimate of the density of that sample so long as the sample Is
well mixed before the portions are removed. However, It Is not logical to use the overall fraction
of all portions of all samples examined as a measure of the mean density of the water flowing
through the distribution system In a month’s time unless the coliform densities follow a Poisson
disfribution or If all portions of any sample are Indeterminate. There are methods to obtain an
overage of MPN estimates but none of them Is simple enough for use in a routine monitoring
program (Thomas, 1952).
The Monitoring Panel recommends that the parameter for evaluation of the microbiological
quality of drinking water be the fraction of the samples which have coliform bacteria present. This
-------
27
parameter can be related in a logical way to the reliability of protection of the system: it is an
easy parameter to calculate; its use could lead to a simplified method for laboratory examination
of samples; and, its use highlights the need for a minimum number of samples which is much higher
than that now required for small systems. The determination of the presence of coliforms in an
individual sample is subject to considerably less error than are attempts to measure coliform
density.
Examination Method. If the arithmetic mean sample density is eliminated as a statistical
parameter of the MCL and the only parameter used Is the fraction of the samples which have
coliform bacteria present, then there is no need for an examination method which will give an
estimate of the number of coliform bacteria in the sample. The only need is for a method which
will demonstrate whether coliform bacteria are present or absent. In this case, It seems reasonable
to recommend that serious consideration be given to a test which could be used to detect the
presence of a variety of indicator bacteria, not just coliforms,
Sample Volume. At the present time, the standard volume for the coliform test is 100 ml for
the MF procedure and either five 10 ml portions or five 100 ml portions for the FT technique.
Collection of a sample of more than 500 ml to give five 100 ml portions for the FT technique is
awkward and this option is little used. Collection of a sample of slightly more than 100 ml is
convenient for handling and such samples are sometimes sent through the mail. Interferences with
the coliform test may occur due to high turbidity or a high density of non-coliform bacteria
(Seidler, 1981). These Interferences are manageable with the present sample volume but would be
more frequent with larger sample volumes.
The Monitoring Panel recommends no change In the sample volumes required under the
current regulations. However, since the recommendation for a statistical parameter for the MCL
is the fraction of samples with coliforms present, there is no need to divide the volume examined
by the FT technique Into five portions. If the MF procedure Is used, the 100 ml sample volume
should be continued and the equivalent density level to distinguish between contaminated and
uncontaminated water would be 2/100 ml.
Riportlng Period. At present the reporting period is one month with allowance of ten extra
days to prepare and submit the report. This is a reasonable period for administration purposes if
-------
28
the sampling results are negative. However, if indicator bacteria ore found in a sufficient number
of samples to cause a violation of the MCL, this should be reported to the State immediately and
remedial action should be token in a more timely manner.
The Monitoring Panel recommends that the reporting period of one month be retained, as in
the present regulations, with an additional provision that if the results of the examination of any
samples indicate a violation of the MCL these results be reported within 48 hours after the
laboratory examination results are obtained. For example, if a water utility is required to collect
ten samples per month and three positive samples will cause a violation of the MCL, the State
should be notified within 48 hours of ter the third positive result is obtained.
Water Quality Evaluation Period. There Is no good evidence to show how long the reliability
of protection of any water system against microbial contamination persists. The Monitoring Panel
recommends that research be undertaken to determine what identifiable occurrences would change
this reliability of protection and, in the absence of such occurrences, how long a period of time the
degree of reliability of protection will remain constant.
The Monitoring Panel recommends that a water quality evaluation period of 2 consecutive
months be adopted until such time that information about the persistence of reliability of
protection against microbial contamination can be collected and evaluated. This recommendation
follows the precedent of evaluating water supplies over a year’s time under the Public Health
Service Drinking Water Standards (Public Health Service, 1962). However, It differs from that
precedent In that each consecutive (2-month period rather than each calendar year Is a water
quality evaluation period.
MCL. The basic standard recommended Is the fraction of all dIstribution system samples
examined with collforms present shall not exceed 0.05 with enough samples examined to give
95 percent confidence that the fraction of the water with coliforms present Is less than 0.10. ThIs
recommended standard differs In concept from the present regulations.
The key to the relationship between the fraction of the samples with coliforms present and
the fraction of the water actually contaminated Is the number of samples examined. if n samples
are collected and p samples hove co liforms present, the ratio, r = p/n, Is used as an estimate of the
actual fraction, p, of the water which Is contaminated. We expect that for any water distribution
-------
29
system, p will be greater than zero and less than one. If n = I, then r can be either 0 to I; if n 2, r
can be 0, 0.5, or I; if n = 3, r can be 0, 0.33, 0.67, or I and so forth. If enough samples are
examined, r is expected to be very close to p, but if only a few samples are examined, the values
which r can have are limited to only a few and r may be quite different from p.
The ratio, r, has a binomial distribution with mean p and variance (I -c /n. If n is greater than
30, the binomial distribution can be approximated by the normal distribution and a 95 percent
confidence interval (one sided) is given by
r + l.645 ’r(l—r)/n
Our recommended standard is that r + l.6L,5Vr(l_r)/n should be less than 0.1 when r is less than or
equal to 0.05. The smallest convenient value for n Is 60 (Table 7), since the panel examined the
question of how many positive samples can be allowed without exceeding the MCL. If n = 60 and p
= 3, then r = 0.05 and the 95 percent confidence interval extends up to 0.098.
For larger distribution systems from which 60 or more samples per month are collected and
examined, there Is no problem In applying the basic standard recommended on a month by month
basis. Smaller systems may experience difficulty with the cost of collecting and examining so
many samples.
A water quality evaluation period of 12 consecutive months is recommended. It is possible
that smaller systems, which now collect and examine one or two samples per month could assume
the cost of examining five samples per month to give a total of 60 samples each 12 months. If
three or less of the 60 samples were positive, the standard would not be exceeded. If four or more
of the samples were positive in any 12-month period, this would be reported to the regulatory
agency at the end of the reporting period (month) in which the fourth posItive sample was obtained.
Corrective action would be instituted at that time.
There is a problem with applying an MCL on the basisof the results from samples collected
over a 12-month period. Violation of an MCL requires notification of the public that the drinking
water may be unsafe, It would be difficult to require public notification based on four samples
with coliforms present distributed over a 12-month period. Therefore, the Monitoring Panel
recommends that the basic standard be established as a protection reliability standard (PRS). If
the PRS Is exceeded, investigation of the cause of the violation should be required and correction
action Instituted.
-------
30
Table 7. Ratios of Positive Samples to Total Samples and 95% Confidence
Interval for Fraction of Water Contaminated (Small Systems).
I 0.20 0.657
0.167 0.582
I Q•f43 0.521
1 0.125 0.471
I 0.111 0.429
I 0.10 0.394
I 0.09) 0.364
I 0.083 0.339
I 0.077 0.3 16
I 0.07 I 0.297
I 0.067 0.279
I 0.062 0.2611
I 0.059 0.250
I 0.055 0.238
I 0.053 0.226
I 0.050 0.2)6
I 0.048 0.207
I 0.045 0.198
I 0.043 0.190
I 0.042 0.183
I 0.040 0.176
I 0.038 0.170
I 0.037 0.164
I 0.036 0.159
I 0.034 0.153
I 0.033 0.149
I 0.032 0,1115
I 0.03) 0.140
2 0.40 0.811
2 0.333 0.729
2 0.286 0.659
2 0.250 0.600
2 0.222 0.550
2 0.20 0.507
2 0.181 0.470
2 0.167 0.438
2 0. 154 0.4)0
2 0.143 0.385
2 0.133 0.363
2 0.125 0.344
2 0.118 0.326
2 0. 1 11 0.310
2 0.105 0.296
2 0.10 0.283
2 0.095 0.271
2 0.091 0.259
2 0.087 0.249
2 0.083 0.240
2 0.080 0.231
2 0.077 0.223
2 0.074 0.2)5
2 0.071 0.208
2 0.069 0.202
2 0.067 0.195
2 0.065 0.188
2 0.062 0.181
1 0.030 0.135
I 0.029 0.131
I 0.029 0.126
I 0.028 0.121
I 0.027 0.116
I 0.026 0.111
I 0.026 0.107
I 0.025 0.103
I 0.024 0099
I t .023 0.095
I 0.023 0.091
I 0.023 0.087
I 0.022 0.083
1 0.022 0.079
I 0.02) 0.075
I 0.021 0.071
2 0.041 0.099
2 0.040 0.097
2 0.039 0.074
2 0.038 0.091
2 0.038 0.088
2 0.037 0.085
2 0.036 0.082
2 0.036 0.080
2 0.035 0.078
2 0.034 0.076
3 0.05) 0.099
3 0.050 0.098
2 0.061 0.174
2 0.059 0.167
2 0.057 0.161
2 0.056 0.155
2 0.054 0.149
2 0.053 0.143
2 0.051 0.137
2 0.050 0.131
2 0.049 0.127
2 0.048 0.123
2 0.047 0.119
2 0.045 0.115
2 0.044 0.112
2 0.043 0.109
2 0.043 0.106
2 0.042 0.103
3 0.061 0.119
3 0.060 0.117
3 0.059 0.115
3 0.058 0.113
3 0.057 0.111
3 0.056 0.109
3 0.055 0.107
3 0.054 0.105
3 0.053 0.103
3 0.052 0.101
4 0.068 0.123
4 0.067 0.121
n = number of samples, p number of positive samples, r =p/n, and I upper limit of 95%
confidence Interval for fraction of water contaminonted. U
-------
31
Since the basic proposed standard of no more than 5 percent of the samples examined with
coliforms present with enough samples examined to give 95 percent confidence that less than
10 percent of the water has coliforms present is recommended as a protection reliability standard,
there is a need for definition of an MCL which could be applied on a month by month basis. Since
the smallest number of samples per month recommended is 5 and we want to allow one sample with
coliforms present in any given month without triggering public notification, the Monitoring Panel
recommends that the MCL be coliforms present in no more than 20 percent of the samples
collected.
Larger water distribution systems will have a greater number of places where microbiological
contamination can enter the system. Under the present regulations (EPA, I 976), the number of
samples per month required to meet the MCL is specified in tabular form as function of the
population served. Population served may be used as a surrogate for other size factors such as area
or total length of pipe in the distribution system. In the absence of specific information about how
the population served is related to other size factors, the Monitoring Panel recommends that the
present sampling frequency table be retained with the modification to make five the minimum
number of samples collected per month for the smallest systems (population served less than
4,900).
Number of Samples. The Monitoring Panel recommends that the minimum number of samples
collected from the smallest water distribution systems be 60 per each consecutive 12-month period
or five per month. This number of samples Is Intended to accompany the criterion that It Is
desirable to have 95 percent confidence that less than 10 percent of the water passing through the
distribution system In any 12-month period is contaminated with no more than 5 percent of the
samples having coliforms present.
It Is recommended that when ten or fewer samples per month are collected from a system,
they all be collected on the same day. This will give a synoptic sampling of the area covered by
the system. it assumes that the water entering the system is adequately dlslnfected and that
monitoring of the distribution system Is on attempt to locate areas of contamination. The synoptic
sampling of the system will give a higher probability of locating areas of contamination, It is
further recommended that if ten or less samples per month are collected, the day they are
collected be as early In the month as feasible. This will allow for check sampling and evaluation of
-------
32
the results of the check samples in a timely manner. It is recommended that check samples be
collected for any positive samples and that there be at least three check samples for each positive
sample found.
Remedial Action. Violation of the MCL requires public notification. Under the present
regulations (EPA, 1976) no further remedial action is specified. Apparently, it is assumed that the
embarrassment of public notification will cause the water utility to take action to remedy the
situation. It might be helpful if other types of remedial action were specified.
Check sampling is not, in itself, a remedial action. Check sampling is, however, a start on an
investigation of the possible sources of the contamination, f the check samples are collected
during the same reporting period as the original positive samples which could cause a violation of
the proposed MCL and the check samples are negative, the additional samples will reduce the
fraction of samples with coliforms present so that the MCL is not violated. However, if violat ion
of the MCL is avoided only by negative results in check samples more than three times in any
consecutive 12-month period, it is almost certain that the proposed PRS will be exceeded.
Short-term remedial action should include increasing disinfectant dosage and flushing of
various parts of the distribution system to obtain more effective use of the disinfectant. Flushing
of the distribution system may dislodge coliforms from wall growth or from sediment in the bottom
of the pipes and result in an even higher fraction of the samples examined having coliforms
present. Such an increase in the positive samples should be temporary and as the system is cleaned
out the sampling results should improve.
Violation of the PRS may require longer term remedial action. This might include
development of new sources of supply, modifications of treatment, disinfection of water in storage
tanks connected to the distribution system, detection and sealing of leaks in water mains, and
construction of new additions to the distribution system. A plan for implementation of such longer
term remedial action should be worked out by the water utility in consultation with the State.
chlorine Substitution
Monitoring for residual chlorine concentration is not a direct substitute for monitoring for
coliform bacteria because the presence of free chlorine residual is not equivalent to the absence of
coliforms (Table 5). Monitoring for free chlorine residual might provide a better measure of the
-------
33
safety of individual portions of the water if the free chlorine Is present every day In all parts of the
system. However, this has not been demonstrated In field studies. The Monitoring Panel
recommends retention of the chlorine substitution option and a field study to determine the
effectiveness of free chlorine monitoring In protecting against waterborne disease.
Turbidily
Some of the materials which cause turbidity also affect health but some do not. High
turbidities can interfere with the coliform test and can protect microorganisms from the action of
disinfectants. Relatively low turbidity in the water entering the distribution system is attainable
at reasonable cost. The Monitoring Panel recommends that the present MCL of I TU at the point
of entry to the distribution system be retained but that the monitoring frequency at filtration
plants be Increased to once every eight hours (once per shift) unless continuous monitoring is
provided. Three tests should be averaged to give the daily result. Where cOntinuous recording Is
provided, the daily result should be the average of the record. In addition, the Monitoring Panel
suggests that when the water is filtered, a treatment objective of a filtered water turbidity of
0.2 TU should be established. Each filter should be monitored for possible breakthrough, especially
at locations where plants operate unattended for portions of each day.
MONITORiNG GUIDELINES
Guidelines are different from MCL’s in that exceeding a guideline value does not constitute a
violation of the regulations. Guidelines are valuable in explaining MCL’s and for establishing goals.
MICTObIOIO 9 ICOI Monitoring During Treatment
With the present availability and low cost of water tróotment processes and equipment there
is no excuse for microbiologically contaminated water to be pumped Into a water distribution
system. It is not enough to assume that the treatment processes will eliminate microbiological
contamination; this should be demonstrated by examination of samples. Of particular concern are
Interruptions of normal treatment processes and construction of new treatment facilities.
Finished Water Monitoring. Table 8 presents a small amount of data obtained from seven
different water treatment plants. From these data it is clear that coliform bacteria can
sometimes be found In treated water. In practically all of the few samples inwhich coliforms were
-------
Table 8. Coliform monitoring results on raw and finished water at seven different treatment plants
System Water
Type of
SamplIng
Raw
Water
Finish
ed Water
Source
Treatment
Period
Number of
Samples
Fraction
Positive
Number of
Samples
Fraction
Positive
CV Surface Coagulation Jan.-Feb. 1979 tO 1.0 77 1)
Settling
Filtration
Chlorination
Well Chlorination April-May 1979 95 0.011 90 0.033
May-June 1981 29 0 58 0
Well Aeration June-July 1979 39 0 .35 0.029
Filtration
Chlorination
Well Aeration June-July 1979 43 0.023 35 0
Filtration
Chlorination
Surface Coagulation Jon.-Feb. 1980 26 1.0 195 0.0 10
Settling
Filtration
Chlorination
BL Well Aeration March-April 1980 35 0.029 95 0
Filtration June 1981 25 0 30 0.033
Chlorination
Well Chlorination Sepf.-Dec. 1981 0 110 0.02
-------
35
found, the count was I per 100 ml or less; the less resulting from some samples from which more
than 100 ml was examined. In all of these treatment plants except BL, there was a free chlorine
residual in all finished water samples examined.
The turbidity of finished water should be less than I TU. With the turbidity this low it is
possible to filter 500 ml of water through a membrane filter and not get a layer of particulate
matter on the filter thick enough to interfere with the development of colonies of indicator
bacteria. Therefore, we recommend, for monitoring water entering the distribution system,
examination of 500 ml samples by the MF procedure or five 100 ml portions by the FT technique.
We recommend that for a period of one year, daily sampling of treated water for coliform
bacteria be performed at a point immediately before it enters the distribution system. If 360
samples are negative, this would provide more than 95 percent confidence that less than I percent
of the 500 ml portions of the water entering the distribution system had one or more indicator
bacteria present. This monitoring requirement would be applied to each source of finished water to
the distribution system. If records show any interruptions or failures of treatment, the one-year
monitoring period should be repeated. After the initial one-year monitoring period, if no indicator
bacteria are found, the sampling frequency could be reduced to one sample per week.
Water Source Considerations. Surface waters are much more likely to have microbiological
contamination than groundwaters. Direct reuse water; i.e., water taken directly from a
wastewater treatment plant Into the water treatment plant, can be expected to be contaminated in
all cases. However, groundwater can also be contaminated and there is the need to demonstrate
that finished water from any source is not microblologically contaminated as it enters the
distribution system. Thus, we recommend that the finished water monitoring requirements be
applied also to those systems which use a groundwater source. Additional monitoring may be
required for direct reuse water or surface waters during periods of unusually severe source
contamination.
Monitorirç of Treatment processes. It Is expected that indicator bacteria densities will be
reduced by each treatment process. If the finished water Is monitored, additional monitoring at
Intermediate stages durIng treatment Is not needed. The exception to this would be treatment
activities which might result in microbiological contamination of the finished water.
-------
36
Some microorganisms are removed during filtration and it may be assumed that a filter which
has been running for some time will contain some indicator bacteria. Bockwashing of the filter
should remove all of the particulate matter accumulated in the filter including the microorganisms.
However, It Is possible that the water In the filter Immediately following backwashing will have
some microorganisms present. Some treatment plants filter to waste for a short period
Immediately following backwashing to avoid introducing turbidity and microorganisms Into the
finished water. If this is not done, there should be monitoring of the filtered water Immediately
following backwashlng.
Monitoring During Treatment Modifications. Construction activities are disruptive of normal
operations. Daily monitoring of the finished water as recommended above should be Instituted at
any time there Is construction at the water treatment plant. The daily monitoring should continue
for one year after the completion of construction In order to demonstrate that the new or altered
treatment processes provide adequate protection against microbiological contamination.
Monitoring Stored Water. Elevated storage tanks and standplpes connected to the distri-
bution system are used to maintain pressure during periods of high water demand. The residence
time of water In these storage devices may be much longer than the residence time In the
distribution system Itself. There is a free air-water interface in the tank or standplpe. This allows
for dissipation of chlorine residUal, introduction of microorganisms from dust In the air or from
birds, and reproduction of bacteria in the tank or standplpe. If coliforms grow in a tank or
standpipe, there will be continual seeding of the distribution system which could interfere with the
detection of other sources of microbiological contamination. Routine monitoring of standplpes and
elevated storage tanks is less Important than synoptic sampling of the distribution system but
sampling of such storage devices should be port of any Investigation to determine sources of
contamination to the system.
Distribution System Monitoring
Seasonal Usage. Some facilities supplying water to seasonal use areas (e.g., campgrounds,
summer camps, and some vacation resorts) may be used only a portion of the year. Without the
continuity of year round operation, the reliability of protection may change from season to season.
-------
37
Thus, the water quality evaluation period for a water system used for only a portion of the year
must be less than 12 months. in order to meet the criteria described previously, the recommended
O samples per water quality evaluation period should be collected over the part of any year during
which the system is in use.
Emergency Conditions. If normal water supply conditions are disrupted due to natural
disasters or large-scale accidents, there is a greatly increased danger of epidemics of waterborne
disease. The traditional response to such conditions is to notify the consumers to boll all drinking
water before use. However, if drinking water is supplied by a government agency by tank truck,
there is a tendency to assume that the water is safe and boiling the water is unnecessary. Ideally,
such water should be tested for bacteriological safety before distribution. However, the one or
two-day period required before the test results are available makes this Impractical in many
Instances. It is probably more effective to stress the boil water advisory and explain that the
water has not been tested for microbiological safety.
COST OF MICROBIOLOGICAL MONITORING
The costs for microbiological monitoring can be allotted to the costs associated with sample
collection and transportation and the costs associated with laboratory examination. Since a
significant increase In the number of samples per month collected from small systems is
recommended, It is appropriate to consider the increase In costs for such systems. If these
recommendations are adopted, the costs of monitoring for large systems will probably be reduced
because of the slmplificatlons in laboratory examination procedures.
Collection ond Tranipuitotion
The smallest water systems usually have one operator. If It is a municIpal system, the
operator of the water system may have other responsibilities such as operating the sewage
treatment plant, reading water meters, billing, maintenance, etc. Collecting one water sample per
month would probably take up an hour of the operator’s time. A water sample can be collected In a
few minutes, most of the time required Is for drIvIng to the site of sampling and returning to the
water plOnt. Increasing the number of samples from one to five per month might increase the time
of sampling from one hour to two. The vehicle cost, operator’s time, overhead, etc. may be
estimated at’ approximately $25 per hour. Thus, the Increase in collection cost would be from $25
-------
38
per month to $50 per month. Of course, the operator’s salary, vehicle cost, overhead, etc. are
fixed costs and there would be no line item increase in the water department budget. The actual
cost would be absorbed In the loss of time devoted to other duties.
The additional check sampling requirements would represent an additional cost increase for
monitoring. Check sampling would be required only if one of the original samples were found to
contain coliforms. If it is assumed that the sampler can collect three check samples in an hour’s
time, the additional cost for check sampling for a single positive sample would be a maximum of
$75 per year. Again this expense would be absorbed in the fixed costs of operating the system.
Transportation of the samples to the laboratory is usually a separate cost item unless the
laboratory is local, If the samples are examined by a commercial laboratory, it is common for a
messenger from the laboratory to pick up the samples at the water plant. The messenger can carry
five samples as well as one and there should be no additional transportation cost in this case. If the
samples are transported by mail there would be additional mailing costs. The additional mailing
costs would be approximately from $2 to $10 per month. If check samples were required three
months in a year’s time, this would be an additional $18 per year.
Ex , J nation Costs
Information on the costs of laboratory examination of water samples for microbiological
contamination has been compiled by Geidreich and Kennedy (1982). According to their figures the
cost for an MF test was $3.32 per sample and for an FT test was $ .2 I per sample based on 1980
material and labor charges. Most of the additional cost of the FT test was ascribed to labor
because of the time required to prepare and sterilize the broth medium. Commercial laboratorIes
charge three to four times the basic cost of the tests. Thus, laboratory fees in 1980 were about
$12 per sample for the MF test and $18 to $21 1. per sample for the FT test.
If the recommendations contained herein are adopted, the usual FT coliform test procedure
will be one 50 ml portion per sample rather than five 10 ml portions. The laboratory cost for five
50 ml samples should be similar to five, 10-mi portions from the same sample so there should be no
significant increase in laboratory examination cost. There would be an Increase in sample
examination cost If the MF technique is used.
The additional check samples might also Increase the laboratory examination cost by $24 per
month for three months of the year. Thus, the total cost for a system now collecting one sample
-------
39
per month and paying approximately $144 per year for laboratory examinations might be as high as
$360 per year.
Violation of the MCL
A violation of the MCL would occur in the smallest systems where fewer than ten samples
per month are collected, if two positive samples were collected in any one month. A violation of
the PRS would occur If one positive sample were collected in any four months out of a 12-month
period. Such violations would trigger investigations to determine the source of microbiological
contamination and ultimately action to eliminate the source and improve the reliability of
protection of the system. Such investigations and actions could be quite expensive relative to the
cost of routine monitoring when a violation does not occur. Such additional costs are justified, in
our opinion, by the need to protect public health.
LITERATURE CITED
Clark, JA. 969. The detection of various bacteria indicative of water pollution by a presence-
absence (P-A) procedure. Can. J. Microblol. j :77 1-780.
Clark, J.A. 1980. The influence of increasing numbers of nonindicator organisms upon the
detection of Indicator organisms by the membrane filter and presence-absence tests. Can. J.
Mlcrobiol. 26: 827-832.
Craun, G. F. 1978. Impact of the coliform standard on the transmission of disease. p. 21-36. In:
Evaluation of the Microbiology Standards for Drinking Water, C. W. Hendricks (ei)
Publication EPA-570/9-78-OOc, Office of Drinking Water, U.S. Environmental Protection
Agency, Washington, D.C.
Environmental Protection Agency. 1976. National Interim primary drinking water regulations.
Publication EPA-570/9-76-003, Office of Water Supply, U.S. Environmental Protection
Agency, Washington, D.C.
Geldreich, E. E. and I-I. Kennedy. 1982. The cost of microbiological monitoring. In: Bacterial
Indicators of Pollution, W. 0. PIpes (ed.) CRC Press, Boca Raton, Florida.
Goshko, M. A., W. 0. PIpes and R. R. Christian. 980. Correlations between collform species
occurrence and chlorine residual In various water distribution system. 1980 Annual
Conference Proceedings, Am. Water Works Assoc., paper 26-3, pp. 959-971.
Public Health Service. 1962. DrinkIng water standards. U.S. Department of Health, Education,
and Welfare, Washington, D.C.
Pipes, W. 0. and R. R. Christian. 1982. Sampling Frequency - Microbiological Drinking Water
Regulations. Office of Drinking Water, U.S. Environmental Protection Agency, Washington,
D.C.
Seidler, R. J., T. M. Evans, J. R. Kaufman, C. E. Warrick, and M. W. LeChevolier. 1981. LimIts of
standard coliform- enumeration techniques. Jour. Am. Water Works Assoc. 73:538-542.
-------
40
Thomas, H. A. 1952. On averaging results of coliform tests. Jour. Boston Soc. Civil Engineers.
39:253-259.
Victoreen, H. 1. (969. Soil bacteria and color problems in distribution systems. Jour. Am. Water
Works Assoc. 61:429-437.
-------
41
APPENDIX A
COMMENTS ON CHECK SAMPLING AND THE MCL
REPRESENTING A MINORITY REPORT BY THE CHAIRMAN
By: Wesley 0. Pipes
-------
42
COMMENTS ON CHECK SAMPUNG AND THE MCL
REPRESENTING A MINORITY REPORT BY THE CHAIRMAN
Wesley 0. Pipes
The Monitoring Panel recommended a minimum of 5 samples for bacteriological examination
each month for the smallest systems covered by the Regulations. For reasons explained in the
Issues Paper, we wanted at least 60 samples each 12-month period and no more than 3 of the 60
(5%) with coliforms present. However, it was difficult to make the 3 out of 60 samples an MCL,
since violation of an MCL requires public notification and the finding of 4 posItive samples
scattered in time over 12 months does not seem to be an adequate basis for notifying the public
that the water may be unsafe to drink. Thus, the 3 out of 60 samples in 12 months was
recommended as a protection reliability standard, violation of which would initiate an investigation
to locate the source of the contamination and eliminate it.
This leaves the question of what to recommend for an MCL still open. In our discussions, we
decided that the first positive sample in a month’s time should not initiate public notification.
Since the smallest systems would be required to take 5 samples per month, the MCL for those
systems becomes no more than I out of 5 samples (20%) each month with coliforms present. This is
the MCL which the Monitoring Panel agreed to recommend.
My interpretation of our discussions is that we did not intend the MCL of no more than 20%
of the samples examined each month with coliforms present to be applied to all systems,
independent of the number of samples per month examined. I believe that we did not discuss
adequately the question of the MCL for larger systems examining more than 5 samples per month.
I also believe that we did not discuss adequately the question of how to use the results of check
samples, although we did agree to recommend that check samples be used in calculating the
fraction of samples with coliforms. At least one member of the Monitoring Panel disagrees with
this interpretaf ion.
The following is my proposal on how to use check sample results In determining whether or
not the MCL Is violated in any month’s time. It involves a change in the MCL recommended by the
Panel as a whole. it is offered as an alternative for consideration in the process of establishing
new regulations.
-------
43
I propose that at least three check samples be collected for every positive sample found. One
check sample should be collected from the same tap where the original positive sample was found
and the other two from nearby sites. A positive check sample would require collection of three
more check samples. Allowing time for transport of the samples to the laboratory, sample
examination, and return of results to the water utility, I assume that the check samples will be
collected about one week after the positive sample was found.
The results of the examination of check samples should be used in calculating the fraction
positive to determine if the MCL is exceeded In any month. The rationale for this is that it
encourages the collection of additional samples when a positive sample is found. Under the present
Regulations the collection of additional samples to “average down” to avoid exceeding an MCL is
allowed but not explicitly recognized.
1 propose the following MCL: if less than 20 original and check samples are collected in any
month, the second positive sample would cause a violation of the M.CL; if between 20 and 39
samples are collected in any month, the third positive sample would cause a violation of the MCL;
if between 40 and 59 samples are collected in any month the fourth positive sample would cause a
violation of the MCL; if between 60 and 79 samples are collected in any month the fifth positive
sample would cause a violation of the MCL; and if 80 or more samples are collected in any month
the MCL is more than 5% of the samples positive. Calculations to illustrate this concept ore given
in the attached table.
For small systems, if the original samples ore collected and examined early in the month and
one or two positive samples are found, the utility has the option of collecting additional samples in
an attempt to avoid violating the MCL. If the additional check samples are negative, this is
evidence, that the contamination is not extensive, If some of the additional check samples are
Positive, this may convince the water utility personnel that there really is a problem which needs
to be corrected. The number of additional check samples needed for “averaging down” increases
with the size of the system. Small systems (5-14 original samples) would need only I to 12
additional negative samples to meet the MCL after the second positive sample is obtained. A
system collecting 68 original samples would need to collect an additional 20 negative samples to
meet the MCL after the fifth positive sample is obtained.
-------
44
TABLE. Calculations to Illustrate Proposed MCL
Original
Samples
Check
Samples
Total
Samples
No.
of Positive
Samples Which
Allowable
Fraction
Meets
Exceeds
MCL
MCL
Positive
5
3
8
I
2
0.125
16
3
19*
I
2
0.053
•
14
6
20
2
3
0.100
33
6
39*
2
3
0.051
31
9
40
3
4
0.075
50
9
59*
3
4
0.051
48
12
60
4
5
0.067
67
12
79*
4
5
0.051
68
12
80
4
5
0.050
*Clear ly, in these cases, If time Is available In any month, the water utility would collect more
samples to ovoid exceeding the MCL.
-------
45
One purpose of this proposal is to avoid going to public notification when there is only a small
amount of evidence about microbiological contamination, but to require public notification when
there is substantial evidence of contamination. An integral part of the proposal is that the State
should be notified as soon as possible for occurrence of samples with coliforms present. If
coliforms occur in the check samples, the question of when to notify the public should be resolved
between the State and the water utility, but it should be no more than four weeks after the first
positive sample is found.
The recognition in the Regulations of the option of taking additional samples to avoid
violating the MCL is desirable. It would institutionalize a current practice. It would encourage
water utility personnel to recognize sampling as a method of investigating problems of micro-
biological contamination. It probably would encourage State agencies to work more closely with
water utilities in solving problems of microbiological contamination.
-------
ANALYTICAL METHODS FOR MICROBIAL WATER QUALITY
Ramon J. Seidler and Thomas M. Evans
Panel Members: James Clark, Barbara Green, Lawrence Leong, Raymond Lundgren, Gordon
McFeters, Pasquale Scarpino, Ray Taylor
ABSTRACT
INTRODUCTION 3
COLIFORM DETECTION 3
Definition of a Coliform 3
FERMENTATION TUBE TECHNIQUE 5
Historical Development of the Presumptive Test 5
Historical Development of the Confirmatory Test
Factors which Affect Sensitivity of the MPN TechnIque 8
Confirmed vs. Completed Test Results 10
Reliability of the MPN Index 10
Potential Limitations of the MPN Technique 12
Summary 12
MEMBRANE FILTER TECHNIQUE 13
Effect of Stress on Coliform Detection IS
Factors Influencing Sensitivity of the MF Technique 17
Summary 19
ALTERNATIVE COLIFORM DETECTION TECHNIQUES AND MEDIA 20
Presence-Absence Test 20
Alternative Fermentation Tube Presumptive Media 21
Alternative Membrane Filtration Media 21
Summary 22
OTHER DETECTION PROCEDURES FOR THE ANALYSIS OF MICROBES OR THEIR
PRODUCTS IN DRINKING WATER 22
Coilform Detection by the B-Galactosidase Assay 23
Detection of Coliforms by Gas Chromatographic Techniques 24
Detection of Microbial Blomau Through Quantification of Cellular ATP Levels 2k
Limulus Amoebocyte Lysate Assay to Detect Lipopolysaccharide 25
Measuring the )-Ieterotrophlc Bacterial Population in Drinking Water 26
Aeromonas, An Indicator Organism or An Interference with Potable Water Analyses 29
Feccil Coliforms as Indicators of Drinking Water Quality 32
Coliphag. as Indicators of Fecal Pollution 33
Automated Electrical Impedence Test for Monitoring Fecal Collforms 3k
Rapid Most-Probable-Number Fecal CoU form Enumeration Procedure 34
Rapid Seven-Hour Fecal Coliform Test by Membrane FIltration 35
TURBIDITY 35
CHLORINE RESIDUAL MEASUREMENTS 38
COLLECTION AND TRANSPORT OF DRINKING WATER SAMPLES 39
SUMMARY 42
RESEARCH RECOMMENDATIONS 43
LITERATURE CITED 44
ABSTRACT
A review is presented of the microbiological techniques, sample transport and storage, and
methods to measure chlorine and turbidity in finished drinking water. The definition of a coilform
is expanded to reflect current knowledge on the ecology and taxonomy of gram-negative lactose
-------
2
fermenting bacteria. The expanded definition represents an attempt to define the indicator group
as a taxonomic assemblage of genera and species which share physiological ecological similarities.
Lactose-fermenting members of the genus Aeromonas would be included in this definition of
indicator bacteria. As defined, the indicator group can be monitored by conventional techniques
which rely on gas formation or distinct colony morphology. However, the committee stresses that
the group can also be monitored by other innovative procedures and media and encourages
development of new techniques to improve rapid and economical detection concepts.
A variety of alternative microbial enumeration techniques and media are reviewed. These
include measuring electrical impedance, analysis of AlP, the presence-absence (P-A) test,
reactivity with Limulus lysate, gas chromatography, and enumeration of coliphage. The committee
felt that no alternative technique except the P-A test has progressed far enough to warrant
possible direct application to routine monitoring of drinking water quality.
An accurate measure of drinking water turbidity is required by the Interim Primary Drinking
Water Regulations (IPDWR) to monitor water quality. Significant problems are associated with
turbidity measurements because of technician error, poor instrument calibration, and unstable
standards. The infrequent microbiological sampling in small water supplies, especially during times
of elevated turbidities, is a serious flaw in the IPDWR. Since turbidities in water can mask
coliform detection by MF, the committee suggests a regulation to require use of the fermentation
tube technique during times in which the MCL for turbidity is exceeded.
The precision and accuracy of ten standard methods to measure residual chlorine were
compared. All techniques and procedures were found to be acceptable except for one commercial
field test kit which was inaccurate.
Quality assurance and training programs must be retained in order lo provide technical
transfer of information to personnel In the water industry. Federally sponsored short training
courses conducted at public facilities are useful, provide dialogues on emerging problems, and
provide visibility for drinking water programs.
Transport and storage of potable water samples result in significant changes in the indicator
population. Storage at room temperature or In the refrigerator for just 24 hr results in changes in
coliform counts in a significant part of the samples. At present there Is no known way to predict
the direction or magnitude of changes in indicator counts with sample storage nor is there an
available method to stabilize the count s.
-------
3
INTRODUCTION
Providing potable drinking water to the public depends on a monitoring program to detect
potential health risks. Monitoring drinking water supplies for total coliforms, residual chlorine, and
turbidity can Identify treatment and distribution deficiencies which may result in disease outbreaks
and may indicate corrective procedures to control an outbreak’s duration. Success in any
monitoring program Is greatly influenced by the efficiencies of the techniques used. In this report,
each technique is analyzed to determine the factors which may influence its sensitivity and
accuracy. The length of any one section should not be taken as on indication of Its Importance but
reflects the information which was available to the authors on each topic.
COLIFORM DETECTION
Definition of a Coilform
Total coliform bacteria are currently defined according to cultural characteristics which are
apparent in the two techniques used for their detection. Gas production with the fermentation
tube technique and sheen formation with the membrane filter technique provide the bases of the
two operational definitions of a cot iform. Some have considered that physiological or operational
definitions of a coliform are too restrictive. Clark and Pagel (1977) concluded that taxonomic
considerations (generic identification) as well as gas or sheen production may be useful in providing
Information about the degree, type and source of fecal pollution in the environment.
The operational definition of a cotiform for use with the multiple tube technique evolved as a
result of investigations into the sanitary significance of coliforms in the environment. Breed and
Norton (1937) considered the conform group as all aerobic and facultative-anoeroblc, gram
negative, non-spore-forming bacilli which ferment lactose with gas formation. This is essentially
the definition adopted by Standard Methods for the examination of water and wastewater
(American Public Health Association, 980). included as part of this definition is the requirement
that gas be produced at 35°C within 48 hr.
Early investigations concerning coflform detection limited Incubation to 48 hr because
bacteria, typically Erwinia and Serratia , which fermented lactose after 1e8 hr were felt to have
little sanitary significance (Prescott, 1946). in addition, these slow lactose fermentors tended to
give variable lactose fermentation reactions when Isolated In pure culture (Mack, 1977). However,
-------
4
slow lactose fermenters should not be summarily dismissed as being of limited sanitary signif-
icance. Parr (1938) has reported the presence of slow lactose fermenters in human feces. An
extension of the incubation time beyond 48 hr could result in greater numbers of coliforms being
recovered. However, the additional numbers obtained may not justify the increased time required
to determine the sanitary quality of the water sample.
The selection of the incubation temperature is based on several considerations. Savage, in the
early 1900’s, classified bacteria found in water as normal inhabitants ( Pseudomonas) , unobjection-
able aliens (bacteria arising from soil such as Bacillus mycoides) , and objectionable aliens (bacteria
arising from feces such as E. coli) (Prescott, 946). Many bacteria of the first two groups as well
as saprophytic bacteria are unable to grow at body temperature. Therefore, growth at 37°C was
considered representative of organisms of fecal origin and could be used to distinguish between
contamination due to fecal and non-fecal sources. The temperature currently used for incubation
is 35°C.
Prescott, ef al, (1946) have reviewed several studies which examined the influence of
incubation temperature on lactose fermentation. Many coliforms have been isolated that ferment
lactose at 30°C or less but not at 37°C. Although data were not shown, Geldreich (1975)
mentioned that the rate of gas production was a function of incubation temperature when con-
ducting MPN analyses of river water samples. The rate of gas production decreased as the
incubation temperature was lowered below 37°C. Taguchi (1960) demonstrated that 35°C was the
optimum temperature for recovery of E. coil from well water and sewage when using an ogar plate
count technique. From this information it is apparent that the justification for using 35°C as the
incubation temperature for coliform analyses is not well documented. Therefore, temperature
other than 35°C should be examined to determine if 35°C is the optimum temperature for coliform
detection and enumeration.
With the membrane filter (MF) technique, the definition of a coliform is based on the
production of a dark colony with a metallic sheen within 24 hr at 35°C on an Endo-type medium
containing lactose (American Public Health Association, 1980). Since sheen production on Endo.
type agar is due to ocetylaldehyde production by qoliform bacteria (Schiff, et al. , 1970), there is
not necessarily a high correlation between sheen formation and gas production (Clark, 1969; Evans,
et al, 1981 b; McCarthy, et al., 1961). The different operational definitions of a coliform for the
-------
5
MF and MPN techniques may result in different species of coliforms being detected by the two
techniques (Evans, et cii., 198 I a; Dutka and Tobin, 1976).
There are obvious problems in having two operational definitions for indicator organisms. It
is agreed that these bacteria share a similar metabolism and have ecological distributions which
are not discontinuous. With one common definition the indicator group could be monitored by any
technique or method that has proven value in recovering representatives of this taxonomic
assemblage. Therefore, it is recommended that a new definition of indicator organisms be
formulated to include lactose fermenting gram-negative, non-sporeforming rods which comprise
the taxonomic assemblage of species now recognized as coliform indicator organisms ( Citrobocter,
Enterobacter, Escherichia, Kiebsiella ) and other metabolically similar species, i.e., Aeromonas,
Hafnio, Serratia, Yersinia that conform to this definition. As presented, the indicator group could
continue to be measured by the conventional techniques which rely on gas formation or distinct
colony morphology. However, the more general taxonomic definition should encourage new and
innovative methods development for finding more rapid and economical means to detect these
bacteria.
FERMENTATION TUBE TECHNIQUE
Historical Development of the Presumptive Test
The earliest detection procedure for coliforms relied on direct plating of the sample onto
litmus lactose cigar and acid production to differentiate between coliforms and non-coliforms
(Prescott, j [ ., I 91 6). Irons, in 1902, reported that preliminary enrichment of the sample in liquid
media was a more sensitive method of detection than direct plating. Additional studies indicated
the enrichment technique could be made quantitative by Inoculating known volumes into a series of
tubes (McCrady, 1915). NutrIent broth containing I percent glucose was commonly used as the
enrichment medium. Since other bacteria besides coliforms fermented glucose, It was necessary to
isolate, using a solid cigar medium, and identify the organisms present (Prescott, et, I 9 1 i6).
The time and materials required for this procedure prompted studies to find more suitable
techniques. Lactose was substituted for glucose In the enrichment medium in the second editlonof
Standard Methods (American Public Health AssociatIon, 1913) with the Idea that a rapid
“presumptive” test for coliforms could be obtained. This first formulation of lactose broth
-------
6
contained bile as a selective agent. Later research indicated bile was inhibitory to some coliforms
(Prescott, et al., 1946), and bile was excluded from the formulation of lactose broth in the third
edition of Standard Methods (American Public Health AssociatIon, 1917).
it should become apparent that non-coliform bacteria, notably aerobic and anaerobic spore-
forming bacilli ferment lactose and cause false-positive presumptive tests (Dominick and Lauter,
1929; Koser and Shinn, 1937; Norton and Barnes, 1928; Thompson, 1927). In addition, the synergis-
tic action of two non-coliform bacteria was discovered as another cause of false-positive tests
(Grer and Nyhan, 1928; 1-lolman and Meekinson, 1926; Leitch, 1925; Sears and Putnam, 1923). These
associations were usually one gram-positive (typically Streptococcus or Staphylococcus ) and one
gram-negative (typically Vlbrio, Salmonella or Proteus ) bacterium utilizing products of inter-
mediary metabolism (Holman and Meekinson, 1926; Sears and Putman, 1923).
The occurrence of high numbers of false-positive results in lactose broth prompted a series of
investigations which were designed to develop an improved presumptive medium. Numerous
presumptive media containing dyes and/or bile as selective agents were compared to lactose broth
(Farrell, 1936; Jordan, 1927, 1932; Ruchhoft, 1935; Ruchhoft and Norton, 1935). These media
included buffered lactose broth, brilliant green bile lactose broth (both 2 and 5 percent bile),
fuchsfn lactose broth, methylene blue-brom-cresol purple broth, crystal violet lactose broth,
formate ricinoleate broth, MacConkey’s peptone lactose bile broth, and trypaflavine broth.
Without exception, lactose broth was Judged the most sensitive presumptive medium for the
detection of coilforms.
Lactos, broth remained the only presumptive medium specllid by Standard ±! 2 until the
introduction of lauryl tryptose broth (LTB). Lauryl tryptose broth Ii a phosphate-buffered lactose
medium and was formulated as a result of studies which determined the effects of culture
ingredients on the growth rate of pure cultures of coliforms (Darby and MaIlman, 1938). A
selective agent, sodium iauryl sulfate Is included in this medium (MaIlman and Darby, 1941).
Comparisons of lauryl tryptose broth and lactose broth indicated the use of louryl tryptose broth
would reduce the number of presumptive tubes needed to be proàessed and Increase the number of
coilform-positive results (HaJna and Perry, 1943; MaIlman and Darby, 1941; McCrody, 1943).
Lauryi tryptose broth was introduced In the ninth edition of Standard Methods (American Public
Health AssocIation, 1946) and replaced lactose broth In the 15th edItion (1980). Recent
-------
7
information has once again added to the controversy as to which presumptive medium should be the
medium of choice. injured total coliform bacteria are detected more effectively with LB than with
LTB (McFeters, et al., 1982).
Historical Development of the Confirmatory Test
Confirmation of gas-positive presumptive tubes was first conducted by subculturing onto
Endo agar (Prescott, et al., I 9k6). With the modification of EMB by Levine (1918), EMB was
included in the fifth edition of Standard Methods as an additional confirmatory medium (American
Public Health Association, 1923). DIssatisfaction with the amount of labor required to process agar
plates and the need to decide which colonies were to be Inoculated into secondary lactose broth
tubes prompted research to develop a broth confirmatory medium. One of the first media to be
tested was brilliant green lactose bile broth (BGLB). In an extensive study conducted by Jordan
(1927), BGLB was found to compare favorably with EMB as a confirmatory medium; approximately
95 percent of the presumptive tubes which confirmed using EMB also confirmed using BGLB. Other
confirmatory media which have been compared to EMB and BGLB include formate ricinoleate broth
and MacConkey broth (Ruchhoft, 1935). With chlorinated finished drinking water, confirmation in
formate ricinoleate detected 82 percent, BGLB 611 percent, EMB 75 percent, and MacConkey broth,
58 percent of the presumptive tubes that were detectably coliform positive by the four con-
firmatory methods. Additional studies involving both raw water and treated drinking water
compared BGLB and EMB (Richey, I 9111) and BGLB, EMB, formate rlclnoléate broth, fuchsln broth,
and crystal violet broth (McCrady, 1937). Two conclusions can be drawn from these studies. The
use of any one confirmatory medium did not detect alt of the presumptive tubes that were coliform
positive. Secondly, BGLB was found to be a very efficient confirmatory medium but only because
the percentage of positive confirmed tests which yielded coliforms was greater with BGLB than
with any other medium. However, other confirmatory media (EMB and formate ricinoleate broth,
m-Endo agar LES) may yield a much greater number of coliform isolations (Ruchhoft, 1935; Evans,
ef al., 1981 c; Seidler, al., 1981).
Prescott, et ci., (I 9k6) have summarized the advantages of a broth as compared” to an ogar
confirmatory medium. The use of a broth confirmatory medium results in the saving of labor and
materials and provides greater uniformity of results. Transfer from a presumptive tube to a broth
and reading gas production does not require judgment as does selecting colonies from a plate for
-------
8
inoculation into secondary lactose broth tubes. These considerations probably resulted in BGLB
replacing Endo and EMB ogcirs as the confirmatory medium in the 14th edition of Standard Methods
(American Public Health Association, 1975).
Factors Which Affect Sensitivity of the MPN TecIvilque
Successful use of an enrichment medium in a fermentation tube technique depends on
coliform bacteria producing gas in sufficient quantities to result in a positive presumptive test.
Any factor which Inhibits cell growth and gas production in an enrichment medium will Interfere
with coliform detection. Bacteria such as Clostridlum can overgrow coliforms in LB and produce
acid at concentrations inhibitory to coUforms (Greer and Nyhan, 1928; Norton and Barnes, 1928).
Other species such as Pseudomonas, Flavobacterlum, Actinomyces, Sorcina , and Micrococcus may
overgrow col iforms In LB (Hutchinson et ci., 1943). Herson and Victoreen (1980) have demon-
strated that Flavobacterlum sp., Acinetobocter sp., and Arthrobacter sp. may Inhibit the growth of
E. coil. However, the Interaction of the coliform and non-coilform was studied in a bacterial
extract medium and not In either of the two presumptive media that may be used In the MPN tech-
nique This work might explain how interactions between coliforms and non-coliforms Influence
coliform die-off within a distribution system.
Indirect evidence for Inhibition of coliforms In presumptive media by bacter)ocin-Ijke
substances was recently reported (Means and Olson, 1981). In that study Flavobacterlum and
Moraxella were found to produce bocterlocln-Ilke substances capable of inhibiting Escherlchlci,
Klebsiella and Enterobocter in an ogar overlay assay. Other conditions such as the presence of
coliphage (Schiemann et ci., 1978) or high concentration of nitrates and/or nitrites (Tubkish, 1951)
in a water sample may also inhibit the growth of coliforms.
IndIrect evidence for the Interference by standard plate count (SPC) bacteria with coliform
detection was reported by Geidreicli, et ci., (1972) in a study of community water supplies. They
reported a decreased Incidence of coliforn, isolation when the SPC population of the water sample
exceeded 500 CFU/rnl. Furthermore, Geldreich, et c ii., (1978) reported that coliforms could be
recovered from presumptive tubes which were gas-negative but turbid. This observation has been
confirmed by other investigators (Evans, et ci., 198 Ic; Olson, 1978).
The quantitative impact of detecting coliforms in turbid, gas-negative presumptive tubes on
the value of the MPN Index has been reported (Evans, et al, 198 Ic; Seidler, et ci., 1981). Coliform
-------
9
isolations from turbid gas-negative presumptive tubes may increase the number of coliform
positive tubes in an MPN analysis by as much as 41 percent for LB and 28 percent for LTB (Seidler,
et cii., 198 I). In that study the magnitude of cot iform interference was found to increase as the
number of SPC bacteria increased when using standard methods.
The efficiency of the confirmatory medium in recovering coliforms from presumptive tubes
can also affect the sensitivity of the MPN procedure. A study conducted by Ruchhoft in 1935
suggested that BGLB may be unsuitable as a confirmatory medium, especially for the examination
of chlorinated drinking water. In one set of samples, where the residual chlorine was neutralized
with a sterile peptone solution, coliform Isolation from BGLB was only 58 percent of the number
isolated with EMB cigar. A more recent study compared BGLB and m-Endo cigar LES as
confirmatory media for recovering cotlforrns from both gas-positive and turbid, gas-negative tubes
of LB and LTB (Seidler, ., 1981). In that study, BGLB gave positive confirmed tests for only
53 percent of the LB and 45 percent of the LTB presumptive tubes that were coliform positive by
m-Endo cigar LES. Clearly, BGLB is inhibitory to the growth of some collforms. Chambers, in
1950, reported that the cell density required to produce gas In BGLB Is greater than that necessary
to produce gas in LB. Direct evidence that BGLB is inhibitory to gas production by coilforms
comes from work conducted by Hajna and Damon (I 955a) who found that coflforms could be
recovered from gas-negative tubes of BGLB.
The sensitivity of the MPt4 technique is also influenced by the completion cigar in detecting
coliforms which grow in the confirmed test medium. Two separate studies concluded that
confirmatory EMB and Endo cigars were equally successful for use in isolating coliforms from LB
presumptive media (Ruchhoft, et al., 1931; Greet, et at., 1928). Later when Mailman and Darby
(1941) were evaluating LTB for use as a presumptive medium, they found that many col iforms
growing In presumptive media were not detected on EMB. They also found that the composition of
the secondary lactose broth Influenced whether the colony isolated and subcultured from EMB-
produced gas. The results of these studies illustrate the serious limitations of the completed test:
failure to detect coflforms when present and the difficulty in demonstrating gas production by the
bacteria isolated, In addition, the failure to detect collform colonies on EMB and Endo cigars by
streaking gas-positive tubes of BGLB should not necessarily be regarded as a false-positive
confirmed test but as a failure of the agar media to give a reaction which can be used to
distinguish coliforms from non—coUforms.
-------
I0
Confirmed vs. Completed Test Results
Standard Methods (American Public Health Association, 1980) and the United States EPA
(Bordner, et al., 1978) require completed tests to be performed on 10 percent of the positive
confirmed tests. The results of several studies indicate that completing only tO percent of the
confirmed tests may not be sufficient to give an adequate indication of the number of coliforms in
a sample. Bowmer and Campbell (1972) reported that only 61 percent of the drinking water
samples that give positive confirmed tests yielded positive completed tests. Geldreich, et at.,
(1965, 1967) found the value of the completed test MPN index was signifIcantly lower than the
confirmed MPN index. These results suggest that the completion test should be performed on oil
samples. However, the data presented by MaIlman and Darby (I 9E I) suggests that EMB agar may
be Inhibitory to some coliform organisms. Therefore, before a completion test is required for all
MPN analyses, the efficacy of using EMB ogar or other ogar media in the completion test should be
determined.
R.Uabflltyof the MPN Index
The mathematics of a maximum likelihood or most-probable-number method of estimating
bacterial numbers were developed by McCrady In 1915. The value of the MPN index is based on
probability theory, i.e., The number of positive and negative results obtained in a sample dilution
series is a function only of the d ensity of the organisms in the sample. There are two principal
assumptions on which the estimate is based (Cochran, 1950). The first requires the organisms to be
randomly distributed throughout the sample. The second is that each sample when inoculated Into
the growth medium will exhibit growth whenever a single viable organism Is present. The second
assumption can be verified by determining The probabilIty of obtaining the particular positive
results from the various dilutions. Woodward (1957) has tabulated the probabilities of obtaining the
number of positive tubes In each dilution series. If improbable results are obtained, either
technician error has occurred or condItions are present which interfere with the growth of
coliforms in the fermentation tube.
One serious lImitation of an MPN technique Is the general lack of precision in the results.
The precision of the fermentation tube technIque can be Illustrated by examining the 95 percent
confidence limits of a given value of the MPN Index. Consider a fermentation tube analysis In
-------
II
which five tubes of three ten-fold dilutions are examined. The lower 95 percent confidence Is
obtained by dividing the index by 33; the upper 95 percent confidence value by multiplying by 3.3
(Cochran, (950). This multiplication factor clearly illustrates the poor precision in estimating
cot iform densities.
Thomas ((955), in a theoretical consideration of the fermentation tube technique, concluded
the value of the MPN index is positively biased, The amount of bias is a function of the number of
tubes in each dilution series and has been determined both experimentally and theoretically. Using
pure cultures of cotiforms, McCarthy, et ci., (1958) found with a five tube dilution series the value
of the MPN index was biased by 29 percent relative to pour plate enumeration. The amount of
positive bias agreed with the 23 percent obtained through a theoretical evaluation of the MPN
technique (Thomas, 1955; Thomas and Woodward, 1955). An MPN index which Is developed using
three tubes per dilution series is biased by 43 percent or about twice as much as a five tube dilution
series (Thomas and Woodward, 1955). it should be noted that the method of statistical analysis of
MPN data will influence the degree of positive bias observed. McCarthy, ., (1958) found that
analysis of data with a harmonic mean resulted in the least amount of bIas (lB percent) and an
arithmetic mean (23 percent) the greatest amount. Thomas (1955) derIved a formula for correcting
an Individual value of an MPN index for bless
.0.005/n
• ‘T
where l is the corrected value of th, MPN Index, 1 T the tabulated valu of the MPN index and vs
is th. number of tubes per dilution series.
Such a correction factor should be used when formal comparIsons of coliform detection by
the MPN and MF techniques are mcd .. When corrections for bias In the MPN technique have been
mode, there have been close agreements between the MF and MPN techniques (Thomas and
Woodward, 1958; Adams, 1957).
Since the tables used to determine the value of the MPN Index have a positive bias included,
the tables should be corrected to give a statistically “true” value of the MPN index. it is the
committee’s understanding that efforts are currently underway (Martin Hamilton, Montana State
University) to correct the tables for bias. Such research should be encour gsd so that the next
edition of Standard Methods con contain corrected tables.
-------
12
Potential Limitations of the MPN Technique
Factors which interfere with coliform detection must be considered when using the MPN
technique to monitor drinking water supplies. These factors occurring alone or in combination can
result in a significant underestimation of coliform numbers. For example, one study reported that
interference with coliform detection in the MPN technique occurred in 80 percent of the drinking
water samples examined and resulted in a five-fold underestimation of coliform numbers (Evans, et
al., I 98 Ic). The major causes of this underestimation were reported as failure of coliforms to
produce detectable gas in the presumptive medium (LTB) and the inhibition of coliforms by BGLB
(Evans, et al., I 98 Ic). Furthermore, the magnitude of coliform underestimation increased as the
number of standard plate count (SPC) bacteria increased, with the greatest underestimation
occurring at SPC densities of 250 ml (Seidler, etal., 1981).
Modification of the MPN technique may be necessary when water supplies contain a high
number of SPC bacteria. Seidler, et al., (1981 b) proposed that both gas-positive and turbid, gas-
negative presumptive tubes of LTB be confirmed using m-Endo agar LES. This modification was
reported as being successful in reducing the underestimation of coliform numbers when using the
MPN technique.
Laboratories which monitor drinking water supplies should determine If the MPN technique is
subject to interferences in their region. If such interferences are found, at least 10 percent of the
turbid, gas-negative presumptive tubes should be submitted to a confirmed test. When coliform
interference is noted, LTB presumptive medium should be used since fewer turbid, gas-negative
presumptive tubes are found with this medium as compared to LB.
The fermentation tube technique is subject to several limitations which reduce its effective-
ness in the enumeration of coliforms. False-positive presumptive tests occur in both LB and LTB
media. However, LTB is generally subject to fewer false-positive reactions than LB. False-
negative presumptive tests occur with about equal frequency In both media. For this reason turbid,
gas-negative presumptive tubes must be submitted to a confirmatory test. Since LTB yields fewer
turbid, gas-negative tubes than LB, LTB should be used when the occurrence of false-negative
presumptive tubes is a problem. However, LTB presumptive medium may not be as effective in
detecting Injured coliforms as LB.
-------
13
The limitations of BGLB as a confirmatory medium should be recognized by all laboratory
workers who use the fermentation tube technique. The use of BGLB as a confirmatory medium
may result in an underestimation of the number of coliforms. Alternative confirmatory media such
as m-Endo agar LES and other media should be examined to determine if the performance of the
confirmatory test can be improved.
The value of the MPN index overestimates the number of coliforms in a water sample. The
amount of bias is a function of the number of coliforms in a water sample. The amount of bias is
also a function of the number of tubes in each dilution series. An MPN index which is developed
using three tubes per dilution series is biased by L 3 percent or about twice as much as a five tube
dilution series. Tubes used to compute the MPN index should be corrected to give a statistically
“true” value.
MEMBRANE FILTER TECHNIQUE
The general application of the membrane filter to the bacteriological analysis of water began
in the early 1950’s with the commercial availability of membrane filters (Goetz and Tsuneishi,
195 I). In the United States, Alexander Goetz developed the methodologies for the manufacture
and use of membrane filters for the examination of water. In 1951, Clark, et oI., reported the use
of a modified Endo medium (EHC Endo) with membrane filters to examine water samples for
coliforms. Their procedure involved a 2-hr preliminary enrichment of the membrane filter on
Albini M medium before the membrane was transferred to an absorbent pad saturated with EHC
Endo broth. This procedure was successful in detecting coliforms but was quantitatively Inferior to
the MPN technique (Clark, et al., 1951; Kabler, 1954; Shipe and Cameron, 1954; Thomas and
Woodward, 1955, 1956). However, because of the inherent advantages of the membrane filter (MF)
over the MPN technique, filtration was adapted as a tentative method In the tenth edition of
Standard Methods (American Public Health AssociatIon, 1955).
The quantitatively inferior nature of the MF technique relative to the MPN technique was
thought to result from certain coliform organisms not producing a sheen on the EHC Endo medium
(Clark and Kobler, 1952) and high numbers of non-coliform bacteria Inhibiting growth or sheen
development by coliforms (Clark and Kabler, 1952; Yee, et ol., 1953). Brilliant green (Kabler,
1954), 8-hydroxyguinollne (Yee, et al., 1953) and sodium desoxycholate (Hojna and Damon, 1955)
were tried with various formulations of EHC Endo broth to reduce the number of non-coliform
-------
14
organisms and enhance sheen development. The sodium desoxycholate medium developed by Hajnc,
and Damon (1955) was the most successful in accomplishing these two goals. This medium,
designated M-HD Endo broth, had the added advantage of not requiring an enrichment step.
Comparisons of coliform detection with the MPN technique and M-HD Endo broth indicated the
MPN technique still recovered greater numbers (Adams, 1957; McCarthy, 1955). McCarthy (1955)
attributed the lack of agreement between the two techniques to failure of some coliform organisms
to produce the typical metallic sheen on the M-HD Endo medium.
Fifield and Schaufus (1958) formulated another Endo-type medium, m-Endo broth, which they
found to be superior to M-HD Endo broth. Enhanced sheen production and an approximate
35 percent increase in coliform recovery as compared to M-HD Endo broth were reported for this
new medium. In addition, the authors reported these results were obtained without a pre-
enrichment step. The m-Endo broth medium replaced E l - IC Endo broth in the I I th edition of
Standard Methods which also accepted the membrane filtration as a standard technique (American
Public Health AssociatIon, 1960). m-Endo broth is one of the two MF media accepted by the latest
edition of Standard Methods (American Public Health Association, 980).
McCarthy and Delaney (1958) used pure cultures of coliforms to evaluate sheen production
and growth on m-Endo medium. Four of 14 coliform species tested gave poor to no sheen on
membranes incubated on m-Endo medium. The coliforms which failed to produce a sheen were
Citrobocter and Aerobocter ( Enterobocter ) species. These authors also quantitatively compared
growth on m-Endo saturated filters to growth on tryptose glucose agar. Depending on th. strain
tested, growth on th. membrane filters ranged from 73 percent to 137 percent, of that on tryptose
glucose 09cr. Additional studies also indicated that m-Endo broth may be slightly inhibitory to
coliforms (McCarthy, ! [ ., 1961). Pure cultures of coliforms, inoculated into m-Endo broth,
showed an increased lag time compared to lauryl tryptose broth.
Based on these studies, McCarthy, Delaney and Grasso (1961) formulated m-Endo agar LES.
The membrane filter procedure they developed involved a 2 h pro-enrichment on pads saturated
with lauryi tryptose broth and transfer of the filter to m-Endo agar LES. Coliform recovery with
the MF technique using m-Endo agor LES compared favorably with the MPN technique for rivers,
lakes, domestic sewage and drinking water samples (Dutka and Tobin, 1976; Lin, 1977; McCarthy, et
at., 1961). A high percentage (87 percent) of the sheen colonies obtained with m-Endo agar LES
-------
15
were found to verify, that is, produce gas when inoculated into lactose broth. The success of rn-
Endo agar LES in the MF technique resulted in this medium being adopted for use in the 12th
edition of Standard Methods (American Public Health Association, 1965).
Comparisons of coliform numbers recovered with m-Endo broth or m-Endo agar LES have
generally indicated that m-Endo agar LES Is superior (McCarthy, 1961; Morgan, et al., 1965).
However, the performance of these two media was not determined under totally comparable
conditions. m-Endo broth was not used with a pre-enrichment step (McCarthy, 1961; Morgan, et
al., 1965) as m-Endo agar LES was, and sheen colonies were not always verified (Morgan, et al.,
1965). Therefore, It is still not clear whether these two media are comparable in the quantitative
recovery of coliforms.
Effect of Stress on Coliform Detection
Coliform bacteria in the aquatic environment are exposed to a series of physical and
chemical stresses that can adversely influence their growth on selective isolation media. Injury
may be related to a number of factors including: time and temperature of aquatic exposure,
chlorine and other disinfectants, heat, freezing, acid mine drainage, heavy metal ions, sunlight and
ultraviolet light, strain of organism, concentration of nutrients, antagonistic standard plate count
bacteria and possibly other undefined chemical and physical parameters (Beuchat, 1978;
Bissonnette, et al., 1975; Busta, 1978; Camper and McFefers, 1979; Fujioka, et al., 1981; Hurst,
1977). In addition, laboratory manipulations involving diluents, selective media and membrane
fitters may cause further underestimations of coliform densities in aquatic environments
(Butterfield, 1932; Geldreich, 1975; McFeters, al., 1982). Injury is an important factor in
underestimating numbers of waterborne Indicator bacteria which may lead to inaccurate public
health assessments in drinking water. Some reports indicate that injured collforms have been
recovered with efficiencies of 10 percent or less on commonly used media (Bissonnette,
1975, 1977; Braswell and Hoadley, 1974; McFeters, etal., 1982).
Recent studies (LeChevallier and McFeters, unpublished data) have demonstrated that
coliform populations occurring in drinking water have significant levels of Injury. A total of 31
samples from four distribution systems in Montana and Massachusetts, Including both chlorinated
and unchlorlnated water, were examined for injury and the mean level of injury was 58 percent
with results ranging from 43 to 86 percent. As a result, over half of the bacteria present were not
-------
16
recoverable with standard MF methods. In environments where total coliform bacteria have been
exposed to chlorine, the MPN technique may detect greater numbers of organisms than the MF
technique (Lin, 1973; Mowat, 1976). This phenomenon is even more pronounced when fecal
coliforms are enumerated by these two techniques (Braswell and Hoadley, 197k). However,
recently McFeters, et al., (1982) have demonstrated that LTB and LB may recover only 34 percent
of the laboratory injured coliforms tested.
A physiological basis for injury has been proposed. Bacteria exposed to chlorine or other
stress are metabolically impaired because of damage to the bacterial envelope (Camper and
McFeters, 1979; Venkobochar, et al., 1977; Zaske, et al., 1980). When exposed to bile salts or
deoxycholate in the detection medium (Endo-type media and LTB) or elevated temperature in the
case of fecal coliforms, these bacteria are not able to produce colonies (Bissonnette, ef al., 1975,
1977; Green, et al., 1977; McFeters, et al., 1982; Rose, 1975).
An important aspect of aquatic stress is that the injury Is reversible (Bissonnette, et qi.,
1975). Incubation of injured cells on a nutrient-rich non-selective medium will result In the repair
of injured coliforms as measured by the ability of these organisms to grow on selective media.
Various procedures have been developed to improve coliform recovery by the MF technique when
examining chlorinated waters. These techniques are outlined in the 15th edition of Standard
Methods (American Public Health Association, 1980). For total coliforms a 2 hr pre-enrichment of
the membrane on LTB is used before incubation on m-Endo cigar LES (McCarthy, et al. , 1961; Lin,
1973). However, this pre-enrlchment procedure has met with limited success when applied to
drinking water samples (Evans, et at., 198 lb). For fecal coliforms, enrichment (Stuart, et at. ,
1977), temperature acclimation (Green, et al., 1977), or both (Lin, I 976b; Rose, et cii., 1975) have
been used. Generally such procedures have resulted In enhanced fecal coliform detection with the
MF technique.
Recently, a new medium for the improved recovery of injured total coliforms from drinking
water has been developed (LeChevallier, et cii., 1982). This medium, termed m-T7 cigar, recovered
nearly three times more verified coliforms from drinking water than the standard MF technique
and over two times more coliforms than the standard LTB resuscitation technique. Research Is
continuing to evaluate m-T7 cigar on a national basis.
-------
17
Factors Influencing Sensitivity of the MF Technique
Numerous factors have been shown to influence the sensitivity of the MF technique in
detecting total coliforms. High numbers of non-coliforms in the drinking water sample may result
in fewer coflform isolations by the MF technique than by other techniques (Clark, 1980; Geldreich,
et al., 1978). Coliforms have also been recovered from apparently coliform negative filters when
water turbidities were5 nephelometric turbidity units (LeChevaHier, etal., 1981).
Lin (I 976a) and Schaeffer, et ci., (1974) reported that the number of total coliforms and fecal
coliforms detected in a water sample varied according to the brand of membrane filter used. The
difference in coliform recovery was thought to be influenced by surface pore morphology which
would influence the rate of diffusion of nutrients to and waste products from the entrapped
bacterium (Sladek, ef al., 1975). Evidence supporting this hypothesis was provided by electron
micrographs which showed that the greatest coliform counts were obtained on filters with the
largest surface pore size (Standridge, 1976). Since the work of these Investigators, membranes
have been designed to optimize surface pore morphology for coLlforrn growth and have displayed
enhanced coliform recovery (Lin, 1976a). However, a recent study has indicated that coliform
recovery may still depend on the brand of the commercially available membranes used (Tobin, et
flj., 1980). The conclusions of that study are open to question because of the statistical
methodologies used to analyze the data (Lemeshaw and Litsky, 1980).
The MF technique was designed to. be a rapid and simple procedure requiring only the
counting and reporting of the number of sheen colonies growing on the membrane filter. Sheen
production by bacteria growing on membrane filters was taken as characteristic of and specific to
the coliform group. For many years, the ME technique was performed In this manner, and the
accuracy of the differential reaction occurring on Endo-type media was not questioned. Recent
investigations have indicated that sheen production may not be specific to the coliform group.
Depending on the water sample examined, the percentage of sheen colonies which produced gas In
lactose broth varied from 44 to 97 percent (Dovenport, !i c ii., 1976; Dutka, 1973; Dutka and Tobin,
1976; Fifield and Schaufus, 1978; Geldrelch, et al., 1967; McCarthy, et al., 1961). Such false-
positive reactions may be due to sheen production by anoerogenic coilforms (Clark and Pegel, 1977)
or by acetoldehyde or proplonaldehyde produced synergistically by two non-coflf arms (Schiff, et 01.,
1970).
-------
18
The ability to produce a sheen on Endo-type media may not be characteristic of all coliform
bacteria. Several studies have indicated that atypical colonies (dark red, pink, or clear and without
any evidence of a sheen) may be aerogenic in lactose broth (Evans, ef al., 1981b; Fifield and
Schaufus, 1958). One report which analyzed a limited number of drinking water samples found
24 percent of the atypical colonies on membrane filters were aerogenic and were representative of
the col iform genera (Evans, et at., 1981 b).
The failure of some sheen colonies to give gas (false-positive reaction) when inoculated into
lactose-containing media has required the adoption of a verification scheme (Bordner, et al., 1978).
This verification scheme consists of subculturing sheen colonies into LTB and BGLB. An evaluation
of this scheme has indicated that coliform organisms may be anaerogenic in LTB or BGLB and
aerogenic in other lactose-containing media (Evans, et at., 1981 b). As a result of this evaluation,
Evans, etal., (198lb) proposed m-LAC broth as a verification medium. The composition of m-LAC
broth was based on the formulation of m-Endo agar LES. Since m-LAC broth contained sodium
louryi sulfate and sodium deoxycholate as selective agents, subculturing of gas-positive tubes of
m-LAC broth into tubes of BGLB was not found to be necessary. Verification of sheen colonies in
m-LAC as compared to LTB/BGLB resulted In a three-fold increase in the number of verified
coliforms in contaminated drinking water samples.
Recently, an alternative to verification of sheen colonies in liquid media has been proposed
(M. Pickeft, 1. J. Leahy, and W. Lltsky, Abstr. Annu. Meet. Am. Soc. Microbiol. 1981, p. 221;
Q124). This method utilizes an enzymatic method of verification. Sheen colonies are examined for
the presence of -galactosiduse and cytochrome oxidase. The developers of this scheme have
reported that colonies which are -galactosidase positive and cytochrome-oxidase negative are
consistent with a taxonomic definition of a coliform. Whether this method is comparable to other
verification methods awaits further field testing.
These investigations involving methodologies for verifying sheen colonies Indicate reactions
occurring on Endo-type media are not adequate for the differentiation of aerogenic and non-
aerogenic coliforms. A new medium should be developed that allows a more distinct differenti-
ation of coliforms from non-coliforms.
Laboratories which use the MF procedure to monitor drinking water supplies should be aware
of potential limitations.with this technique and of means by which the sensitivity of the technique
-------
19
can be improved. The MF technique should not be used when sample turbidities exceed five
nephelometric turbidity units (NTU) (LeChevallier, et at., 1981). When samples of >5 NTU are
encountered, coliform analyses should be performed with the MPN technique.
Atypical colonies that appear on membrane filters should be submitted to a verification
scheme, to determine whether coliform and non-coliform organisms are adequately differentiated
on Endo-type media. Five atypical colonies from a minimum of one plate per month or 10 percent
of all plates examined on a monthly basis should be tested. Whenever typical colonies are absent, a
minimum of five atypical colonies when possible, should be verified. Verification of typical and
atypical colonies in LTB only and not in LTB and BGLB may also improve the sensitivity of the MF
technique (Evans, et al., 1981 b).
The membrane filter technique has enjoyed widespread use In the last 30 years. The
advantages of this technique relative to the MPN ore simplicity of the procedure, obtaining results
in a , shorter time period, and greater precision In the data obtained. However, this technique
should not be applied without recognizing its limitations. Injured total coliform bacteria in
drinking water do not appear to be effectively detected by this technique. In fact 50 percent of
the col iforms present in a chlorinated water sample may be injured and go undetected. New MF
media such as rn-Ti should be evaluated to determine If the sensitivity of the MF test can be
improved.
The sensitivity of the MF technique Is reduced relative to the MPN when the sample contains
either high numbers of non-coliform bacteria or elevated levels of turbidity. The MF technique
should be used when sample turbidlties exceed 5 NTU.
Recent evidence indicates coliform and non-coliform bacteria are not adequately differ-
entiated on Endo-type media. Therefore,’ both typical and atypical colonies should be submitted to
a verification scheme. Verification should be accomplished In LTB or m-LAC broth. The use of
BCLB in verification should be discontinued because of the inhibitory nature of BGLB to gas
production by co liforrns. A new procedure which examines colonies for the presence of
B -galactosldase should be examined as an alternative to verification in liquid media.
-------
20
ALTERNATIVE COLIFORM DETECTION TECHNIQUES AND MEDIA
Presence-Absence Test
Clark (1968, 1969) has developed a presence-absence (P-A) test using a modified MacConkey
broth for monitoring drinking water supplies. The P-A test is essentially a large volume
fermentation tube procedure where a 50 ml or 100 ml volume of drinking wclter is inoculated into
bottles containing MacConkey broth and fermentation tubes and incubated up to five days at 35°C.
Presumptive tests are read daily and gas positive tests are confirmed using BGLB and ECbroth. A
comparison of the number of verified coliforms obtained by the MF technique and confirmed test
results by the P-A test indicated the P-A test detected coliforms in approximately twice as many
samples as did the MF technique (Clark, 1968).
Several disadvantages of the P-A test relative to the MF technique are evident. The
incidence of false-positive presumptive tests with the P-A tests (gas production without con-
firmation) was found to be seven times the number of false-positive tests with the MF technique
(sheen production without verification of any colonies) (Clark, 1968). However, it should be
remembered that the P-A test is not as subject to false-negative results as is the MF procedure
(Clark, 1968). The time required to obtain results with the P-A test may be longer than the ME
technique. Clark and Vlassoff (1973) have reported that 61 percent of the P-A tests were positive
only after 24 hr of incubation. However, with only a 24 hr incubation the P-A test still gives a
greater estimate of the number of coliforms in the sample than the ME technique (Clark and
Vlassoff, 1973).
The concept of the P-A test is an attractive one because of its simplicity and sensitivity in
detecting coliforms. In addition, the results of the P-A test can be used to give a maximum
likelihood estimate of coliform density in a water sample. A number of samples are collected
throughout the treatment or distribution system during the same sampling event and analyzed. The
positive and negative results obtained are considered as one sample and may be converted to a
maximum likelihood estimate, for the number of coliforms within the distribution system. Clark
(personal communication) has developed tables which allow the conversation of presence-absence
data to a most-probable-number of coliform density. A simple formula developed by Cochran
(1950) gives the same results. Possibly, modifications of the presumptive and/or confirmatory
-------
21
media should be examined to decrease the incubation time required to obtain results and reduce the
incidence of false-positive presumptive tests. Such modifications, if possible, would increase the
usefulness of the P-A test and provide a technique which is a suitable alternative to the MF
procedure. Clark (personal communication) has recently modified the medium used in the P-A test
because of the variable performance of MacConkey broth. A comparison of the new medium and
MacConkey broth is currently being conducted and results will be published shortly.
Alternative Fermentation Tthe Prestxnptive Media
Several different media in addition to LTB have been used outside the United States with
successful results. Two chemically defined media have been formulated to optimize gas production
by coliforms. One of these media, formulated by Path, el al., (1977), was found to be superior to
MacConkey broth for the detection of coliforms in treated drinking water supplies. However, when
this medium was compared to LTB, the MPN analyses with LTB generally detected greater numbers
of coliforms in a variety of sewage, surface water, and seawater samples (Dutka and Tobin, 1976).
The second medium, minerals-modified glutamate media (MMGM), is used in England for MPN
analyses of drinking water supplies (Public Health Laboratory Service, 1968). Field testing of
MMGM has indicated that it is superior to MacConkey broth and Teepol broth for detecting
coliforms in chlorinated drinking water supplies (Public Health Laboratory Service, 1968). In
addition, completed test results indicated that MMGM was superior to LTB for the examination of
chlorinated drinking water (Committee, 1980). The superiority of MMGM to LTB was especially
evident when coliform densities were less than 50 coliforms per 100 ml. Also, MMGM was found to
be a better presumptive medium for a fecal coUform ( Escherichia coli) MPN analysis than LTB.
The minerals-modified glutamate medium should be evaluated in the United States as an
alternative to LTB.
Alternative Membrane Filtration Media
Media other than m-Endo broth and m-Endo agar which have been Used to detect coliforms in
drinking water or fresh water Include membrane enriched teepol broth (Joint Committee, 1979),
MacConkey agar (Grabow and DuPreez, 1979) and MC agar (Dufour and Cobelli, 1975). Grabrow
and DuPreez (1979) evaluated m-Endo agar LES, Teepol broth and MacConkey agar for detecting
total coliforms in river water and wastewater. They concluded that m-Erido agar LES yielded the
-------
22
highest average counts of the three media tested. However, It is not clear whether the membrane
filter counts were verified if not, the actual number of coliforms detected could change
substonlially, thereby affecting the conclusions reached in this study.
The MC agar was developed by Dufour and Cabel 11(1975) to enumerate the component genera
of coliforms found in seawater. This medium was compared to m-Endo agar LES in a study
conducted by Dutka and Tobin ($976). In their Investigation which involved sewage, sewage
effluent, and surface water, the highest verifIed coliform counts were obtained with m-Endo agar
LES. The results of these comparison studies indicated that m-Endo agar LES was the best medium
of the ones tested for use with MF technique.
Numerous methods and media other than those currently recommended In Standard Methods
exist for the detection of coliforms. Minerals modified glutamate medium (MMGM) appears to be a
reasonable substitute for LTB in a fermentation tube technique. Coliform detection with MMCM is
greater than that obtained with LTBI The MMCM medium should be evaluated in the United States
as an alternative to LTB.
A number of Investigators have compared the performance of membrane enriched teepol
broth, MocConkey agar and MC ogor to that of Endo type media. None of the media outperformed
the Endo-based media for the detection of coliforms. Therefore, there does not appear to be an
alternative (other than rn-Ti which was discussed previously) to Endo-type media.
The P-A test Is an attractive alternative to both the MF and MPN techniques. The P-A test
is simple to perform and the sensitivity appears to be greater than that of the MF technique. In
addition, quantitative results can be obtained with the P-A test. However, the high Incidence of
false-positive results In the P-A test indicates that a more suitable medium Is needed.
OTHER DETECTION PROCEDURES FOR THE ANALYSIS OF
MICROBES OR THEIR PRODUCTS IN DRINKING WATER
Alternative methods for the convenient or rapid analysis of microbes or their metabolic
products have received much attention In the last ten years. Alternative methodologies seem
motivated by a goal to modernize sanitary microbiology, to reduce detection times In cases of
emergency surveillance needs, and to lower costs. As with most new or alternative techniques
intended to replace old standard procedures, change comes slowly and Is generally resisted. The
-------
23
following section attempts to summarize some methods developed or updated within the last ten
years. No attempt was made to prepare an exhaustive review of all alternative methods to monitor
drinking water, but those discussed represent a sampllng of current interests which show potential
promise for application.
Coliform Detection by the B-Galaetosldøse Assay
One of the classical techniques for the rapid detection of lactose-fermenting bacteria
involves an assay for -galactosidase activity (I c Minor and ben Hamlda, 1961). The induceable
enzyme cleaves lactose Into glucose and galactose and Is metabolically associated with lactose
metabolism in coliforms as well as other bacteria. Assays of the enzyme activity have been made
convenient and accurate through the use of a chromogenic substrate, ortho-nitrophenyl
-D-galactopyranos1de (OHPG). This substrate Is colorless, but following enzyme hydrolysis one of
the hydrolysis products, ortho-nitrophenol, is yellow. Dense cell suspensions grown on a lactose-
containing medium to Induce enzyme activity, can be mixed with ONPG and a yellow color con be
detected within 20 mm to 1 hr. General surveys for enzyme activity made In Enterobacterloceae
have indicated virtually all coliform species tested were -galactosIdase positive. Representatives
of genera other than coliforms can be positive as well, Including Arizona (now a Salmonella
species), Shigelia, Vibrlo, Aeromonas , and Posteurella . Some coliforms which normally require
several days to ferment lactose become positive In the ONPG assay within 3 hr (Ic Minor and ben
Hamida, 1961). The concept Is readily adoptable to analyzing fecal coliforms as well (Warren, et
2i• 1978).
A unique application of the -galactosIdase assay as a rapid coliform detection technique was
recently described (Cundell, et al ,, 1980). In this assay, the cell suspension Is first exposed to
isopropylthjo- -.D-galactopyranosIde (IPTG) to Induce -galoctosidase. The induced cells are then
mixed with a fluoresce In-di- -D -galactopyronosIde (FDG) substrate and sprayed onto c i microscope
slide. A drop of silicone oil Is placed onto a coversllp which Is then placed over the cell suspension.
Cells with -golactos1dase activity hydrolyze the substrate releasing the fluorescent dye,
fluoresceiri, whIch becomes concentrated into the silicone droplets. When examlt4ed with a
fluorescence microscàpe’ ‘the number of fluorescing droplets can be counted and equated to the
relativi number of ceIl ’wlth -galoctosidase activity.
-------
24
This procedure holds promise since it is rapid and can be automated. However, high cell
densities are required (106 cells/mI) and there is the possibility of false positive reactions from
noncoliform bacteria.
Detection of Coliforms by Gas Chromatographic Techniques
Tests for detecting both total and fecal coliforms in water samples have been developed using
the measurement of metabolic end-products such as ethanol (Newmcinn and O’Brien, 1975). The
basic procedure involves the inoculation of a water specimen into a defined mineral salts medium
supplemented with organic digests (casamino acids, yeast extract) and lactose. Samples are
incubated at 37°C for total coliforms and 44.5°C for fecal coUforms. Samples are removed from
the culture medium at 30 mm intervals and analyzed by gas chromatography (GC) for ethanol. The
time of first ethanol production is quantitatively related to the original indicator density in the
water sample. In the original study, best repeatable results occurred when the initial cell density
exceeded five total or 50 feccil coilforms/mI. Minimum detection times were about 9 hr. The
presence of coliform bacteria was confirmed in 79 out of 80 water samples positive for ethanol
production. Several questions remain from the orIginal 1975 study. For example, the authors do
not explain why an incubation temperature of 37°C was used for total coliforms. Increased
sensitivity achieved by concentrating cells by filtration is mentioned by the authors but results of
field trials are not reported. It is also not clear if non-coliform organisms are capable of producing
ethanol In the relatively nonselective broth employed. Any application to potable water supplies
seems doubtful due to the relatively poor detection sensitivity of this procedure.
The concept of using gas chromatography to detect metabolic products produced by indicator
organisms does, however, have merits because of the potential for automation. It Is clear, though,
that further work is necessary before this concept becomes reliable as a sensitive means to detect
coilforms In potable waters.
Detection of MiCrObial Blomass Through Quantification of CeHulor AlP Levels
Numerical estimates of various cellular components have been explored as an lndI!ect means
of estimating the numbers of microbes in natural ecosystems. Of those chemical analyses
examined, measuring the odenoslne-5’-trlphosphote (ATP content) seems to be one logical method
of enumerating viable cells. This technique shows promise sInce ATP, a natural substance, Is
universally present in all living cells and Is rapidly decomposed enzymoticauly In non-living cells.
-------
25
The major drawbacks to this technique are that ATP extraction techniques vary in their
efficiency of recovering non-hydrolyzed AlP molecules and the high cost of the instrument. The
relationship between ATP content of a sample and cell number is based on an estimate of specific
numbers of AlP molecules per cell. The growth rate of the bacterial population and the
physiological state of the cells will also influence the intracellular concentration of AlP.
For analyses involving drinking water, samples will require concentration. One study
attempted to automate the firefly luciferase ATP assay and apply the procedure to quantitation of
bacteria in potable water (Picciolo, et al., 1981). The major problem encountered was achieving a
rapid and convenient means to concentrate the water and provide sufficient numbers of undamaged
cells (lOs) to meet the detection limits of the assay. A hollow fiber membrane filtration device
provided promising results. Unfortunately, the study was terminated prior to automating the AlP
bioossays with a flow system. Field evaluations were also not conducted.
The firefly luciferase enzyme bloassays for quantitating AlP do offer promise for a rapid
estimation of biomass in aquatic samples. Suitable filtration techniques apparently are available
and the question of adaptability of the technique awaits field trials on potable water supplies. One
obvious limitation now is not In the technique, but, rather In the concept of accepting an estimate
of total blomass In water as a substitute for water quality now given by the conventional coliform
analysis. In addition, the Intracellular levels of ATP may vary considerably within individual taxa
and within species. With bacteria, the coefficient of variation in the ratio of cell carbon to AlP
may be as much as 129 percent (Karl, 1980).
Limukis Amoebocyto Lysate Assay to Detect LIpopolysacch side
In 1968 Levin and Bang reported the clotting properties associated with blood cells from the
horseshoe crab ( Ijmulus ) when they were mixed with lipopolysaccharides or with gram-negative
bacteria. This assay has since been widely utilized to quantitate the amount of lipopolysaccharide
(LPS) present in pharmaceuticals, waters, and other substances (Jorgensen and Alexander, 1981).
The assay requires about 2 hr to perform and Is capable of detecting as little as one picogram of
LPS which corresponds to about tOO cells/mi. Two assays are available, the clot method and the
turbid suspension method. The latter is more sensitive and amenable to automation,
Several studies have been conducted to determine whether the LAL assay results con be
correlated quantitatively with conventional microbiological measurements. In one study, Evans,
-------
26
al., (1978) examined the bacteriological counts in a high quality raw water source and found excel-
lent correlation coefficients of 0.91 between coliform numbers and bound endotoxin and 0.95
between heterotrophic bacterial numbers and bound endotoxin. Bound endotoxin represents the
portion of LPS which sediments following centrifugation at 12,000 x g for 10 mm (Evans, et al.,
1978). Jorgensen, et ol., (1976) indicated that the LAL assay holds promise as a quick and efficient
procedure to quantitate bacteria in renovated water and In finished drinking water supplies.
However, no comparisons of LAL determination were mode with biological counts thus making It
impossible to evaluate the claims of this report. Later, however, the same group studied the
performance of LAL for quantitating bacteria in reclaimed wcjstewater (Jorgensen, et ol., I 979).
Correlation coefficients were only 0.620 between standard plate count bacteria and bound
endotoxin. Total coliforms and bound endotoxin exhibited a correlation coefficient of 0A 19. The
conclusion reached indicated the Limulus assay is not an adequate predictor of the sanitary quality
of reclaimed or highly treated wastewater.
Recently, a commercially available instrument was described for the automated analysis of
LPS using the LAL assay (Jorgensen and Alexander, 1981). The assay Is based on an hcrease in
adsorbance with time of the Limulus lysate-endotoxin mixture. The assay was developed using
purified endotoxin from E. coIl. Automated detection was based on the time a sudden optical
density increase was recorded by the instrument. This time ranged from 10 mm with 100 ng of
endotoxin to about 30 mm with 0.01 ng, the minimum level detected in the automated assay. The
instrument is capable of analyzing up to 176 samples in one assay. The instrument will probably be
widely used in a variety of clinical and pharmaceutical situations where the presence of bacterial
endotoxin will be used as an indication of microbial contamination of a product.
It is clear that the technology for rapid and automated LAL assay is available. Conclusive
studies are still needed on a national basis before one can evaluate or justify potential applications
of the LAL assay to the routine testing of potable water supplies.
Measuring the Heterotrophic Bacterial Population In Drkiklng Water
The National Academy of Sciences recognized the value of a viable plate count procedure In
the routine examination of bacterial quality In drIn’ ing water (Drinking Water and Health, 1977). The
relative size of the heterotrophic bacterial count determined by any reasonable standard protocol
can be invaluable in checking the efficiency of water treatment and In assessing possIble water
quality deterioration in the distribution system (Geldreich, et ., 1978).
-------
27
A renewed interest in the viable plate count (“standard plate count”) came as a result of an
initial 500 organisms/mI limit in the proposed Interim Primary Drinking Water Standards. Although
this limit was deleted from the regulations, much information has been summarized on the signif i-
conces of this microbial population in drinking waters (Geidrelch, 1973; Geldreich, et [ ., 1972;
Geidreich, et c i., 1978). There are two major Issues In this section. First, how should the
heterotrophic bacterial populations be counted and should thIs population become part of the regu-
lations and carry a specific maximum contaminant level (MCL)?
The so-called “SPC” or standard plate count population Is currently defined operational by
those bacteria capable of growth within 48 hr Incubation at 35°C on plate count agar (American
Public Health Association, 1980). The acceptable method of enumeration is by a pour plate
procedure. Various criticisms have been made against the pour plate because the hot molten agar
Inactivates a significant portion of the cell population (Klein and Wu, 974). Other studies have
Illustrated that longer Incubation periods at lower temperatures allow for higher viable counts than
those obtained with the current pour plate method (Reasoner and Geldreich, 1981). The composition
of the medium also Influences cell numbers (Reasoner and Geldreich, 1981). in addition to pour
plate procedures, spread plate and membrane filtration has also been used to enumerate th. viable
cell component in drinking water. Taylor and Geldreich (1979) developed a new method used with
membrane filtration involving a 48-72 hr Incubation at 35°C. Counts are generally higher than
those obtained by the pour plate procedure and the filtration provided greater flexibility and
convenience in processing water samples of larger volumes than the I ml allowed by the pour plate
procedure.
A choice will have to be made as to which medium, temperature, and procedure should be
used to enumerate bacteria in potable waters. The detection of ID- or tOO-fold increased levels of
bacteria following eight days or longer incubation periods cit 25°C seems to have little relevance to
the routine monitoring of water for bacteriological quality, treatment efficiency, etc. In the case
of monitoring water supplies, more Is not necessarily better If time is an element of importance.
Viable plate counts should be conducted to provide maximum Information In the shortest time and
with the smallest cost as possible. Greater efforts are needed to concentrate on these needs and
Information of a comparative nature must accompany and justify these developmental efforts.
Many will agree to the general significance provided by monitoring standard plate count bacteria
-------
28
using conventional methods (Geldreich, ef al., 1978; Geldreich, et al., 1972; Seidler, et al, 198Gb).
If the methods are changed to more efficiently measure the viable cell population, we must be
certain that comparable information is being provided by the new technique. Thus, is there more
meaning to a plate count of I 0 4 /mI obtained after 12 days incubation at 20°C than a count of
l0 2 /ml obtained in 48 hr at 35°C?
One of the most important problems associated with elevated numbers of standard plate
count (SPC) bacteria in water is their ability to interfere with the detection of coliforms. Standard
plate counts In excess of 500-1,000 ml can mask or prevent the detection of coliforms in
contaminated water (Geldreich, et al., 1972; Geldreich, et at., 1978). Seidler, et al., 1981 have
illustrated that coliform recovery from water containing 500 SPC or more per ml give a geometric
mean cotiform count of 2.2/100 ml while in fact, over 14 coliforms/l00 ml were present.
The presence of large numbers (over 500 ml) of SPC organisms signifies deteriorated finished
water quality. High SPC numbers can result from changes in pressure which dislodge microbes
from dead-end locations, or from surfaces of distribution pipes (Allen, et al., 1980). A significant
portion of the SPC population in distribution waters were also multiply antibiotic resistant and
could pose a health threat to the compromised patient (Armstrong, ef al., 1981). Numbers of SPC
organisms appear to be readily controlled by the level of residual chlorine in water (Geldreich, ef
al., 1978). Data published by Geidreich, et al., (1975) indicate that a free chlorine residual of
0.1 mg/I or greater is generally effective in limiting bacterial numbers to less than 500 ml.
The question still remains as to whether an “SPC” measurement should be added to the
regulations of the Safe Drinking Water Act. A suitable technique exists which can provide counts
In 48 hr (Taylor and Geldreich, 1979) while the standard coliform enumeration procedures require
48 to 96 hr. A comparative cost analysis of microbiological monitoring techniques indicates the
SPC is less expensive to perform than coliform analyses ($2.49 for SPC vs $2.77 for MF and $5.22
for MPN) (Geldrelch, personal communication). Guidelines for measuring SPC bacteria should be
offered to water supply operators and formulated so as not to contain specific density limitations.
A variety of factors such as source water type, characteristics of the distribution system and
seasonal influences will all affect the number of SPC bacteria in any system. The committee
recommended that monitoring of SPC numbers be mandatory but no density limitation be
stipulated. The following explanatory information is given as a means of defining the significance
-------
29
of SPC numbers: 100 CFU/ml, achievable goal for all systems; 100-500 CFU/mI, anticipated
during seasonal increase or at certain locations within the system (dead-end locations); 500, poor
microbiological quality. If the results of a single analysis show a 5- to tO-fold increase over
commonly achieved numbers, the cause should be investigated. Greater than 10-fold increases
should be reported to the regulatory agency. The choice of methods should be left to the
monitoring personnel, but the same method must be used consistently. Any method outlined in
Standard Methods or an equivalent procedure could be used as long as incubation conditions are
L 8 hr at 35°C. The number of samples analyzed for SPC bacteria should correspond to the number
required for total cotiform analyses.
Aeromonas, An Indicator Orgcmism or An Interference with Potthle Water Analyses
Aeromonas are similar to the Enterobacteriaceae in their facultatively anaerobic
metabolism. However, unlike the conventional definition of a coliform organism, Aeromonas are
pokirly flagellated and cytochrome oxidase positive. The latter two traits were thought to be
taxonomiccilly significant in the early classifications which placed Aeromonas In the
Pseudomonadaceae group. This mistaken taxonomic grouping may have contributed to the failure
of many sanitary and environmental microbiologists to recognize the significance of Aeromonas us
a causative agent of infections in humans and animals.
Much information is now available to assess the taxonomic, ecological and clinical signif-
Icance of Aeromonos . The genus is placed in the family Vlbrionaceae due to its facultatively
anaerobic metabolism, polar flagellaflon and oxidase activity. The taxonomy of the species is in a
dynamic state with one of the latest recommendations recognizing the species A. solmonicida , A.
hydrophila , and A. sobria (Popoff and Veron, 1976). Both of the latter species cause human
infections while A. salmonicida is a pathogen of cold-blooded animals. There is genetic
heterogeneity In both of the human pathogenic species indicating eventual addition of new biotypes
and/or species will be In order (Popoff, etal., 1981).
Aeromonas are among the most common single species present in surface and potable
drinking water supplies (Bragow and DUPreez, 1979; LeChevallier, et al., 1980; Seidler, et al.,
l980a). Cell densities are strongly influenced by the temperature and trophic level of the water
(Lupo, ef al., 1977; Rlppey and Cabelll, 1980; Seldler, et al., 1980a) Illustrating the capacity of
Aeromonos species to multiply under favorable environmental conditions.
-------
30
Once believed to be an opportunistic pathogen of low virulence, Aeromonas species are not
recognized as a primary pathogen (Ljungh, et at., 1977; Daily, et al., 1981). Gastrointestinal and
soft-tissue infections often result from water related accidents and/or the consumption of con-
taminated water (Hanson, et at., 1977; Joseph, ci., 1979; Cumberbatch, et al., 1979). Virulence
associated factors produced by at least some Aeromonas strains include hemolysin, proteases,
cytotoxic, and enterotoxic activities (Ljungh, Wretlind, and Wadstrom, 1978).
Aeromonas recovered from the stools of patients with diarrhea were shown to produce
enterotoxln-like activity which induces fluid accumulation In rabbits (Ljungh, et al., 1977;
Cumberbatch, et ol., 1979). The immunological distinctiveness of the Aeromonas enterotoxin and
its mode of action apparently distinguished it from that of enterotoxic E. coil and V. choleroe
(Donta and Hoddow, 1978).
The limited data available illustrate that . sobria is less common in raw surface waters than
is A. hydrophila but A. sobria is more common in clinical specimens (Seidler, et ci., l980a). The
environmental and clinical isolates of both species contain subgroups possessing a variable number
of virulence associated factors. These factors include the production of extracellular enzymes,
cytotoxlcity for tissue culture cells, enterotoxin production, and the ability to adhere to the
surfaces of tissue culture cells (Daily, flj., 1981). Two groups were distinguished by virulence in
mice (LC lO ) and production of cytotoxin activity. Ten of the 15 most virulent strains were
sobria. Aeromonas strains from aquatic sources may possess as many virulence associated factors
as clinical strains. However, more clinical cultures than environmental isolates were virulent in
the test assay (Daily, etal., 1981).
Aerogenic strains of Aeromonas have been labelled nuisance organisms which interfere with
coliform enumeration. These aerogenic strains can produce gas from lactose and typical green
sheen colonies on Endo type media (Lupo, et al, 1977; Grawbow and DuPreez, 1979). Due to these
factors, it has been demonstrated that aerogenic lactose-fermenting Aeromonas biotypes con
adversely influence the enumeration of coliforms. These observations prompted some to
recommend the oxidase test be used to distinguish positive aeromonads from oxidase negative
coliforms. However, it may not be appropriate to consider Aerosnonas as merely a nuisance
organism which “interferes” with coliform detection.
-------
31
Aeromonas is present in high numbers in raw and treated sewage. The elimination of
Aeromonas from sewage coincides with the elimination of S lmonelIa and other pathogens
(Schubert and Schafer, 197 I). The elevated numbers of Aeromonas in sewage could be useful as a
sensitive indicator of sewage contamination in potable waters. Aeromonas grow in nutrient rich
waters containing organic waste materials (Hendricks, 1978). Consistent with these general
observations, it has been quantitatively demonstrated that Aeromonas densities reflect the overall
trophic state of lakes and ponds (Rippey and Cabelli, 1980). The relative proportions of aerogenic
and anaerogenic Aeromonas (tested with glucose) is proportional to the degree of wastewater
contamination (Schubert, 1976). Thus, in contaminated waters, very few aerogenic strains are
Isolated.
The presence of Aeromonas in drinking water could falsely signal contamination of sanitary
significance because of its ability to multiply, especially at temperatures over 15°C. However, it
must be recalled that coliforms also regrow under similar circumstances and the regrowth problem
Is not unique to Aeromonas (Hendricks, 1978). Arguments used to counter the use of Aeromonas as
an indicator of contamination have one significant flaw. Aeromonas is a potential pathogen and
some strains can cause primary infections in the human population, Including diarrhea. The
synthesis of cytotoxic virulence factors are significant in an organism’s ability to cause disease
since these factors can destroy erythrocytes and other mammalian cells (Liungh, Wretlind, and
Wadstrom, 1978). The cytotoxin activity of Aeromonas seems to be present in 30 to kO percent of
the environmental strains (Seidler, !i q [ ., 1980). This prevalence is greater than that found in
random samplings of Enterobocteriaceae (about I to 3 percent; Donta and Hciddow, 1978).
We agree with the comments of Grawbow and DuPreez (1979). The total coliform count
should be considered mainly as an indicator of proper water treatment and proper maintenance of
sanitary practices within the distribution system. This count should not be regarded as a specific
indicator that fecal contamination has occurred. The presence of bacteria which overgrow
membranes and interfere with coliform measurements indicate inefficient treatment, of tergrowth,
or contamination in the distribution system. The exclusion of one segment of the bacterial
population from thIs measurement ( Aeromonas ) will limit the sensitivity, efficiency, and reliability
of the total coliform Indicator concept. Therefore, Aeromonas should be included in the total
coilform Indicator group and should not be considered as on interference with coliform analyses of
potable water.
-------
32
Fecal Coilforms Indicators of Drinking Water Quality
The use of total coliforms as indicators of fecal pollution has been criticized because some
total coliform bacteria may be of non-fecal origin. Citrobacter, Enterobacter , and Kiebsiella may
originate from such diverse sources as soil, vegetation, or insects (Geldreich, et al., 1962;
Geldreich, et c i i ., 1964). As a result, the fecal coliform test has been advocated as being the best
means of determining the sanitary quality of water (Dufour, 1977).
Fecal coliform tests which rely on gas production in EC broth or blue colony development on
m-FC agar at 44.5°C are not necessarily specific for organisms originating from fecal sources of
pollution (Dufka, 1977; Grabow, et al., 1981). Only membrane filter and fermentation tube tests
which Include a test for indole production allow the detection of Escherichia coli (Delaney, et al.,
1962; Mara, 1973). Such tests which specifically detect E. coli are the safest methods specific for
recent fecal contamination where non-point sources of pollution impact receiving waters.
A series of factors must be considered when determining the relative merits of fecal
coliforms, E. coli or total coliforms as indicators of drinking water quality. E. coIl is found in
most human fecal samples and is the most numerous cot iform occurring at approximately
io8 cells/gram (LeClerc, et al., 1977; Dufour, 1977). Other coliforms may also be found in human
feces. Enterobacter, Citrobacter , and Klebsiella were found at I0 to 106 cells/gram in the fecal
flora of 30 adults (LeClerc, 1977). The incidence at which these coliforms were found In the fecal
samples varied, with Citrobacter being present in 66 percent of the samples, Kiebsiella in
50 percent of the samples and Enterobacter in 10 percent of the samples. While E. coli may be
expected to outnumber other coliforms in feces by 100 to I, this is not the case for sewage. In
samples of raw sewage, primary treated effluent and secondary treated effluent, Klebsiella,
Cltrobacter , and Enterobacter species may comprise 65 percent of the coliforms detected (Duf our,
1977). In other aquatic environments total coliforms greatly outnumber fecol coliforms (Cohen and
Shuvol, 1973; Gallagher and Spino, 1968; Smith, et gj., 1973).
To assure adequate safety of drinking water supplies, an indicator must be monitored which is
present in excess of any pathogen. Such an indicator should provide a high margin of safety In any
assessment of drinking water potability. Since coliforms are representative of any type of
contamination (treatment failure, secondary contamination) that would be expected in drinking
water and outnumber fecal coliforms in most biologically hazardous materials, total coliforms are
a better indicator of water potability than are fecal coliforms.
-------
33
Coliphoge Indicators of Fecol Pollution
The rapid detection and enumeration of coliphoge has been used as an indicator of fecal
contamination. A quantitative relationship was established between coliform bacteria and their
coliphages in 150 raw water samples (Kenord and Valentine, 197k). More recent studies indicated
that the coliphage procedure, ARCAT® (A Rapid Coliphage Analysis Technique) is easy to
perform and gives results within 6 hr (Isbister, et ci., 1981). The quantitative relationship between
coliphages and E. coil was confirmed in 12 locations where sampling of natural waters was
conducted. The original testing procedure was simplified and the sensitivity of the ARCAT was
improved (Scott, et flj., 1979) but until recently the test was not sensitive enough for applications
to potable water. A refined ARCAT method using positively charged Zeta Plus 60S filters for
bacteriophage concentration increased the sensitivity of the technique to I phage/ 100 ml of water
(lsbister,etal., 198 1).
The standard ARCAT procedure involves mixing I ml of a thawed cell suspension of an E.
oII C culture with 5 ml of the test wafer and adding the mixture to molten agar. The contents
ore mixed and poured Into a petrl dish and Incubated at 35°C for 6 hr. The plates are set up as four
replicates, each containing 5 ml of the test water. The number of coliphoge in the sample is
calculated as: total plaque forming units (pfu) on four plates (20 ml sample) x 5 = pfu/lOO ml. The
increased sensItivity of the techntque’wcs accomplished by concentrating phage on filters followed
by their eluflon Into a small volune of fluid. The sensItivity of the technique was also Increased by
an “amplification technique.” The amplification amounts to an enrichment procedure whereby the
host . coil C Is Incubated with the water sample long enough for one round of Iytic phoge growth
to occur prior to the plating and enumeration of pfu.
The assay system has not been used in field studies to monitor coliphoge In potable water
supplies. Comparisons of coliphoge recovery with densities of indicator organisms using the refined
ARCAT were based on a limited number of samples collected from the Potomac River over a
20-day period. During this perIod, eIght samples of raw Potomac River water were analyzed and
numbers of total coliform and collphoges recovered exhibited a correlation coefficient of 090
(Isbister, et al., 1981). However, of these eight samples analyzed, four hod no total coliforms
present (one lob occident, three were zero) so the value of the correlation coefficient for raw
water appears to be based on four determinations. Fecczl coliform counts determined In parallel on
-------
34
these samples exhibited higher counts than total coliforms in four samples, a theoretically
impossible result. Therefore, some very basic research must be conducted to determine if
coliphage can be used as surrogates for animal viruses in determining the efficiency of animal virus
destruction or removal in the treatment process.
The costs of the ARCAT and standard MF coliform tests were given as follows (based on 1980
costs): Standard ARCAT, $2.95; Zeta concentration ARCAT, $4.12; and Standard MF, $3.24
(Isbister, etal., 1981).
Automated Electrical Impeckmce Test Far Monitoring Fecal Coilforms
Fecal coliforms growing in a selective lactose broth Incubated at 44.5°C generate a change in
electrical impedance when the cells reach 106 -l0 ml (Silverman and Munoz, 1979). Changes in
the impedance ratios can be automatically measured with a “Bactometer”, model 32 (Bactomatic,
Inc., Palo Alto, CA). Although this test system is specifically developed to monitor fecal and not
total coliforms, there may be special cases where fecal coliforms should be monitored in drinking
water (Reasoner, et al, 1979).
Impedance ratios were measured in a pair of tubes, one inoculated and the other a sterile
control. The Bactometer can measure up to 32 sample and reference pairs and readings require
only three seconds to complete. Automatic readings are recorded on a strip chart and processed in
c i computer. Detection time corresponds to a definite and continual increase In the Impedance
ratios which occur over three or more hours after the tubes are inoculated. Results illustrate that
200 fecal coliforms are detected in 5.8 to 79 hr and I is detected in 8.7 to 11.4 hr. Variations in
detection times for a specific number of fecal coliforms were found among different sewage
treatment plants and also with samples collected at one location over a one month period. This
suggested that standard curves should be periodically developed for determining the most accurate
relationship between impedance and numbers of fecal coliforms.
R ld Most-Probthle-Nuniber Fecal Coilform Enumeration Procedure
In this procedure the appearance of growth (turbid tubes) is the sole criterion used for
designating fecal coliform positive samples (Munoz and Silverman, 1979). The test requires only
18 hr Incubation at 445°C. Standard methods procedures now require 48 to 72 hr. Growth can be
determined automatically through changes In electrical impedance or simply by visual examination
-------
35
of the tubes. The medium used is a modification of that developed for the 7-hr fecol coliform
count involving membrane filtration (Reasoner, et al., 1979). Recoveries comparable to the
Standard Method’s MPN procedure were obtained In samples containIng 10 to 1O 7 fecol coil-
forms/lOO ml. Overall 87 percent of the positive (turbid) tubes were confirmed as containing fecal
col Iforms.
Rq 1d Seven-Hair Fecal Cofifarm Test By Membrane Ffttraticn
A procedure Is available which uses conventional membrane filtration to enumerate fecal
coliforms after seven hours of Incubation (Reasoner, ., 1979). The chosen incubation
temperature of 41.5°C was based on temperature gradient studies which shows this to be the
optimum temperature for quantifying feed coliforms In a 7-hr period. The medium contains
lactose and mannitol, organic digests, two selective agents and two pH indIcators. Coliform
colonies appear yellow against a purple background. Counts are made after seven hours using a
lox or l5X power dissecting microscope. Verified fecal coliform càunts from yellow colonies
appearing on the new medium designated rn-i-h EC were equal to or greater than the conventional
standard methods m-FC ogar counts recorded after 24 hr of Incubation.
One very interesting observation revealed that a large number of Isolates without yellow
colonies on rn-7-h FC were also fecal coliforms. Substantial numbers of fecal cot lforms,
therefore, go undetected on the standard m-FC agar as well. Since verified fecal coliform counts
on the new medium are at least as high as the conventional medium but can be counted
substantially sooner, the new medium should serve well in emergency conditions where a fast
estimate of water quality is needed.
The committee felt that at present no other detection procedure involving any of these
bloassays, chemical analyses, or automated electronic systems were ready to replace “conven-
tional” Indicator organisms analyses.
TURBIDITY
Turbidity Is a general optical property of water which can be measured by estimating the
degree of light scattered by particles in suspension. The amount of light scattered is a complex
function of particle size, shape, and density. Turbidity in drinking water is measured by the
Nephelometric Method in accordance with American Public Health Association recommendations
-------
36
(1980). The recommendations also specify instrument sensitivities which must be less than
2 percent error tolerance (detection of 0.02 NTU difference in water of less than I NTU). It seems
doubtful that the commonly available nephelometers are capable of such precision.
Turbidity is one of the primary regulated parameters for potable drinking water served to the
public (U.S. Environmental Protection Agency, 1975). The monthly average turbidity shall not
exceed one nephelometric unit and readings must be recorded on a daily basis. The regulations
apply to wafer systems using surface water sources in whole or in port. The control of turbidity in
drinking water is justified since the goal of the Act is to deliver adequately disirifected water to
consumers (LeChevallier, et al., 1981).
The intent of the regulation was to promulgate an MCL of I NTU. Since the regulations
specify “one” (not 1.0), monthly averages must technically be allowed to reach 1.49 (Bill Mullen,
personal communication) before a supplier can be considered out of compliance.
Water samples are taken from representative entry point(s) to the distribution system. The
regulations do not specify exactly where this is to be for various types of water treatment systems.
Significant problems are associated with turbidity measurements because of technician
carelessness, poor instrument calibration and instability of standards, and changes in turbidity
which occur during sample storage (Geldreich, 1977; McGirr, 1974; I-loch, 1972). Despite these
shortcomings, turbidity has been shown to be an excellent indicator of water quality (Geldreich,
1977; LeChevallier, etal., 1981).
In 1979, the Region X office of the U.S. Environmental Protection Agency launched a
program to evaluate the precision and reliability of nephelometers used in the Pacific Northwest by
establishing a turbidity measurement quality control program. Over 140 water suppliers in Oregon
participated in the study (Bill Mullen, personal communication). Several major conclusions were
drawn from the study. Commonly used instruments (Turner,’ Hach, and HF/Fisher) are susceptible
to operator error. 1-fach nephelometers use a secondary standard for instrument calibration. The
Hach kit 10 NTU standard is subject to deterioration. Nearly 80 percent of the HF nephelometers
(especially DRT- 15 units) were found to be malfunctioning. The quality control program was
successful in increasing operator awareness to potential problems of inaccurate turbidity measure-
ments. A commercially available turbidity standard is marketed by Amco International or
formazin solutions can be used as turbidity standards (American Public Health Association, 1980) to
establish and evaluate quality control programs for turbidity measurements.
-------
37
When a sample exceeds the MCL (one) a repeat sample is collected In less than I hr. Results
from the repeat sample are used for calculating the monthly average. It is not clear why repeat
sample averages are used to calculate the average. One possible reason may be to permit greater
flexibilities in the MCL standard.
A serious problem with once a day turbidity readings occurs in smaller water supplies which
do not chemically treat or filter the surface water prior to disinfection. Storm events bring runoff,
turbidity, and organic matter which increase the chlorine demand of the water. If the chlorine is
applied at a constant dose, the demand can exceed the dose and improperly disinfected water
enters the distribution system (LeChevallier, et al., 1981). Consumers face the threat of water
contamination with each runoff event. This is very evident with small systems where the
regulations do not provide these consumers the same protection as larger water supplies provide
their customers. Since small systems cannot afford complete water treatment, the turbidity can
come and go during the night and carry a slug of inadequately disinfected, turbid water to the
consumers during the morning peak water use. With less than 1,000 population served, only one
sample per month is collected for microbiological quality. This situation is nothing short of
scandalous. The regulations should address this problem from the point of providing consumers
greater microbiological protection during times when the turbidity MCL is exceeded. This might
be reflected in a requirement to increase microbiological sampling frequency at strategic locations
In the distribution system following major runoff events or peaks in turbidity occurrences.
Investigators have suspected that modest levels of turbidity can Interfere with collform
detection by membrane filtration (MF). Fryt (1979) noted that differences between MF and MPN
results tended to Increase with turbiditles over 1.8 NTU. Geidreich, et at., (1978) noted that more
coliforms could be detected in waters of lower turbidity (1-5 NTU) than with high turbidity.
LeChevailler, et ci., (1981) developed procedures to quontitate coliform masking on membrane
filters. The incidence of false negative results increased as water turbidity Increased. At
turbidities of 5 NTU, 115 percent of the filters which were Initially free of typical coliform
colonies, were found to be coliform positive. These results collectively suggest that MF is an
undesirable technique to monitor indicator bacteria when the water contains turbidlties greater
than 2 NITU. Serious consideration should be given to monitoring coilforms by the fermentation tube
procedure when water turblditles exceed 2 NTU. At times of high water turbidity, high chlorine
-------
38
demand, and high indicator organism survival (LeChevoliler, et al., 1981) one cannot afford failures
in the indicator organism enumeration technique.
CHLORINE RESIDUAL MEASUREMENTS
Under specially approved circumstances, a public water supplier may substitute chlorine
residual monitoring for up to 75 percent of the monthly required bacteriological tests. However, at
least four chlorine residual measurements are required for each substituted coliform analysis.
When the supplier exercises the option of substitution, chlorine residual measurements must be
conducted daliy. Levels of free chlorine shall not be less than 0.2 mg/I throughout the distribution
system. The specific method by which the chlorine concentrations are determined is not specified
by the National Interim Primary Drinking Water Regulations other than to use one of those In the
13th edition of Standard, Methods (U.S. Environmental Protection Agency, 1975).
A comparative study has been conducted using ten methods described in Standard Methods to
measure total available residual chlorine in different types of water samples (Bendor, 1978).
Precision of the methods was determined by seven replicates and accuracy was compared to the
lodornetric starch titration technique and expressed in terms of percent yield. Analyses were
conducted on various samples inciuding drinking water. lodometric, amperometric, DPD, and flux
monitoring analysis comprised seven laboratory analytical methods while the remaining three
techniques (selective ion electrode, two DPD commercial kits) were amenable to field analyses.
For analyzing chforine residuals in drinking water, all techniques and procedures were found to be
suitable except for one commercial field test kit which gave readings over 30 percent higher than
actual chlorine concentrations. This test kit (Hach CN-66) Is popular In routine analysis in field
situations. In addition to it being Inaccurate, the color wheel on the Hach CN-66 test kit also
fades In the presence of sunlight (Bendor, 978).
One chlorine analysis procedure described In the I 3th edition of Standard Methods (American
Public Health Association, 1971) is no longer recommended. The orthotolidlne method is poor in
accuracy and precision and was deleted from the I 1&th and 15th editions of Standard Methods .
Problems with that technique perhaps should be publicized to water suppliers If Indeed this has not
already been done.
Cautions should be exercised with most of the other commonly used chlorine analytical
detection procedures because of the potential for Interference. With the DPD methods, high
-------
39
monochloramine concentrations interfere and register as free chlorine. This interference is
especially common in the presence of organic contaminants (American Public Health Association,
1980). With the leuco crystal violet method, nitrite and monochloramine in combination, or
oxidized manganese interfere with measuring free available chlorine (American Public Health
Association, 1980). Colorimetric or spectrophotometric methods to measure chlorine are also
compromised by turbidity and colored matter In the water.
COLLECTION AND TRANSPORT OF DRINKINC WATER SAMPLES
The training of personnel to collect, ship, and/or analyze laboratory water samples is one of
the most crucial components of a workable and reliable Safe Drinking Water Act. If the water
supply is properly engineered and water is efficiently treated and distributed, the occurrence of
false positive coliform samples due to careless collection or Incorrect processing of specimens
becomes inefficient and intolerable. If samples are not collected, shipped, and processed In an
efficient and professional fashion, the entire monitoring and MCL requirements of the Safe
Drinking Water Act are doomed to failure. For these reasons various training manuals, short
courses, and certification programs have been developed by the U.S. Environmental Protection
Agency and other associations. The committee feels it is beyond the scope of this paper to
critically evaluate all these programs. However, It cannot be overemphasized that these quality
assurance training programs must be retained in order to provide updated and uniform certification
standards for working personnel and training for new individuals entering the water quality field. It
is naive to place this training burden on public colleges and universities because most training
programs are too specific and cannot be justified on a cost analysis because of the limited number
of students served. However, there are special short training courses which are conducted at
public facilities and these are typically funded through registration fees and usually some federal
support. These programs have been very successful and must be continued to maintain
professional, and economic application of methods crucial to the collection and examination of
water specimens.
Sample storage and transport time are issues of current practical concern in potable water
analyses. There are two issues relevant to the allowable time between sample collection and
processing in the laboratory. First, must samples be analyzed when they are returned to the
laboratory in the late afternoon or can specimens be stored overnight In the refrigerator? Second,
is the restriction of a maximum transport time justified for samples shipped through the U.S. mail?
-------
40
Water samples must be placed on ice if they cannot be processed within one hour of
collection (American Public Health Association, 1980). Raw water samples must be held below
10°C and transport time must not exceed six hours. Parallel instructions for potable waters are
not specified and one presumes that this specification applies to those samples as well. Samples
sent by Carrier must be analyzed within 30 hr from collection. Changes which result in a loss of
indicator organism are the most serious problems arising from delays in sample transport and/or
storage.
The allowable limits from sample collection to process time differ from country to country.
Guide! ines recommended by NATO (Committee of the Challenges of Modern SocIety, 1981) indicate
samples should be transported to the laboratory as quickly as possible, ideally within six hours.
Samples should never be examined more than 24 hr from collection. French regulations indicate
analyses should begin within eight hours from collection (Vial and Geoff ray, 1980) whIle Sweden
allows up to 36 hr if samples are maintained at less than 10°C.
Coliform counts appear to vary upward, downward, or not at all following storage of various
samples at refrigeration or room temperatures. in one series of experiments, Vial and Geoffray
(1980) studied the changes In total coliform, fecal coliform, and fecal streptococci in 262 tap water
samples. All but 29 of the samples contained coliforms when detected by membrane filtration.
Water samples were analyzed less than eight hours after collection and again after refrigeration
for 24 hr. The conclusions indicated that numbers of indicator organisms did not significantly
change after the refrigerator storage period. Although the tendency was for 21e hr counts to be less
than the Initial count, the conclusion was that decreases were too small to be of hygienic
significance.
Studies from the Public Health Laboratory Service in England (1952, 1953) lead to a different
conclusion concerning 24 hr storage at refrigeration temperatures. To increase precision, analyses
were conducted with 70-tube MPN tests involving seven sets of two-fold diminishing concentra-
tions of sample. In all, 151 samples were analyzed soon after collection and subdIvided In two.
One split sample was stored in the dark at room temperature (16-23°C) whIle the other was
refrigerated for 24 hr. Criteria for defining a difference in counts after storage were based on a
change of 9 or more positive tubes after the storage period. This change corresponds to a doubling
or halving In the Initial MPN count.
-------
141
After storage at room temperature, 66 percent of the samples showed no change In the total
cot if orm count, 19 percent increased, and 15 percent decreased. After refrigeration, 76 percent of
the samples remained unchanged, 17 percent increased, and 7 percent declined in coliform
numbers. There were more increases than decreases and more variations in counts were noted with
room temperature storage. The committee concluded that overnight storage of a water sample,
even In the refrigerator, will lead to a significant change in counts in at least 25 percent of the
samples. However, one should not be concerned with Increases In the coliform count since a single
dilutIon 5 tube fermentation test is not a very precise quantitative measure. The real danger Is in
the complete loss of coliforms, but only I of the 151 collform contaminated samples dropped to
zero following storage.
Waters of differing microbiological quality were examined for Indicator stability following
storage at room and refrigerator temperatures in one other study (Lonsame, Parhad, and Roo,
1967). Recovery comparisons were made with both the fermentation tube and membrane filtration
techniques. The Influence of storage was found to be significantly dependent upon the original
number of coliforms in the sample. Thus, when the total coliform counts were less than 20/ tOO ml,
there was no change in the counts after 214 and 148 hr of storage at room or refrigeration
temperatures as measured by either technique. Although these results appear to differ from those
of the Public Health Laboratory Service, the enumeration precision (15 vs 70-tube MPN) arid
climatic and geographic differences in the two study areas (England and India) might account for
the disparities In coliform persistence.
One of the most serious flaws in the guidelines for sample transport time restrictions Is the
recommendation not to analyze samples held for over 30 hr (Bordner and Winter, 1978). This
creates an unnecessary delay in sample analysis and discarding the sample is a potentially wasteful
mistake. It Is strongly recommended that such samples be processed and efforts made to Initiate a
request for a fresh specimen. If the delayed sample shows coilforms to be present, a repeat
(“check”) sample will already be on its way to the testing laboratory. A study completed by Jon
Standridge of the State Laboratory of Hygiene in Wisconsin (personal communication) Irpd)cated 35
of 1,299 samples received late (after 30 hr from collection) actually contained cotiforms indicating
that potential intermittent contamination problems could have otherwise gone unnoticed were
these samples discarded as recommended.
-------
L 2
One solution to the Instability of microbial counts is to find a suitable way to preserve
samples during transport. In this context studies have been conducted to evaluate the role of
specific chemicals on stabilizing coliform counts (Brodsky, et al. , 1978; Toombs and Connor, 1980).
By the addition of sulfonilomide or sodium benzoote to buffered water, it becomes possible to
stabilize E. coil or Enterobacter cloacae for t 8 hr. These stabilized cell suspensions are used In
proficiency testing of water samples in public health laboratories. Perhaps similar stabilizing
agents could be used to maintain counts of indigenous coliforms in water samples. Sensitivity to
various bocterlostatic agents differs among coliforms and the benzoote and sulfanilomide did not
maintain the viable counts of all coliform genera in one published study (Toombs and Connor, 1980).
SUMMARY
Review of the standard microbiological methods for routine water analyses revealed the poor
sensitivity of techniques In detecting coliforms. The problems occur with both the fermentation
tube and MF techniques. 4 eseorch efforts are currently underway to improve coliform detection
by both methods.
The presence-absence fermentation tube test holds promise as a sensitive and convenient
alternative to the standard fermentation tube test. Comparative field testing should be
encouraged.
The definition of a coliform has been changed In an attempt to encourage further innovation
in new procedures f or their detection and enumeration. The change simple indicates that the group
Is a taxonomic assemblage of related gram-negative genera and species which do not have
discontinuous ecological boundaries and do share the ability to ferment lactose.
Concern was expressed with the failures of the ME coliform detection technique which
results when suspended matter enters a distribution system. The concern applies to those supplies
which utilize surface waters which are not treated, only dlsinfected. It was suggested that
bacteriological analyses be conducted as soon as possible when the turbidity reading exceeds
5.0 NW. The analysis must be conducted by the fermentation tube technique. Because of the
specific problems associated with turbid drinking waters, Increased bacteriological surveillance was
recommended when the turbidity MCL Is exceeded.
The usefulness of the standard plate count In documenting changes In distribution water
quality and water treatment efficIency was discussed and a regulation was recommended to
-------
43
enumerate this microbial group. Poor quality drinking water would occur whenever the standard
plate count exceeded 500 ml.
if It is necessary to ship the specimen via commercial carriers or mall, the current 30 hr
maximum transport time should be retained until suitable reagents are found which can stabilize
microbial counts.
RESEARCH RECOMMENDATIONS
I • There is a definite need to develop modern, rapid, and automated coliform analyses. Non. of
th. current alternatives are satisfactory from a point of recovery, expense, and time.
2. There should be a continued development of new media for conventional MF and MPN
coliform analyses. It must be recognized that some current media and techniques suppress or
underestimate coliform numbers and steps need to be mode to counter the.. problems.
3. Comparative studies are required to sort out the coliform definition by comparing gas
production, ONPH and oxidase reactions for verification plus other metabolic/physiological
processes. In short, encourage new and Innovative techniques with the assurance that the
same indicator group Is being detected with new conventional procedures.
4. Field testing of current coliform techniques should be compared to that of th P-A test, It Is
essential, however, that media known to be inadequat, for conventional fermentation tube
analyses NOT be used In the development of the P-A test.
5. StudIes are needed to determine whether pathogen. are Injured by the some processes as
coliforms In drinking water; if they are, are pathogen. itllI vjrulent?
6. AdditIonal studies are required to determine th. extent of bacterial attachment to portia .
ulates arising from Intake turbidity carbon particulate., floc breakthrough and slough off
from dlstributlàn mains arid the magnitude of bacterial underestimation. these problems
Impose.
7. More Information Is also needed concerning the factors responsible for the suppression of
indicator organisms In water. These include injury, thi r.s.nc. of excessive numbers of SPC
bacteria, and others. It Is crucial to determine whether similar factor, influence coliform
detection to the same degree In different regions of the country. For example, cell injury Is
a big problem in hard waters, and higher SPC numbers are found predominantly in warmer
climates.
-------
44
8. A serious problem arises in non-compliance when sample transport times are exceeded.
There is a ncessity to find reagents which can stablize coliform counts during periods of
sample storage or transport.
LITERATURE CITED
Adams, R. B. 1957. Comparison of standard dilution and membrane filter methods. J. Am. Water
Works Assoc. 49:1452-1458.
Allen M. J., E. E. Geidreich, and R. H. Taylor. 1980. The occurrence of microorganisms in water
main encrustaceans, p. I 13-135. In: American Water Works Association Water Quality
Technology Conference ProceedingC Advances in laboratory techniques for quality control.
Am. Wat. Works Assoc., Denver, Cob.
American Public Health Association. 1913. Standard methods for the examination of water and
sewage , 2nd ed. American Public Health Association, Inc., Washington, D.C.
American Public Health Association. 1917. Standard methods for the examination of water and
sewage , 3rd ed. American Public Health Assoc. Inc., Washington, D. C.
American Public Health Association. 1923. Standard methods for the examination of water and
sewage . 5th ed. American Public Health Assoc., Inc., Washington, D.C.
American Public Health Association. 1946. Standard methods for the examination of water and
sewage , 9th ed. American Public Health Assoc., Inc.
American Public Health Association. 1955. Standard methods for the examination of water,
sewage, and industrial wastes , 10th ed. American Public Health Assoc., Inc., Washington,
D.C.
American Public Health Association. 1960. Standard methods for the examination of water and
wastewater , 11th ed. American Public Health Assoc., Inc. Washington, D.C.
American Public Health Association. 1965. Standard methods for the examination of water and
wastewoter , 12th ed. American Public Health Assoc., Inc., Washington, D.C.
American Public Health Association. 1971. Standard methods for the examination of water and
wastewater , 13th ed. American Public Health Assoc., Inc., Washington, D.C.
American Public Health Association. 1975. Standard methods for the examination of water and
wastewater , 14th ed. American Public Health Assoc., Inc., Wash [ ngton, D.C.
American Public Health Association. 1980. Standard methods for the examination of water and
wastewater , 15th ed. American Public Health Assoc., Inc., Washington, D.C.
Armstrong, J. L., D. E. Shigeno, J. J. Calomiris, and R. J. Seldler. 1981. AntibIotic-resistant
bacteria in drinking water. Appi. Environ. Mlcrobiol. 42:277-283.
Bender, D. F. 1978. Comparison of methods for the determination of total available residual
chlorine In various sample matrices. U.S. Environmental Protection Agency. Cincinnati,
Ohio. EPA-600/4-78-OI 9.
Beuchat, L. R. 1978. Injury and repair of gram-negative bacteria, with special consideration of
the involvement of the cytopicismic membrane. p. 210-243. In: Advances in applied
microbiology, D. Per lman (ed.), Academic Press, New York.
-------
45
Bissonnette, G. K., J. J. Jezeski, G. A. McFeters, and D. G. Stuart. 1975. Influence of
environmental stress on enumeration of Indicator bacteria from natural waters. Appl.
Microbiol. 29:186-194.
Bissonnette, G. K., J. J. Jezeski, G. A. McFeters, and D. G. Stuart. 1977. Evaluation of recovery
methods to detect coliforms in water. AppI. Environ. Microbial. 33:590-595.
Bordner, E., J. Winter, and P. Scarpino (eds.). 1978, Microbiological methods for monitoring the
environment. U.S. Environmental Protection Ageeny. Cincinnati.
Bowmer, E. J., and J. A. K. Campbell. 1972. Index of drinking water pollution-total coliform MPN
tests: confirmed test versus completed test. Can. J. Pub. Health. 63:335-358.
Brasweil, J. R. and A. W. Hoadley. 1974. Recovery of E. coli from chlorinated secondary sewage.
Appi. Microblol. 28:328-329.
Breed, R. S., and J. F. Norton. 1.937. Nomenclature for the colon group. Am. J. Pub. Health
27:560-563.
Brodsky, M. H., B. W. Ciebin, and D. A. Schlemann. 1978. Simple bacterial preservaflon medium
and its application to proficiency testing in water bacteriology. Appi. Environ. Microbial.
35:487-49 1.
Busta, F. F. 1976. Practical implications of injured microorganisms in food. J. Milk Food Technol.
39:138- 145.
Butterfield, C. 1. 1932. The selection of a dilution water for bacteriological examinations. J.
Bacteriol. 23:355-368.
Camper, A. K. and G. A. McFeters. 1979. Chlorine injury and the enumeration of waterborne
coliform bacteria. Appi. Environ. Microbiol. 37:633-641.
Chambers, C. W. 1950. Relationship of coliform bacteria to gas production In media containing
lactose. Pub. Health Reports 65:619-627.
Clark, H. F., E. E. Geidreich, H. L. Jeter, and P. W. Kabler. 1951. The membrane filter in sanitary
bacteriology. Pub. Health Reports 66:951-977.
Clark, H. F. and P. W. Kabler. (952. The membrane filter in water quality tests. Am. J. Pub.
Health 42:385-388.
Committee on the Challenges of Modern Society (NATO). 1981. Pilot Study on Drinking Water
Supply Problems. Draft report. 528 pp.
Cumberbatch, N., M. J. Gurwlth, C. Langston, R. B. Sack, and J. L. Brunton. (979. Cytotoxlc
enterotoxin produced by Aeromonas hydrophila : relationship of toxigenic isolates to dkirrheal
disease. Infect. Immun. 23:829-837.
Cundell, A. M., A. Plsan, and E. FindI. 1979. A rapid method for detection of coliform bacteria.
In Developments in Industrial Microbiology, Vol. 20, pp. 571-577.
DaIly, 0. P., S. W. Joseph, J. C. Coolbaugh, R. I. Walker, B. R. Merrell, D. M. Rollins, R. J. Seldier,
R. R. CoIwelt, and C. R. Llssner. 1981. AssociatIon of Aeromonas sobria with human
Infection. J. Cliri. Microbiol. 13:769—777.
Darby, C. W. and W. L. Mallmann. 1939. Studies on media for coliform organisms. J. Am Water
Works Assoc. 31:689-706.
Davenport, C. V. ,, W. B. Sparrow, and R. C. Gordon. 1976. Fecol indicator bacteria persistence
under natural conditions in an ice-covered river. AppI. Environ. Microbiol. 32:527-536.
-------
46
Delaney, J. E., J. A. McCarthy, and R. J. Grasso. 1962. Measurement of E. coil type I by the
membrane filter. Water Sewage Works 109: 289-294.
Dominick, J. E. and C. J. Lauter. 1929. Methylene blue and brom cresol purple in differentiating
bacteria of the colon-aerogenes group. J. Am. Water Works Assoc. 21:1067-1075.
Donta, S. 1. and A. D. Haddow. 978. Cyfotoxic activity of Aeromonas hydrophila . Infect.
immun. 21:989-993.
Dufour, A. P. 977. Escherichia coil: the fecal coliform, p. 48-58. In: Bacterial
Indicators/health hazards assoàlateTwith water. American Society for testing and
Materials Publication No. 635, A. W. Hoadley and B. J. Dutka (eds.), American
Society for Testing and Materials, PhIladelphia, Pa.
Dufour, A. P. and V. J, CabelIl. 1975. Membrane filter procedure for enumerating the component
genera of the coliform group in seawater. Appi. Microbiol. 29:826-833.
Dutka, B. J. 1973. Coilforms are an inadequate index of water quality. J. Environ. Health :39-
48.
Dutka, B. J. and A. El-$haarawi. 1980. Microbiological water and effluent sample preservation.
Can. J, Mlcrobiol. 26:921-929,
Dutka, B. J. and S. E. Tobln. 1976. Study of the efficiency of four procedures for enumerating
coliforms In water. Can. J. Mlcrobiol. 22:630-635.
Evans, T. M., J. E. Schillinger, and D. C. Stuart. 1978. Rapid determination of bacteriological
water quality by using Limulus lysate :376-382.
Evans, T. M., M. W. LeChevallier, C. E. Waarvick, and R. J. Seidler. 1981a. Coliform species
recovered from untreated surface water and drinking water by the membrane filter, standard
and modified most-probable-number techniques. Appl. Environ. Mlcroblol. 41:657-663.
Evans, T. M. R.J. Seidler, and M. W. LeChevallier. l98lb. Impact of verification media and
resuscitation on the accuracy of the membrane filter total coliform enumeration technique.
Appl. Environ. Microbiol. 41:1144-1151.
Evans, T. M., C. E. Woarvick, R. J. Seldier, and M. W. LeChevaliler. l98 1c. Failure of the most-
probable-number technique to detect coliforms In drinking water and raw water supplies.
Appl. Environ. Microblol. j: 130-138.
Farrell, M. A. 1936. The comparative study of nine presumptive test media. J. Am. Water Works
Assoc.28:6 1 1-622.
Fifield, C. W. and C. P. Schaufus. 1958. Improved membrane filter medium for the detection of
coliform organisms. J. Am. Water Works Assoc. Q: 193-196.
Fujioka, R. S., H. H. Hasimoto, E. B. Siwak, and H. F. Reginald. 1981. Effect of sunlight on
survival of Indicator bacteria in seawater. AppI. Environ. Mlcroblol. 41:690-696.
Gallagher, T. P. and D. F. Spino. 1968. The significance of numbers of coliform bacteria as an
indicator of enter Ic pothogens. Water Res. 1:169-172.
Geldr.Ich, E. E. l 73. Is the total count necessary? Sections VII- I -VIl-2. Proc. A.W.W.A., Water
Qualify Technology Conference, Cincinnati, Ohio.
Geldreich, E. E. 1975. Handbook for evaluating water bacteriological laboratories. U. S.
Environmental Protection Agency, Cincinnati, Ohio.
Geldreich, E. E. 1977. TurbidIty of drinking water and Its relation to bacteriological standards.
Annual Environmental Laboratories Seminar, Springfield, IllinoIs.
-------
47
Geidreich, E. E., M. J. Allen, and R. H. Taylor. 1978. Interferences to coliform
detection In potable water supplies, p. 13-20. In: Evaluation of the microbiology
standards for drinking water, C. W. Hendricks fld.), U.S. Environmental Protection
Agency, Washington, D.C.
Geidreich, E. E., H. F. Clark, C. B. Huff, and L. C. Best. 1965. Fecol coliform-organism medium
for the membrane filter technique. J. Am. Water Works Assoc. 57:208-214.
Geldreich, E. E., C. 3. Huff, R. H. Bordner, P. W. Kobler, and J. F. Clark. 1962. The fecal coil-
aerogenes flora of soils from various geographic areas. J. AppI. Bacteriol. 25:87-93.
Geldreich, E. E., B. A. Keener, and P. W. Kabler. 1964. The occurrence of coliforms, fecal
coliforms and streptococci on vegetation and insects. AppI. Microbiol. 12:63-69.
Geldreich, E. E., H. L. Jeter, and J. A. WInter. 1967. Technical considerations in applying the
membrane filter technique. Health Lab. Sd. k:l 13-125.
Geidreich, E. E., H. D. Nash, D. J. Reasoner, and R. H. Taylor. 1972. The necessity of controlling
bacterial populations in potable waters: community water supply. J.Am . Water Works Assoc.
64:596-602.
Goetz, A. and N. Tsuneishi. 1951. Application of molecular filter membranes to bacteriological
analysis of water. J. Am. Water Works Assoc. 43:943-984.
Grabow, W. 0. K., and M. DuPreez. 1979. ComparIson of m-Endo LES, MacConkey, and Teepol
media for membrane filtration counting of total coliform bacteria In water. Appl. Environ.
Mlcrobol. 38:351-358.
Grabow, W. 0. K., C. H. Hilner, and P. Coabrough. 1981. Evaluation of standard and modified
m-FC, MacConkey, and Teepol media for membrane filtration counting of fecal coliforms in
water. App). Environ. Microbiol. 42:192-199.
Green, B. L., E. M. Clausen, and W. Litsky. 1977. Two-temperature membrane filter method for
enumerating fecal coliform bacteria from chlorinated effluents. App). Environ. Microblol.
33:1259-1264.
GreeT, F. E., R. E. Noble, F. V. Nyhan, and A. E. O’Neal. 1928. The sanitary significance of
lactose fermenting organisms not belonging to the B. coil groups; 7. Mediums and methods. J.
Infect. DIseases 62:556-567.
Greer, F. E. and F. V. Nyhan. 1928. The sanitary significance of lactose fermenting bacterIa not
belonging to the . coIl group. 3. BacterIal associations In cultures containing lactose
fermenting bacteria. rinfect. Diseases 62:525-536.
I-loch, C. C. 1972. UnderstandIng turbidity measurement. indust. Water Eng. 9:18-22.
Hajno, A. A. and S. R. Damon. l955a. New method for detection of coliform organisms. J. Am.
Wafer Works Assoc. 47:631-636.
Hajna, A. A. and S. R. Damon. l955b. Coliform detection in water with a single step technique
using the membrane filter. Pub. Health Rep. 69:58-60.
Hajno, A. A. and C. A. Perry. 943. Comparative study of presumptive and confirmative media
for bacteria of the coliform group and for fecal streptococci. Am. J. Pub. Health 33:550-556.
Hanson, P. G., J. Standridge, F. Jarett, and D. C. Maki. 977. Freshwater wound fnfecflori due to
AeromonOs ydroPhIia . J. Am. Med. Assoc 238:1053-1054.
Hendricks, C. W. 1978. ExceptIons to the coliform and fecal coliform tests, pp. 99-145. In:
indicators of Viruses in Water and Food. C. Berg, (ed.), Ann Arbor Science.
-------
48
Herson, D. S. and H. 1. Victoreen. 1980. Hindrance of coliform recovery by turbidity and non-
coliforms. U.S. EPA, Cincinnati, Ohio. EPA-600/2-80-097.
Holman, W. L. and 0. M. Meekinson. 1926. Gas production by bacterial synergism. J. Infect.
Diseases 39:145-171.
Hurst, A. 1977. Bacterial injury: a review. Can. J. Microbiol. 23:936-942.
Hutchison, D., R. H. Weaver, and Scherago. 1943. The incidence and significance of micro-
organisms antagonistic to Escherichia coli in water. J. Baceriol. 45:29.
Irons, E. E. 1902. Neutral red in the routine examination of water. J. Hyg. 2:314-319.
Isbister, J. D., J. L. Al , R. Foutch, A. DeSouza, R. S. Wentsel, and J. F. Kitches. 1981.
Increasing ARCAT test sensitivity for examination of potable waters. U.S. EPA,
Cincinnati, Ohio, Contract No. 68-02-2914.
Joint Committee of the Public Health Laboratory Service and the Standing Committee of Analysts.
1979. Membrane filtration media for the enumeration of coliforms and Escherlchia coli in
water: comparison of Tergitol 7 and lauryl sulphate with Teepol 610. J. Hyg. (Camb.).
85:181-19 I.
Jordan, H. E. 1927. Brilliant green bile for the detection of the coli-aerogenes group. J. Am.
Water Works Assoc. 8:337-346.
Jordan, H. E. 1932. Brilliant green bile for coll-aerogenes group determinations. J. Am. Water
Works Assoc. 24:1027-1053.
Jorgensen, J. A. and G. A. Alexander. 198 I. Automation of the Limulus amoebocytelysate test by
using the Abbott MS-2 microbiology system. Appl. Environ. Mlcroblol. 41:1316-1320.
Jorgensen, J. H., J. C. Lee, C. A. Alexander, and H. W. Wolf. 1979. Comparison of Limulus assay,
standard plate count, and total coliform count for microbiological assess nent of renovated
water. Appl. Environ. Microbiol. 37:928-931.
Jorgensen, J. H., J. C. Lee, and H. R. Pahren. 1976. Rapid detection of bacterial endotoxins in
drinking water and renovated wastewater. Appl. Environ. Microbiol. 32:347-351.
Kabler, P. 954. Water examinations by the membrane filter and most probable number
procedures. Am. J. Pub. Health 44:379-386.
Karl, D. M. 1980. Cellular nucleotide measurements and application In microbial ecology.
Microbiol. Rev. 44:739-796.
Kenard, R. P. and R. S. Valentine. 1974. Rapid determination of the presence of enteric bacteria
In water. Appl. Microbiol. 27:484-487.
Klein, D. A. and S. Wu. 1974. Stress: a factor to be considered in heterotrophic microorganism
enumeration from aquatic environments. Appl. Mlcrobiol. 27:429-431.
Koser, S. A. and W. C. Shinn. 1937. Aerobic spore-forming bacilli which ferment lactose. J. Am.
Water Works Assoc. 18:328-336.
LeChovallier, M. W., S. C. Cameron, and G. A. McFeters. 1982, A new medium for the Improved
recovery of injured coliforms from drinking water. Abstracts Annual Meet. Am. Soc.
Microblol. Atlanta, Georgia.
LeChevallier, M. W., T. M. Evans, and R. J. Seidler. 1981. Effect of turbidity on chlorination
efficiency and bacterial persistence in drinking water. AppI. Environ. Microbiol. 42:159-167.
-------
49
LeChevalfler, M. W., R. J. Seidier, and T. M. Evans. 1980. Enumeration and characterization of
standard plate count bacteria in chlorinated and raw water supplies. AppI. Environ.
Microbial. 40:922-930.
LeCierc, H., D. A. Mossel, P. H. Trinel, and F. Gavini. 1977. Microbiological monitoring-a new
test for fecol contamination, p. 21-31. In: Bacterial indicators/health hazards associated
with water, A. W. Hoadley and B. J. Dutk 1eds.), American Society for Testing and Materials
Publication No. 635. American Society for Testing and Materials, Philadelphia, Pa.
leMinor, L. and F. benHamida. 1961. Advantages de Ia recherche de Ia -galactosidase sur ceile
de Ia fermentation du lactose en milieu compiexe dons le diagnostic bacterioi ogique, en
particular des Enterobacterlaceae . Ann. Inst. Past. 102:267-277.
Lietch, G. B. 1925. Disturbing factors in the presumptive test for B. coIl. J. Am. Water Works
Assoc. 13:I86- 192.
Lemeshaw, S., and W. Litsky. 1980. Ranking membrane filters. Am. Society Microbiol. News.
46:601.
Levine, M. 1918. Differentiation of B. coil and B. aerogenes on a simplified eosine methylene
blue agar. J. Infect. Diseases 23:4 - 7 —
Levin, J., and F. B. Bang. 1968. Clottable protein in Limulus : its localization and kinetics of Its
coagulation by endotoxin. Thromb. Diath. Haemorih. 19:186-197.
Levin, M. L. 1973. Gas-forming Aeromonas hydrophila Infection in a diabetic. Postgrad. Med.
54:127-129.
Lin, S. 973. Evaluation of coliform tests for chlorinated secondary effluents. J. Water Poliut.
Control Fed. 45:498-506.
Lin, S. D. I 976a. Evaluation of Miwpore HA and HC membrane filters for the enumeration of
indicator bacteria. Appi. Environ. Mlcrobiol. 32:300-302.
Lin, S. D. 1976b. Membrane filter method for recovery of fecal coliforms In chlorinated sewage
effluents. Appi. Environ. Microbiol. 32:547-552.
Lin, 5. 1977. Evaluation of methods for detecting coliforms and fecal streptococci in chlorinated
secondary sewage effluents, p. 113-131. In: Proceeding of the Symposium on the Recovery
of Indicator Organisms Employing Membrane Filters, R. H. Bordner, C. F. Frith, and J. A.
Winters (eds.), Report No. EPA-600 19-77-024. US. Environmental Protection Agency,
Washington, D.C.
Ljungh, A., B. Wretlind, and T. Wodstrom. 1978. EvIdence for enterotoxin and two cytolytic toxins
in human isolates of Aeromonas hydrophlla , p. 947-960. s Proceedings of the 5th
International Symposium on To*Ins: Animal, Plant, and Microbial, P. Rosenberg (ed.).
Pergomon Press, New York.
Ljungh, A., M. Popoff, and T. Wadstrom. 1977. Aeronionas hydrophila in acute dlorrheal diseases:
detection of enterotoxin and blotyping of strains. J. Clin. Microbiot. 6:96-tOO.
Lonsane, G. K., N. M. Porhad, and N. V. Roo. (967. Effect of storage temperature and time on the
col iforms In water samples. J Water Research, Vol. I, pp. 309-316. Pergamori Press, Great
Britain.
Lucking H. E. 1967. Death rate of coliform bacteria In stored Montana water samples. J.
Environ. Health 29:576-580.
Lupo, L., E. Strickland, A. Dufour, and V. CabeW. (977. The effect of oxidase positive bacteria on
total coliform density estimates. Health Lab. Sd. j : 117-121.
-------
50
Mack, W. N. 1977. Total coliform bacteria, p. 59-64. In: Bacterial indicators/health hazards
associated with water, A. W. Hoadley and B. J. Dutka (eds.), American Society for Testing
and Materials Publication No. 635. American Society for Testing and Materials, Philadelphia,
Pa.
MaIlman, W. L. and C. W. Darby. I 94$. Uses of a lauryl sulfate tryptose broth for the detection of
coliform organisms. Am. J. Pub. Health 31:127-134.
Mara, D. D. 1973. A single medium for the rapid detection of Escherichia coil at 44°C. J. Hyg.
71:783-785.
McCarthy, J. A. 955. Comparative coliform densities in water by membrane filter test and by
multiple tube technic. J. Am. Pub. Health Assoc. 45:1569-1577.
McCarthy, J. A. and J. E. Delaney. 1958. Membrane filter media studies. Water Sewage Works
105:292-296.
McCarthy, J. A., J. E. Delaney, and R. J. Grasso. 1961. Measuring coliforms in water. Water
Sewage Works 108:238-243.
McCrady, M. H. 1915. The numerical interpretation of fermentation tube results. J. Infect. Dis.
17:183-212.
McCrady, M. H. 1937. A practical study of procedures for the detection of the presence of
coliform organisms in water. Am. J. Pub. Health : 1243-1258.
McCrady, M. H. $943. A practical study of lauryl sulfate tryptose broth for detection of the
presence of coliform organisms in water. Am. J. Pub. Health 33:1 199-1207.
McFeters, G. A., S. C. Cameron and M. W. LeChevallier. 1982. Influence of diluents, media, and
membrane filters on the detection of injured water-borne coliform bacteria. Appi. Environ.
Mlcrobiol. 43:97-103.
McGlrr, D. J. 1974. Interlaboratory Quality Control Study No. 10. Turbidity and filterable and
nonfilterable residue. Report Series No. 37, inland Waters Directorate, Canada Centre for
inland Waters. Burlington, Ontario.
Means, E. G. and B. H. Olson. 1981.. Collform inhibition by bacteriocinlike substances in drinking
water distribution systems. Appl. Environ. Microblol. 506-5 12.
Morgan, G. B., P. Gubblins, and V. Morgan. 1965. A critical appraisal of the membrane filter
technique. Health Lab. Sd. 2:227-237.
Mowat, A. 1976. Most probable number versus membrane filter on chlorinated effluents. J. Water
Pollut. Control Fed. 48:724-728.
Munoz, E. F. and M. P. Silverman. 1979. Rapid, single-step most-probable-number method for
enumerating fecal coliforms in effluents from sewage treatment plants. AppI. Environ.
Microblol. 37:527-530.
National Interim Primary Drinking Water Regulations. 1978. U.S. Environmental Protection
Agency. Wash., D.C. EPA-570/9-76003.
Newman, J. S. and R. 1. O’Brien. 1975. Gas chromatographic presumptive test for coliform
bacteria in water. AppI. Environ. Microbiol. 30:584-588.
Norton, J. F. and M. Barnes. 1928. Interference of Clostridium welchli with B. coil tests. J. Am.
Water Works Assoc. 19:729-730. —
Olson, B. H. $978. Enhanced accuracy of coliform testing in sCawater by a modification of the
most-probable-number method. Appi. Environ. Microbiol. 36:438-444.
-------
5,
Parr, L. W. 1938. The occurrence and succession of coliform organisms in human feces. Am. J.
Hyg. 27:67-87
Patil, M.D., N. Shivaraman, K. Subbarao, P. V. R. C. Ponicker, N. M. Parhad, and N. U. Rao. 977.
A presumptive synthetic medium for the enumeration of coliforms in water. md. J. Environ.
Health 19:1-15.
Picciolo, G. L., E. W. Chappell, J. W. Darning, R. R. Thomas, 0. A. Nibley, and H. Okr.nd. 981.
Firefly (uciferase ATP assay development for monitoring bacterial concentrations In water
supplies. Project Summary. U.S. EPA Cincinnati, Ohio. EPA-600/32-8 1-04.
Picciolo, G. L., R. R. Thomas, J. W. Deming, and E. W. Chappelle. Environmental applications of
the firefly luciferase ATP assay flow techniques for monitoring the wastewater effluent and
of drinking water supplies, pp. 547-564. In: Proceedings of the Second BI-Annual ATP
Methodology Symposium, SAl Technology, 40 Sorrento Valley Blvd., San Diego, California.
Popoff, M. V., C. Coynault, M. Kiredjian, and M. Lemelln. 1981. Polynucleotide sequence
relatedness among motile Aeromonas species. Curr. Microbial. :lO9-l 14.
Popoff, M. and M. Veron. 1976. A taxonomic study of the Aeromonas hydrophila-Aeromonas
punctata group. J. Gen. Microblol. 94:11-22.
Prescott, S. C., C. A. Winslow, and M. H. McCrady. 1946. Water bacteriology. John Wiley and
Sons, Inc., New York.
Public Health Laboratory Service. 1953. The affect of storage on the coliform and Bacterium coil
counts of water samples. Storage for six hours at room and refrigeration temperature. J
Hyg. 51:559-57 I.
Public Health Laboratory Service Standing Committee on the BacterIological Examination of Water
SupplIes. 1968. A comparison of MacConkey broth, teepol broth and glutamic acid media
for the enumeration of coliforms in water. J, Hyg. (Camb.)66s67.82.
Public Health Laboratory Service, Water Subcommittee. 1952. The effect of storage on the
collform and Bacterium coil counts of water samples. J. Hyg. Q :107 -l25.
Reasoner, D. J. and E. E. Geldreich. 1981. Influence of medium method and incubation time and
temperature on the bacterial count of potable water. Bact. Proc. p. 177.
Reasoner, D. J., J. C. Blannon, and E. E. Geldreich. 1979. Rapid seven-hour fecal coliform test.
App I. Environ. Mlcrobiol. 38:229-236.
Richey, D. 1941. Relative value of 2 percent and 5 percent brilliant green bile confirmatory
media. J. Am. Water Works Assoc. 33:649-658.
Rippey, S. R. and V. J. Cabelli. 1980. Occurrence of Aeromonas hydrophila in limnetlc
environments: relationship of the organism to trophic state. Microb. Eóol. s45-54.
Rose, R. E., E. E. Geldreich, and W. Lltsky. 1975. Improved membrane filter method for fecal
collform analysis. AppI. Mlcroblol. 29:532-536.
Ruchhoft, C. C. 1935. ComparatIve studies of the media for determination of the coll-aerogenes
group In water analysis. J. Am. Water Works Aasoc. z 1732-1745.
Ruchhoft, C. C., J. G. Ka las, B. Chlnn, and E. W. Coulter, 1931. Coli-aerogenes differentiation in
water analysis. J. Bacterlo, j:4Ol-4kO.
Ruchhoft, C. C. and J. F. Norton. 1935. Study of the selective media for coll-aerogenes isolation.
J. Am. Water Works Assoc. 27:1134-1142.
-------
52
Schaeffer. D. J., M. C. Long, and K. G. Jonardcin. 1974. Statistical analysis of the recovery of
coliform organisms on Gelman and Millipore membrane filters. Appi. Microbiol. 28:605-607.
Schiemamn, D. A., M. H. Brodsky, and B. W. Ciebin. 1978. Salmonella and bacterial indicators in
ozonated and chlorine dioxide-disinfected effluent. J. W iter Poll. Control Fed. 50:158-162.
Schiff, L. J., S. M. Morrison, and J. V. Mayeux. 1970. Synergistic false-positive coliform reaction
on m-Endo MF medium. AppI. Microbiol. 20:778-781.
Schubert, R. H. W. and E. Schafer. 1971. Vergleichende Untersuchungen uber die eliminierung von
viren, Salmonellen und Aeromonaden in vorfluter. GWF-Wasser/Abwasser. 112:401-403.
Schubert, R. H. W. (976. Der Nachweis von Aeromonoden der hydrophila-punctato-grupp. In:
Rahmen der Hygienischen Trink wasserbeurteilung. Abi. Bakt. Hyg., I Abt. Orig. B l6l:4 -
497.
Scott, W. M., P. E. O’Neill, M. J. Wilkinson, and J. F. Kitchens. 1979. Evaluation of coliform
bacteria and bacteriophage relationships in assessment of water quality. Final Technical
Report. NSF Grant No. PFR78- 19 116, Atlantic Research Corporation.
Sears, H. J. and J. J. Putnam. 1923. Gas production by bacteria in symbiosis. J. Infect. Diseases
32:270-279.
Seidler, R. J., D. A. Allen, H. Lockman, R. R. Coiwell, S. W. Joseph, and 0. P. Daily. 1980.
Isolation, enumeration and characterization of Aeromonas from polluted waters encountered
during diving operations. AppI. Env. Microbiol. :I0l0-l0l8.
Seidler, R. J., 1. M. Evans, J. R. Kaufman, C. E. Waarvick, and M. W. LeChevallier. 1981. New
directions in coliform methodology, p. 16 1-172. : American Water Works Association Water
Quality Technology Conference Proceeding. Advances in laboratory techniques for quality
control. American Water Works Association. Denver, Co.
Silverman, M. P. and E. F. Munoz. 1979. Automated electrical impedance technique for rapid
enumeration of fecal coliforms in effluents from sewage treatment plants. AppI. Env.
Microbiol. 37:521-526.
Shipe, E. L. and G. M. Cameron. 1954. A comparison of the membrane filter with the most
probable number method for coliform determinations from several waters. AppI. Microbiol.
2:85-87.
Sladek, K. J., R. V. Suslovich, B. I. Sohn, and F. W. Dawson. 1975. Optimum membrane structures
for growth of coliform and fecal coliform organisms. Appl. Mlcrobiol. 30:685-691.
Smith, J., R. M. Twedt, and L. K. Ftanigan. 1973. Relationships of indicator and pathogenic
bacteria in stream water. J. Water Pollut. Control Fed. 45:1736-1745.
Standridge, J. H. 1976. Comparison of surface pore morphology of two brands of membrane
filters. Appl. Environ. Microbiol. 31:316-319.
Stuart, D. G., G. McFeters, arid J. E. Schillinger. 1977. Membrane filter technique for the
quantification of stressed fecal coliforms in the aquatic environment. AppI. Environ.
Microbiol. 34:42-46.
Taguchi, K. 1960. Experimental studies on the examination of coliform organisms in water. I.
Fundamental studies on incubation temperatures. Bull. Inst. Pub. Health. 9:165-175.
Taylor, R. H. and E. E. Geldreich. (979. A new membrane filter procedure for bacterial counts in
potable water and swimming pool samples. J. Am. Water Works Assoc. 71:402-406.
Thomas H. A. Jr. 1955. Statistical analysis of coliform data. Sewage Industrial Wastes 27:212-
222. —
-------
53
Thomas H. A. and R. A. Woodward, 1955. Estimation of coliform density by the membrane filter
and the fermentation tube methods. Am. J. Pub. Health 45:1431-1437.
Thomas, H. A. and R. L. Woodward. 1956. Use of molecular filter membranes for water potability
control. J. Am. Water Works Assoc. 48:1391-1402.
Thompson, ft E. 1927. Irregularities in the test for B. coD in Water. J. Bacteriol. 13:209-221
Tobin, R. S., P. Lomax, and D. J. Krushner. 1980. Comparison of nine brands of membrane filter
and the most-probable-number methods for total coliform enumeration in sewage-
contaminated drinking water. Appi. Envi orn. Microbiol. 40:186-191.
Toombs, R. W. and D. A. Conrior. 1980. Proficiency test sample media for single and mixed pure
cultures of water pollution indicator bacteria. Appi. Environ. Microbiol. 40:883-887.
Tubiosh, H. S. 1951. The anoerogenic effect of nitrates and nitrites on gram-negative enteric
bacteria. Am. J. Pub. Health 41:833-838.
U.S. Environmental Protection Agency. 1975. National Interim Primary Drinking Water Regu-
lations. Fed. Regis. 40:59566-59588.
Venkobachar, D, L. Lyengar, and A. V. S. P. Rao. 1977. Mechanism of disinfection: Effect of
chlorine on cell membrane functions. Water Research 11:727-729.
Vial, J. and C. Geoffray. 1980. Effect of sample holding time on the enumeration of fecal
indicators in samples of water for distribution. Tech. et Sd. Municipales 75:1 10-115.
Warren, L. S., R. E. Benalt, and J. A. Jessee. 1978. Rapid enumeration of fecal coliforms in water
by a colorimetric -gaIoctosidase assay. AppI. Environ. Microbiol. :l36-l4I.
Woodward, R. L. 1957. How probable Is the most probable number. J. Am. Water Works Assoc.
49:1060-1068.
World Health Organization Environmental Health Criteria Programme. 1980. Guidelines for
drinking water quality-microbiological aspects. EHE/EHCI8O.2 IA/rev. 2.
Vee, G. S., W. B. Krabek, and C. P. Schaufus. 1953. A new medium for bacteriological analysis
with molecular filter membranes. J. Am. Water Works Assoc. 45:945-951.
Zaske, S. K., W. S. Dockins, and G. A. McFeters. 1980. Cell envelope damage in E. coil caused by
short-term stress In water. Appl. Environ. Microbiol. 41:386-390.
-------
SOURCE, TREATMENT, AND DISTRIBUTION
Richard S. Engelbrecht
Panel Members: V. Argaman, John Gaston, Ed Geldreich, Charles Haas, John Hoff, John Johnson,
Rick Karlin, Ed Kreusch, Gary Logsdon, James Manwaring, Blame Severin, Otis Sproul, Jerry
Swanson, John Trenary
ABSTRACT
INTRODUCTION 2
SOURCE 3
Selection of Raw Water Source - Quantity and Quality 4
Surface Water 5
Protected Surface Water 5
Quality Deterioration 5
Microbial Quality of Surface Water 7
Groundwater 8
Protected Groundwater 8
Quality Deterioration 8
Protection of Raw Water Source to
Quality of Source Water and Treatment Requirement I I
Summary and Recommendations 13
Research Needs 15
TREATMENT 6
Raw Water Storage 16
Aeration 19
Prech lorinatlon 20
Coagulation-Flocculation and Sedimentation 21
Hardness ReductIon 23
Filtration 24
Activated Carbon 27
Disinfection 29
Chlorine 30
Ozone 34
Chlorine Dioxide 38
iodine 41
Ultraviolet Light 42
By-Product Formation
Operation of Water Treatment Facilities 46
Summary and Recommendations 46
esearch Needs
FINISHED WATER STORAGE AND DISTRIBUTION 50
Disinfectant Residual in Distribution Systems 52
Open and Closed Reservoirs 52
Distribution System Piping and Appurtenances 53
Microbial Regrowth and Transient Microorganisms 54
Back Siphonage and Cross-ConnectIons 55
Operation of Distribution Systems 55
Summary and Recommendations 56
Research Needs 57
LITERATURE CITED 58
ABSTRACT
The microbial quality aspects of potable water supplies as affected by the raw water source,
treatment processes, and the dIstribution network are reviewed.
-------
2
Regarding raw water sources, It is recommended that continuous monitoring of pH and
turbidity be practiced; that treatment requirements be related to source water quality, and that
further studies be undertaken to determine whether a reliable relationship can be established
between raw water quality and required degree of treatment.
In view of the importance of water treatment in meeting the microbial water quality goals, It
Is recommended that all water supplies be disinfected unless it Is proven that microbial
contamination Is not a problem; all surface water supplies should be pretreated and filtered unless
shown that this degree of treatment is not required; and all surface water supplies should be
continuously monitored for disinfectant residual. With respect to potential formation of THM upon
disinfection, It is recommended that the microbial standards take precedence over the MCL
established for THM. It is not suggested, however, that compliance with the established MCL for
THM be abandoned. A number of research topics are indicated in the area of various treatment
processes and their relation to microbial water quality.
The water distribution system provides many opportunities for microbial degradation and
contamination of water supplies. To minimize these opportunities, it is recommended that a
disinfectant residual be maintained In all portions of the distribution system; all new finished water
reservoirs be covered; new and repaired mains be properly disinfected before being placed In
service; an active and effective cross-connection control program be implemented for every public
water supply; and research be undertaken to study microorganisms colonization in distribution
systems and Its prevention, and the relationship between various physical/chemical water char-
acteristics and the occurrence and density of microorganisms.
INTRODUCTION
The choice of raw water supply, the treatment of the wafer, and transport of the water to
the consumer, together with conscientious surveillance, Including Inspection, certification, and
monitoring, represent a continuum of natural and artificial barriers to the transmission of
waterborne diseases. With the onset of increased water requirements resulting in greater need for
direct and Indirect water reuse, the continuum becomes nearly cyclic as the water supply becomes
increasingly stressed by the wastes generated by the consumers. Any breakdown In any of the
barriers to disease transmission, from the mishandling of waste products which may affect the
quality of the raw water to the transport of treated water, ultimately endangers the integrity of
the system.
-------
3
Predicting the survival of pathogenic microorganisms in the entire water system, from source
to consumer, is very complex. Some of the barriers to disease transmission, such as water
filtration, are fairly well understood while other barriers are less so. For example, current
controversy exists over the importance of disinfecting secondary wastewater effluents, the
practice of prechlorination of surface water supplies and the consequent danger of trihalomethane
formation, and the efficacy of maintaining a disinfectant residual in transporting wafer to the
consumer through the distribution system. The results of alterations in these practices on the
transmission of disease are essentially unknown.
The above line of discussion leads ultimately to the question of the minimum treatment and
handling requirements for a given water system. Obviously, some water systems must necessarily
be subject to stricter requirements than others.
Finally, although it is not the immediate topic of this paper, the enforcement of water
quality standards, together with effective monitoring programs for water quality assurance, is an
essential requirement for disease control.
SOURCE
The major sources of raw water available for potable use may be identified as surface waters,
groundwaters, desallnlzed waters and reuse or reclaimed waters. A blended water supply is defined
as any combination of waters from two or more sources. Surface waters include natural lakes and
Impounded or unimpounded rivers or streams. Groundwaters are those waters trapped In
underground aquifers above Impermeable geological formations which become available through
springs or wells. Brackish or saline water treated by evaporation, electrodialysis, osmosis or other
means to remove salts constitutes a desallnlzed source. Water reuse may be either direct such as
the use of treated wastewaters, or may be indirect such as the recharge of treated wastewater Into
groundwater aquifers.
ApproxImately 75 percent of the U.S. communities having public water supplies use ground-
waters as a raw water source, while 7 percent use a blend of ground and surface waters (McCabe,
et al., 1970). Of the existIng water supplIes, 63 percent are surface waters on a volume basis;
22 percent are groundwater; and IS percent are salIne waters (CuIp, 1979). Water reuse In the U.S.
amounts to 250 bIllion gallons a year with roughly 62 percent being used in agrIculture, 31 percent
In Industrial processes, 5 percent goIng to groundwater recharge, and 2 percent for recreational
purposes.
-------
The demand for water is projected to increase, with groundwater supplies being developed to
c i higher degree than surface waters (U.S. Government Printing Office, 1980). While most areas of
the country will be able to meet demands in the year 2000, severe shortages are projected for
certain areas of the U.S., particularly in the western states (Cuip, et cii., 1979). Shortages are
expected to result from inadequate distribution systems, groundwater overdrafts, quality degrci-
dation of groundwater and surface water, institutional constraints and competition between users.
Selection of Raw Water Source - Quantity and Quality
The choice of a raw water source is usually based upon the utilization of that source which
provides the best quality water to meet the demand for the required quantity. In the case of a
public water supply, the raw water source must be adequate to meet all municipal, institutional,
and industrial requirements plus the fire fighting demand of a community. In the case of private
supplies, the source must only meet the immediate demand of the consumers. The quality of a
private source, used solely by industry, may be either higher or lower than that required by the
general needs of a community. In many cases, a single source of surface water or groundwater will
adequately meet the demands. However, in some cases more than one source must be developed.
The major raw water sources available for use as a potable water supply are surface waters
and groundwaters. The choice of either a surface water or a groundwater dictates the level of
treatment required in order to provide the quality of water required. In general, a groundwater
will have a more consistent and better microbial quality than a surface water. In many cases,
especially with small systems, only disinfection prior to distribution of the water Is required for a
groundwater source. With groundwater sources of poor quality, or those where the general
demands of the community or Industry require further treatment of the water, the typical
treatment systems might include aeration and filtration for control of Iron, manganese and gases,
or hardness removal and filtration. The quality and quantity of surface waters are usually less
consistent than groundwaters. Further, surface waters usually possess a higher level of organic
matter, microbial contamination, and particulates. Typical treatment systems for surface waters
consist of storage primarily to provide a constant volume of water, coagulation-flocculation and
settling, filtration and disinfection. With some surface water sources, prechlorination may be used
for the oxidation of organic matter or control of high levels of microbial contamination,.
-------
5
Stwface Water
A watershed Is defined as the entire region or area which contributes drainage or runoff
waters to a river or lake. A protected watershed is a drainage area in which human activities, such
as industrial or municipal waste discharges or agriculture, are controlled so that the quality of the
surface water is not deteriorated.
Protected Surface Water. A great variety of microorganisms, including bacteria, algae,
protozoa, fungi, and viruses, are indigenous to a watershed and the associated surface waters.
Some of these organisms are natural Inhabitants of water, whereas others are transients, gaining
entrance to the water system from the air or by way of runoff waters. It is generally accepted
that even a protected watershed b subject to an influx of potential human pathogens, such as the
contribution of Giardia lamblia from Infected animal reservoirs. The survival of organisms in
water is dependent upon various ecological factors, Including the season of the year which affects
the temperature and the available sunlight, pH, microbial competition and antagonism, the
presence of microbial nutrients, and other factors.
Quality Deterioration. The major sources of microbial pathogens entering a surface water
supply are domestic and Industrial wastewater discharges, agricultural runoff, septic tank seepage
and leachate from solid wastes disposal sites. The type and quantity of pathogens present in these
discharges is dependent upon the type and quality of the discharge, the degree of treatment
imposed prior to disposal, and other factors. For example, the presence of pathogens in
wastewater discharges is dependent on the time of year and the economic status and general health
of the contributing population. It must be recognized that the quantity and type of pathogens
reported to be present may not Include all of those actually present, primarily because of
inadequate sampling andfor recovery methods; this Is particularly true in the case of the viruses,
protozoa and helminths. A secondary impact of waste discharges, runoff, and leachates Is on the
quality of the water reaching the consumer, In that any change in source water quality may affect
th. efficiency of the water treatment processes,
A partial listing of pathogens potentially present in raw municipal wostewater Is presented in
Table I (Environmental Protection Agency, 1980). Among the waterborne bacteria, Shigella and
Salmonella continue to have th. greatest health impact. The viruses of greatest Importance oppear
-------
6
Table I. Infectious Agents Potentiall Present in
Raw Domestic Wastewater
Organism Disease
Bacteria
Shigella (1 spp.) Shigellosis
(Bad I lary Dysentery)
Salmonella typhi Typhoid Fever
Salmonel Ia (1,700 spp.) Salmonel losis
Vibrio cholerae Cholera
Escherichia coil Gastroenteritis
Yersinia enterocolitica Yersinlosis
Leptospira (spp.) Leptospirosis
Campylobacter Gostroenteritis
II. Viruses
Enteroviruses (71 types) Gastroenteritis, heart anomalies,
meningitis
Hepatitis A virus Infectious hepatitis
Adenovirus (31 types) Respiratory disease
Rot avi rus Gastroenteritis
Reovirus Not clearly established
Norwalk-like Gastroenteritis
III. Protozoa
Entamoeba histolytica Amebiasis (Amoebic Dysentery)
Giardia lamblia Giordiasis
Balantidium coil Balantidiasis (Balantidial Dysentery)
IV. Helminths
Ascaris lumbricoldes Ascarlasis
Ancylostoma duodenate Ancylostomiosis
Necator americanus Necatorlosis
Ancylostoma (spp.) Hookworm
Strongyloldes stercoralis Strongyloidlasis
Trichuris trichiura Trichuriasis
Taenla (spp.) Taen lasis
Enterobius vermicularis Enterobiasis
Echinococcus granuloss Hydatidosis
°From EPA (1980).
-------
7
to be those associated with infectious hepatitis and gastroenteritis. In terms of the number of
reported cases, the most prevalent waterborne disease is giardiasis, indicating that Giardia lamblia
is perhaps the most important waterborne pathogen. Entamoeba histolytica is also recognized as
an important pathogenic protozoan. Although it is impossible to give the density of the different
pathogenic agents that might be found in raw municipal wastewater, the variation and/or order of
magnitude of the density of certain pathogens may be illustrated: Salmonella up to l0 /l;
protozoan cysts, up to IO /l; helminth (ova), up to I0 /l; and enteric virus (pfu), up to
(Environmental Protection Agency, 1980). The quantity of organisms discharged to source waters
is dependent upon the degree of treatment the wastewater receives. Certain processes employed
in treating wastewaters remove a significant number of microorganisms, including the enteric
pathogens.
Industrial wastewaters also pose a potential threat to the microbial quality of surface waters.
For example, pathogenic microorganisms may be found in certain industrial wastewaters, such as
slaughterhouse, dairy and tanning wastes and, thus, may be discharged directly to surface waters.
Industrial wastewaters may also enhance the survival of pathogenic microorganisms discharged to a
surface water from another source.
The intermittent runoff of waters from urban and rural areas and landfill seepage constitutes
a potential source of pathogenic bacteria (Gauf in, 1974; Geldreich, 1972; Olivieri, et at., 1977).
Urban runoff, especially that associated with combined sewer overflows, may contain untreated
waste of human and animal origin. Rural runoff is likely to contain waste material from domestic
animals and septic tank seepage. The use of phosphate and nitrogen fertilizers in agricultural areas
may result in the discharge of these nutrients to water sources. These inorganic compounds may
indirectly affect the microbiology of a source water. Surface seepage from mproperly constructed
and/or operated, landfills and the runoff from lands receiving sludges produced in wastewater
treatment may contribute human pathogens to surface waters.
Microbial QuaLity of Surface Water. Human pathogens may be present In discharges of human
waste products and these organisms can enter surface waters and become the cause of waterborne
diseases. Between 1946 and 1970, 54,935 cases of waterborne disease were attributed to
contaminated surface waters (Craun, et al., 1973).
-------
8
Surface waters near metropolitan areas are generally contaminated with human wastes. As
an example, in 1981 there were 56 water supply intakes along the 977.8 miles of the Ohio River and
of these, 43 were within 5 miles of an upstream wastewater treatment plant effluent (Geldreich,
1981). In a survey of 20 cities using surface water supplies, the estimated wastewater component
of the source water during normal flow ranged from 2.3 to 16 percent and in certain cases
increased to predominantly wastewater during low flow periods (Swayne, 1980)..
The presence of coliform organisms in wafer is accepted as an indication of the presence of
fecal pollution. Geldreich (1966, 1972) has reported a high degree of correlation between the
presence of fecal colf forms and the occurrence of pathogens in surface waters and water distri-
bution systems. Fecal coliform counts at raw water intakes along the Missouri River, serving
Omaha, St. Joseph and Kansas City, have frequently exceeded 2,000/100 ml (Environmental
Protection Agency, 1971).
Groundwater
• The potential presence of pathogens in groundwater Is based upon the concept of a pristine
aquifer and subsequent deterioration of the water quality, such as by land disposal of wastes,
groundwater recharge, etc.; also, overdraftlng of an aquifer often accelerates the deterioration
process.
Protected Groundwater. The microbial qualify of groundwaters Is usually acceptable since
microorganisms are removed by filtration and adsorption as the water moves through the soil.
Percolation of water through a few feet (6-7) of fine gralned soil Is generally adequate to remove
enteric pathogens (Hirshleifer, et al., 1960). For example, removal of bacteria and viruses in
excess of 99 percent have been observed in percolating a raw surface water through a dune
Infiltration system (Kool, 1978). In general, the infiltration process Is similar Ln efficiency and
mechanism to slow sand filtration.
Quality Deterioration. The rate and extent of contamination of groundwoters is site-specific,
depending on the local topography, type of welt structure, rate of water withdrawal, type of soil,
and type of pollutant. The movement of groundwater Is considered to be very slow, on the order of
several tens of feet per year In granular, porous media aquifers. Under such conditions, a plume of
contamination may travel a great distance in a groundwater but the time of travel is likely to be
-------
9
very great. This would not be the case with solid rock aquifers with cracks or fissures or with a
limestone aquifer having solution channels. Consolidated rock aquifers with joints, fissures, or
solution passages may permit rapid travel of the groundwater with the result that microbial
contaminants may not be filtered out. Studies performed In suburban Minneapolis-St. Paul in 1959-
1961 showed that where limestone aquifers were near the ground surface, all the wells were
generally affected by chemical contamination from wastewafer (nitrates and surfactants) and
bacterial contamination was common (Woodward, 1961). At the time of the study, Minneapolis-St.
Paul suburban homes generally were served by individual water supplies or wastewater disposal
systems or both. A shallow limestone aquifer at Posen, Michigan, was found to be a contributing
factor In an Infectious hepatitis outbreak In 1959 (Vogt, 1961).
The persistence of enteric pathogens in soil Is usually less than several months. Therefore, In
order for enteric pathogens to be a problem In groundwater, the site of the contamInation must be
relatively close to the point of withdrawal. The practice of overdrafting a supply, however, can
alter the natural flow of water plumes, causing a well supply to become contaminated. The
survival time of microbial pathogens In groundwaters may be greater than in surface waters. In
groundwaters, the lock of exposure of the water to ultraviolet light, lessened microbial antagonism
and predation, and constant low temperatures favors Increased survival time. Gerba, et ., (1975)
reviewed the survival time of various bacteria in groundwaters under different conditions and
summarized their findings as follows (unless otherwise stated, recovery of the test organism was
minimal): Escherlchla coil in groundwater stored in the laboratory, 4-4.5 months; 9 jj in
groundwater in the fIeld, 3-3.5 months; E. coIl in a recharge well, 63 days; total coliforms in well
water, 17 hours with 50 percent reduction; Salmonella In Infiltration waters In sand columns,
44 days; Shigella In waters In sand columns, 24 days; Shigella flexnerI In well water, 26.8 hours with
50 percent survival; and VlbrIo cholerue , 7.2 hours with 50 percent survival; no information on
water temperature was given. Wellings, etal., (1975) demonstrated that enteroviruses In shallow
groundwater below a wastewater irrigation site could survive for at least 28 days. Veager and
O’Brien (1977) reported that I I to 14 days were required for 90 percent removal of enterovlruses In
groundwater. It Is recognized that these agents are, usually removed close to their entry paint into
the soil and would not normally be observed at the corresponding distances in soil represented by
these times of survival. Organisms would be observed, however, in groundwater flowing for these
times in fractured or fissured rocks or in solution cavIties.
-------
10
There are many reports of the Isolation of microbial pathogens and indicator bacteria from
groundwaters. Allen and Geldreich (1975) cite literature indicating the isolation of Shigella
flexneri , S. sonnei, Salmonella typhi and polioviruses from different groundwaters. According to
Allen and Geldreich (1975) and McCabe, et ci., (1970), coliform organisms are also routinely
detected in groundwaters. On the other hand, the absence of coliform organisms in a groundwater
does not assure the absence of microbial pathogens (Boring, et oh, 1971; McFeters, et al, 1974).
From a review of the literature on viruses in groundwater, Keswick and Gerba (1980) found that the
viruses reported as being isolated and identified from drinking water wells and beneath land
treatment sites included poliovirus, echovirus, coxsackievirus, and rotavirus, as well as bacterial
viruses.
In reviewing the outbreaks of waterborne disease between 1971 and 1977, Craun (1949)
determined that 49 percent of the 192 reported outbreaks were attributable to untreated or
inadequately treated groundwater. Cretin and McCabe (1973) concluded that 49.4 percent of the
waterborne disease outbreaks were due to unsatisfactory well construction and improper location
of wells, resulting in contamination of the source. Thus, It would appear that a large fraction of
the incidence of woterborne disease is caused by contaminated groundwaters which were either
untreated or given minimal treatment.
Protection of Raw Water Source
The U.S. Public Health Service Drinking Water Standards (1946) stated that “The water supply
shall be: (a) obtained from a source free from pollution; or (b) obtained from ci source adequately
purified by natural agencies; or (c) adequately protected by artificial treatment.” For the purpose
of discussion in this paper, the protection of a source water means any actions which are taken to
assure that a potential pollutant is not permitted to enter the water source to an extent that It
cannot be contrOlled by natural purification mechanisms. The natural purification mechanisms
Involved In groundwater sources and surface water sources are generally taken to be different.
The U.S. Public Health Service (1969) IdentIfied the following measures to be taken In
protecting a groundwater source. All groundwater withdrawal points should be located a “safe”
distance from sources of pollution including septic tanks and other wastewater disposal systems,
sewers, industrial discharges, land drainage, farm animals, fertilizers, and pesticides. A “safe”
distance was defIned as the distance that ensures no contaminant will be drawn or will flow Into
-------
II
the withdrawal point when the conditions for pollution, e.g. withdrawal rates and the water table
level, are most adverse. The direction of flow should be determined and wells should be placed at
an elevation higher than any points of pollution. All withdrawal wells should be properly sealed
against surface water contamination, and meet other construction standards, such as those set
forth by the AWWA (1966).
It is becoming Increasingly difficult to find a surface water source that is not polluted. To
protect a surface water source, the Public Health Service (1969) recommended complete ownership
of the Impoundment area and if possible, complete ownership of the watershed by the utility. The
latter, however, is Impractical on large river systems; here the practice is to provide the longest
lag time or distance of travel between water supply intakes and known upstream point sources of
pollution. The use of reservoirs or impoundment areas apart from the river system can help in
protecting the source by Increasing the time available for natural purification. In addition, when
the qualify of the water in the river is poor, the Impounded water can be used until such a time
that the quality of the river water Improves.
QuoUty of Source Water and Treatment Requfrement
In order to assure that a water Is free from detectable pathogens, It Is often required that the
water be treated by a series of processes. However, no single process, nor In fact any series of
processes, can be expected to produce a finished water consistently free from pathogens if the raw
water source is of sufficiently poor microbial quality. Because of this, several national and
International committees have fried to define raw water microbiological standards. These
standards, for various reasons, are based on levels of indicator organisms such as coliforms, plate
count, etc., rather than on the enumeration of pathogens. There is sufficient, evidence available to
question whether the absence of Indicator organisms Implies the absence of pathogens, especially in
groundwater sources. Perhaps the ultimate goal of a treatment standard should be to define the
minimum treatment requirements of a wafer based on the quality of the raw water. However, the
establishment of such a relationship would be a large task, based on the fact that the removal of
microorganisms In the various water treatment processes Is Intimately fled to the organic,
Inorganic and suspended solids qualities of the raw water, It seems ixilikely, therefore, that
standards based on the presence of Indicator organisms alone will adeqvately define the minimum
treatment required. Even though the adequacy of a microbial raw water quality standard based
-------
sole’y on indicator organism densities may be questioned, It is of Interest to examine the various
proposed standards and, where possible, the rationale for their development.
The U.S. Public Health Service (1969) identified three types of source waters: those usable
without treatment, those requiring disinfection only, and those needing complete treatment.
Conventional complete treatment was defined as including coagulation-f locculof Ion and sedimen-
tation, rapid granular filtration, and disinfection. The bacteriological standard for waters requiring
no treatment was set at one total coliform per 100 ml on the basis of on arithmetic monthly
average, as defined in the Public Health Service Drinking Water Standard (1962). For waters
requiring disinfection only, the raw water upper limit was set at 100 total coliformIIOO ml or
20 fecal coliforms/lOO ml on the basis of an arithmetic monthly average. The Public Health
Service (1969) recommended that raw water containing more than L ,00O fecol coliforms/l00 ml, on
a monthly geometric mean basis, should not be employed for drinking water, even with complete
conventional treatment.
The National Academy of Sciences and National Academy of Engineering (1973) recom-
mended that the geometric mean density of coliforms in raw surface waters used for a public water
supply should not exceed 2,000 fecal collformsl tOO ml and 20,000 total coliforrns/ 100 ml. No
sampling frequency was established. The rationale for this recommendation was based on the
observations of Geldreich (1970) and Geldreich and Bordner (197 I). In stream samples, It was found
that the frequency of Salmonella detection increased sharply when the fecal coliform densities
were greater than 2,000/100 ml. For 123 samples with fecal coilform density above 2,000/100 ml,
positive detection of Salmonella occurred in 97.6 percent of the samples. Further, when fecal
coliforms exceeded 2,000/100 ml, Salmonella , echovirus 7 and 33, and poliovirus types 2 and 3 were
recovered at several water intakes along the Missouri River (Environmental Protection Agency,
1971).
The mast recent set of standards proposed for a raw surface water source appears to be that
of the Commission of the European Communities (1975; 1978). Three classes of raw water were
defined: those requiring filtration and disinfection; those requiring coagulatlon-flocculatlon,
filtration, and disinfection; and those requiring extensive treatment, including coogulatlon-
flocculation, filtration, supplemental treatment, such as activated carbon, and disinfection.
Maximum microorganism levels for each of the three classes of row water were, respectively, 50,
5,000, and 50,000 coliforms/lOO ml; 20, 200, and 20,000 thermotolerarit collforms/lOO ml; or 20,
1,000 and 10,000 fecal streptococci/lOO ml.
-------
13
While the establishment of minimum treatment requirements based upon the microbial
quality of the source wafer is an interesting and potentially meaningful approach to regulating
finished water quality, information currently available does not appear to be adequate to set such
standards.
Summary id Recommendaflcns
The major raw water sources for a potable water supply are currently either surface waters
or groundwaters and, on occasion, a blend of the two. In selecting a raw water source, it is normal
practice to use that source(s) which provides the best quality water to meet the demand for the
required quantity. Of the two sources, a groundwater generally will have a more consistent and
better microbial quality than a surface water.
A wide range of microorganisms, in terms of both type (bacteria, algae, protozoa, fungi, etc.)
and density, may be found in surface waters. Some are indigenous to the water environment, while
others may be transients, gaining entrance to the water from the air or by way of wastewater
discharges or runoff from urban or rural areas. The major sources of microbial pathogens affecting
a surface water ore domestic and Industrial wostewater discharges, both treated and untreatedi
agricultural and urban runoff; septic tank seepage; and leachote from solid waste disposal sites, Of
these, It s likely that domestic wastewater will continue to be the primary cause of the
degradation of the microbial quality of surface waters, including the presence of human pothogens.
However, even with a protected watershed, microbial pathogens may be found in a surface water
because of human activity in the area or as a result of contribution of pothogens from animals,
The type and density of pathogens in the various sources, as well as In surface waters, will vary
with the nature of the source and the potential for survival which is a function of a variety of
ecological factors.
Depending upon soil type, depth and associated factors, the microbial quality of natural
waters found In many aquifers is acceptable because of the removal of microorganisms by filtration
and adsorption as the water percolates through the soil. On the other hand, groundwaters are
subject to microbial contamination on a site-specific basis, as influenced by the local topography,
type of soil, nature of the contaminant, rate of water withdrawal, type of well structure, etc. The
movement of groundwater may be very slow in granular, porous media aquifers to very rapid In
aquifers consisting primarily of rock with cracks and fissures or with limestone aquifers having
-------
‘4
solution channels. Microorganisms are not removed in these cracks and fissures since they do not
contain fine materials capable of acting as a filtering medium. Although the predominant form of
contamination of groundwaters is chemical, microbial contamination is possible. For example,
leachate from solid waste disposal sites or drainage from domestic wastewater or sIudge spread on
land may lead to the degradation of the microbial quality of a groundwater. Further, the survival
of human pathogens in soil and aquifers may be on the order of several months which may be
greater than in surface waters.
It is becoming increasingly difficult to find a surface water source that is free of man-made
pollution which often affects the microbial quality. Likewise, the potential exists for more
extensive pollution of aquifers so that the quality of groundwaters may become more impaired in
the future. It is mandatory, therefore, that every effort be mode to improve and/or protect the
quality of the source waters used for public water supplies. The production of a microbiologically
safe water supply must begin by controlling continuous and intermittent sources of contamination
affecting both surface and groundwaters. If this is not successfully accomplished, a greater burden
will continue to be placed on the other barriers associated with producing a microbiologically
acceptable water supply, i.e. the natural purification process and water treatment facilities.
A water treatment facility, which normally includes a series of processes for improving the
quality of the withdrawn raw water, constitutes the major barrier to the transmission of
waterborne disease. In general, the quality of the raw water dictates the processes included in the
treatment system. Thus, a raw water having a poor microbiological quality must receive more
extensive treatment if the objective of producing a finished water free of detectable pathogens is
to be achieved. For this reason, attempts have been made to equate the microbiological quality of
the raw water to the degree of treatment required to meet an established standard for potable
water. Although the development of minimum treatment requirements or standards based upon the
microbiological quality of the raw water is interesting and potentially useful, it is not currently
possible to do so because of the lack of adequate information.
Having reviewed and considered the microbial aspects associated with raw water sources with
respect to producing a potable water supply, it is recommended that:
I. Regulatory agencies consider the raw water quality relative to real and potential
sources of contamination, turbidity, bacterial quality, etc. when specifying the required
-------
‘5
treatment of a particular source water. The issue of establishing a water quality index,
relating raw water quality and treatment requirement, has been raised. The proposed
indices attempt to relate treatment requirement only to the bacteriological quality of
the row water, It is believed that other parameters besides the bacterial quality should
be considered and that there are insufficient data available regarding these other
parameters to permit the establishment of an index relating raw water quality and
treatment.
2. Continuous monitoring of raw surface water sources for turbidity and pH be practiced.
Mandatory raw water bacteriological monitoring would not provide real time informa-
tion for control purposes and should, therefore, not be required. A historical record of
the bacterial quality of a raw water source, however, would be valuable but not
essential,
R earth N
Increasing land develoment and use within watershed areas may be expected to
adversely alter the microbial characteristics of surface waters and shallow ground-
waters. The elimination of or even a reduction in the practice of disinfecting domestic
wastewaters may also Impact the microbial quality of raw water sources. Thus, the
margin of safety associated with existing water treatment plants for producing a
microbiologically acceptable finished water may be compromised; the extent of this risk
is unknown. Research is needed to evaluate the effect of further quality deterioration
of source waters on the efficiency of removing/inactivating microorganisms by the
various unit processes employed in water treatment and by treatment facilities as a
whole.
2. AuthoritIes hove attempted to classify water sources based upon quality and then to
specify the required degree of treatment of each classification. While the available
information does not support this approach, It is possible that further research may
provlde the data base upon which to develop a reliable relationship between raw water
quality and the specification of water treatment technology. If this type of relationship
can be developed, It would serve as a guideline. If Is not meant to Infringe on the
professional judgment of the design engineer. Research Is needed to determine whether
-------
16
a reliable relationship between raw water quality and the required degree of water
treatment be established and if so, what is the relationship.
TREATMENT
In the following sections, the various unit processes and operations employed in the treatment
of water supplies are reviewed with respect to their ability to remove and/or inactivate
microorganisms. The effect of each treatment unit including raw water storage, aeration,
prechlorination, coagulation-flocculation, hardness reduction, filtration, activated carbon treat-
ment, and disinfection on the microbiological quality of water is considered, where possible, with
respect to the four broad groups of pathogenic organisms, i.e. bacteria, viruses, protozoa, and
helminths.
Table 2 is a generalized summary of the expected efficiency of various physical-chemical
treatment processes to remove microorganisms from water, based upon available information.
Although the Indicated removal efficiencies may be somewhat conservative, the Information should
still be accepted with some degree of caution. It should be recognized that most of the data in the
literature on removal of microorganisms by the various treatment processes have been derived
from carefully controlled laboratory experiments. Because of variable operating conditions, field
data would suggest less removal than observations obtained from laboratory studies. Also, the
performance of certain treatment units may be influenced by processes placed either before or
after them. For example, the efficiency of rapid filtration in removing microorganisms is
dependent upon the water being pretreated by coagulation-flocculation. It should be noted also
that a water treatment facility may not necessarily Include all the treatment units listed In
Table 2.
Raw Water Storage
Raw water storage reservoirs are useful for insuring an adequate quantity of water during
drought conditions or In minimizing the fluctuation of the quality of a raw water supply, e.g. the
occurrence of upstream spill or accidentlal discharges. Holding a raw water supply for long
periods of time also can affect the microbial quality of the water.
The major factors affecting the removal of microorganisms in a reservoir are sedimentation,
natural predation, sunlight, and Increased temperatures. The removal of particulate matter or
-------
17
Table 2. Expected Efficiency of Physical and Chemical Water Treatment
Processes to Remove Pathogenic Microorganisms.
Percent Removal
Unit Process Bacteria Viruses Protozoa• Helminths
Storage Reservors 80- 90 90- 90 NK NK
Aeration NK NK NK NK
Coagulation- 90- 99 90- 99 >90 >90
Flocculation plus
sedimentation
Hardness Reduction
high lime 90-99.9 99-99.9 NK NK
low lime 90- 99 90- 99 NK NK
Rapid Filtration
with coagulation- 90-99.9 90- 99 90-99.9 NK
flocculation plus
sedimentation
without coagulatIon- 0. 90 0- 50 0- 90 NK
flocculation plus
sedimentation
with coagulation- 90- 99 90- 99 90- 99 NK
flocculation; no
sedimentation
Slow Filtration 0 90-99 9 90-99.9 NK NK
Diatornoceous Earth 90- 99 95 99 NK
Filtration
with pretreatment and
precoating of filter
Activated Carbon NK 10- 99 NK NK
NK - Not known
-------
18
turbidity has two effects on microorganisms in water. As particulate matter settles out of the
water, it carries with it many absorbed microorganisms. The second effect is to increase the
penetration of ultraviolet light (UV) into the water column. The inactivation of microorganisms by
ultraviolet light may be counterbalanced, however, by the presence of algal blooms, which would
interfere with the penetration of UV light into the water. Natural predation may be accomplished
by protozoa, rotifers, and phage. Increases in temperature in the supply water tend to Increase the
activity of predators as well as increase the natural die-off rates of microorganisms.
The reduction of total coliforms by raw water storage appears to be quite variable. Coliform
reductions, ranging from 13 to 97 percent, have been observed at water treatment utilities taking
water from the same river system (Ohio River Valley Water Sanitation Commission, 980). Similar
investigations on E. coIl indicate ranges of 77 to 99.8 percent removal in reservoirs, depending on
the season, with the least removal in fall or winter and improved removals in the spring or summer
(Metropolitan Water Board, 197 1-1973; Thames Water Statistics, 1976). In another study,
95.8 percent of the natural population of E. coIl was removed In a reservoir In one week (Poynter
and Stevens, 1975). Total coliforms were reduced to a somewhat lesser degree than E. coil, i.e.,
85 percent In one week and 90 percent after two weeks. The die-off of fecal streptococci In
bottled river water was reported In one study to be similar to that of 2 jj (Burman, et al, 1978).
Where Increases in bacterial numbers in a reservoir have been observed, they appeared to coincide
with peak algal activity (Kool, 1977; Price and Valadon, 1970). In farm pond water stored in bottles
at 21-29°C, it was found that Salmonella survived for periods of 14-16 days (Andre, et al., 1967).
Where farm yard wastewater was used to supplement irrigation water, holding tank storage for a
20-day retention period in summer or 60-day period In winter was observed to produce a
“ Salmonella free” supply (Braza, 1964). Other data indicate a reduction of 90 percent is possible
with S. lyphi after about two weeks of storage and that a high or low p1-I Increased the death rate
above that observed at neutrality (Cohen, 1922; Fennell, et al., 974). VIbrlo cholerae Inoculated In
stored river waters for one week (7-18°C) was removed to the extent of 99.9 percent (Houston,
1908-1911). When V. choleroe was Introduced Into either a water containing common saprophytes
or added to the high density bacterial flora of wastewater or activated sludge, there was a
significant suppression of the pathogen, often resulting In 99 percent inactivation in 6 hours (Pillal,
et al., 1952).
-------
19
Storage also affects the survival of enteric viruses. Significant removals of viruses from
river water have been reported, with survival being similar to that for the enteric bacteria at
water temperatures above 10°C (Slade, 1977). On the other hand, with poilovirus type 3 it was
found that at a temperature of 50 to 6°C, more than nine weeks were required to achieve
99.9 percent removal.
Although data appear to be lacking, it may be speculated from the size and density of
protozoan cysts and helminth ova that they should be removed from water by storage, primarily
through sedimentation of suspended matter. The specific gravity of both helminth ova and
protozoan cysts Is generally greater than 1.06.
In summary, bacterial and viral removals of 80-90 percent may be attained of ter one or two
weeks of storage under suitable conditions. In addition to the effect of the quality of the water
Itself, factors influencing removal or die-off of microorganisms include sedimentation, natural
predation, inactivation by UV light, etc. Considering the limited amount of definitive information
available, the effect of raw water storage on microbial quality needs further study, particularly as
related to pathogenic organisms and their correlation with indicator systems.
Aeration
While the ability of an aeration process to directly affect the survival of pathogens Is
uncertain, the microbiology of a water supply may be indirectly affected by aeration. Ground-
waters are deliberately aerated for a variety of reasons, e.g. the control of Iron and manganese
which, if present, often makes the water aesthetically unacceptable. The removal of Iron might
also control the growth of iron bacteria in distribution systems. ft may be suggested that in waters
rich In reduced iron, the aeration of the water would oxidize the iron and, thus, improve the
efficIency of subsequent filtration by providing iron fioc Impregnation of the filter bed. Aeration
Is also used to reduce the concentration of C0 2 , H 2 S, and NH 3 as well as CHk and other volatile
organic compounds that may be present in groundwaters. It Is generally recognized that the
development of microbial growths In distribution systems can cause a deterioration In the chemical
quality of the water. By removing the gases, C l -I 4 , C0 2 , NH 3 , H 2 S, and the reduced iron, nutrients
otherwise available for the growth of microorganisms in a distribution system may became limited.
The removal of NH 3 , H 2 S, and certain organic compounds would also enhance the efficiency of
subsequent disinfection. On the other hand, It should be recognized that aeration or the addition of
oxygen to a water increases Its corrosive tendency.
-------
20
Prechiorinotlan
Prechiorination in the treatment of water supplies is defined as the application of chlorine to
water prior to another treatment process used before terminal disinfection. The objectives of
prechlorination vary according to the individual situation and may include improved filter operation
and performance by accomplishing a reduction in the bacterial and algal load; control of the slime
growths; improved coagulation; reduction of taste, odor, and color producing materials by oxidation
and retardation of microbial decomposition; and the provision of an additional safety factor in
disinfecting heavily contaminated waters. It would appear that improved filter operation and
coagulation rather than infectious disease control were the original reasons for the use of
prechlorination (American Water Works AssociatIon, 1925; Streeter and Wright, 1930).
Prechiorination was first used in the early 1900’s and soon became an established water
treatment practice. Enslow (1928) reported on the use of prechiorination at a number of treatment
plants. His observations indicate that the density of coliform organisms in finished water was
consistently lower when prechiorination was used. Streefer and Wright (1930) compared the
recovery of coliform and plate count bacteria after various stages of treatment with and without
prechlorination. The chlorine was added shortly after the addition of coagulant. At the point of
application of the water to sand filters, coliform levels had been reduced substantially both with
and without prechiorination, but levels In the prechiorinated water were consistently lower, often
by an order of magnitude or more. This difference also was observed In waters following filtration
and final chlorination, with coliform and plate count bacteria almost always tower in the finished
water that had been prechlorinated. Analysis of the unit processes on a percentage basis showed
that the efficiency of filtration and postchlorinatlon was much lower In prechiorinated water than
in nonprechlorlnated water. They stated that differences in filtration efficiency may have been
due to disturbon e of the biological condition of the filter by the presence of chlorine and that the
apparent difference In postchtorinatlon results may have been due to the presence of low numbers
of bacteria having a higher degree of resistance to chlorine. It was concluded that the main
advantage of prechlorlnotlon, with regard to bacterial reduction, was one of Improving the overall
effectiveness of the rapid sand filtration process.
Walton (1961) studIed the effectiveness of water treatment processes for removing or
inactivating coliform bacteria In more than 80 operatIng plants. The effects of prechlorination
-------
2$
were studied In three plants. in one plant, he compared coliform densities in filter effluents with
and without the prior addition of chlorine or chioramine over a range of water coliform densities.
Coilform densities in the filter effluent were consistently I to 2 orders of magnitude lower in
water that had been prechiorinated. in two other plants, brief periods of discontinuing
prechiorinatlon resulted in detection of coliforms in filter effluent, whereas when prech$orlnation
was employed, coliform levels in filter effluent were consistently <2.2/100 ml. Both of these
plants used primary settling and coagulation in addition to rapid sand filtration.
Although prechlorination continues to be used, recent concerns about the formation of
trihalomethanes (THM) in water supplies hove caused its use to be reconsidered. As a result, the
effects of changing the point of chlorination to late stages in the treatment system have been
studied with respect to THM formation and the resulting microbiological quality of the water
(Symons, 2 j., $981). These studies again demonstrated that the elimination of prechlorlnation
resulted in a higher density of coliform and plate count organisms in the wafer in subsequent
treatment stages. However, neither a measurable change In the microbiological quality of the
finished water, nor any apparent in-plant problems developed as a result of eliminating pre-
chlorination, such as the occurrence of algae and slime growths in filters and sedimentation basins.
Caagulatlan-Flacculatkn and SeWmsntatlon
Chemical coagulation-flocculatlon is a process in which collolds and other suspended
materials ore chemically destabilized and contracted for floc growth. This process is normally
used for pretreating waters for subsequent gravity sedimentation and filtration. Th. process
requires the addition of coagulants to the waters the major coagulants include aluminum sulfate,
ferric iron salts, and polyelectrolytes or polymers. The additIon of lime for hardness reduction may
also accomplish some coagulation-f locculation.
The work of Foligust and Doncoeur (1975), Manwarlng, et al. (1971), Schaub and Saglk (1975)
and Thorup, st c i. ($970) an virus removal by chemical coagulation-flocculation has been reviewed
in a report prepared by the Safe Drinking Water Committee of the National Academy of Sciences
(1977). It was concluded that both aluminum sulfat. and ferric chloride wer. effective primary
flocculants for virus removal. Polycationlc polymers were found to be better In removing viruses
than nonionic or anionic polymers when u.d as coagulant aids. Clays, when added to water as a
coagulant aid, were also observed to absorb viruses and thus, increase virus removal. However,
-------
22
organic matter was found to interfere with virus removal. it was concluded that under carefully
controlled conditions, coagulation-flocculation can remove 90 to 99 percent of the viruses present
in water. A review of the removal of enteric viruses by chemical coagulation-flocculation by
Sproul (1980) indicated greater than 86 percent removal of virus under a variety of test conditions.
However, virus particles removed by coagulation-flocculation may remain viable and may be
desorbed from the floc.
Little definitive information seems to be available on the effectiveness of the coagulation-
flocculation process to remove bacteria. In full-scale plants, Streeter (1927, 1929) observed
bacterial removals of 46 to 83 percent while Cummins and Nash (1978) found the total coliform
removal to be 42 percent. Chang, et at., (1958, 1958) reported 99 percent bacterial removals.
These data would indicate that the removal of bacteria by coagulation-flocculation fluctuates
widely. In a more recent study, Engelbrechf et a!., (1979) reported on the removal of these
organisms under optimal conditions for turbidity (Koolinite clay or natural solids) removal. With
ferric chloride or aluminum sulfate, removal of total coliforms amounted to 60 to 99.5 percent,
with most results being 93 to 98 percent; acid-fast organisms and yeasts were removed to the
extent of 89 to 99.1 percent and 78 to 99 percent, respectively. In experiments with lime, the
yeasts were most resistant with removals of 97 to 99.8 percent, while acid-fast and coliform
removals exceeded 99.7 percent. The high pH associated with the lime experiments (pH 10.7-10.9)
may have accounted for the high removals.
The removal of protozoan cysts, Glardla muris and Entamoeba histolytica , by coagulation and
sedimentation has been studied (Arozarena, 1977; Cram, 1943; Parkhurst, 1977; Spector, et a!.,
1934). Under optimum conditions, It was found that removal of protozoan cysts exceeded 90
percent. It can be expected that because of their size and high specific gravity helminth ova may
also be removed from water by sedimentation.
It should be realized that the operation of the coagulation-flocculation process is optimized
for the removal of turbidity rather than microorganisms. On the other hand, the alteration of
certain control parameters might optimize the process of organism removal. Parameters such as
pH, coagulant dosage and floc size, and the use of coagulant aids might be varied and, as a result,
Improve microorganism removal.
-------
23
Existing data indicate effective removal of enteric viruses and protozoan cysts by the unit
process of coagulation-flocculation and sedimentation. It is speculated that helminth eggs would
also be effectively removed. Bacteria may also be removed but results may be expected to be
variable. For this reason, information on the removal of the full range of potential pathogenic
organisms in water and their correlation with indicator systems should be collected. The studies to
generate this information should be performed under conditions where all organisms are present in
the same experimental reactor system. The results of these studies should be verified In the field.
I-krdness Reduction
Hardness in water is caused primarily by the ions of calcium and magnesium and is more
likely to be a problem in groundwaters than in surface waters. The acceptance of hardness in
wafer is usually dictated by consumer sensitivity. Hardness levels above 100 mg/I as CaCO 3 are
usually objectionable while levels of 60 to 80 mg/I are considered moderate. Besides increasing
soap consumption, hardness may result in excessive scale formation in pipes. The reduction of
hardness in public water supplies is usually accomplished by the lime-soda process, although cation
exchange is an alternative.
There are two major variations of the lime-soda process. The low lime process reduces the
concentration of only the carbonate hardness associated with the calcium ion while the high
(excess) lime process also reduces the hardness associated with magnesium. A reduction In the
concentration of magnesium requires the pH of the water to be adjusted to near II, while calcium
hardness is reduced at near pH 10.3. Soda ash (Na 2 CO 3 ) is added to the wafer to reduce calcium In
the case of non-carbonate hardness, It Is generally accepted that microorganisms may be removed
with the calcium carbonate precipitate in the low lime process, white In the high lime process
inactivation accompanies removal.
The density of viruses in water may be reduced by either the low or high lime processes,
particularly if the p 1 -1 Is Increased to 10 or greater (Thayer and Sprout, 1966; Wentworth, et at.,
1968). In reviewing existing data, Sprout (1971) concluded that modest virus removal occurred with
the low time process, but at pH values near I I with the high lime process, 99 percent removal of
virus could be attained. In experiments with wast’ewater seeded with potiovirus type 3, high lime
treatment at pH 11.5 was found to remove 99.87 percent of the seeded virus (Berg, 1977).
-------
24
In an experimental wasfewater reclamation plant, lime treatment at pH 11.2 resulted in a
reduction of two orders of magnitude (2 log) of fecal coliform and a 3 LI log reduction of viruses.
Greater than a 5 log reduction of fecal coliform and a 4-5 log reduction of viruses was reported
with lime treatment at pH 11.5 (I-Iattingh, 1978). In another advanced wastewater treatment plant
designed for water reclamation, lime clarification (pH I 1.3) resulted in a 5 log removal of total and
fecal coliforms and a 98 to 99.9 percent reduction of viruses (McCarty, 1978; 1980). Recognizing
that this information is related to treating secondary wastewater effluent, it is recommended that
similar studies be performed to obtain data directly appucable to water supplies.
Filtration
Historically, water filtration was the first unit process recognized as a barrier to the
transmission of waterborne diseases. This observation was linked to the 1892 cholera outbreak in
Hamburg, Germany. The cities of Altona and Hamburg, Germany, obtained their drinking water
from the River Elbe and, although the water intake of Altona was downstream of Hamburg, the
populace of Altona was not severely affected, while 8,600 cholera deaths was reported in Hamburg,
The extent of the epidemic coincided with the boundary of the water distribution system of
Homburg and Altona. Altono used slow sand filters to treat its water white Hamburg provided only
settling (Hazen, 1913; 1-luisman and Wood, 1974).
Filtration Is defined as the removal of suspended and colloidal impurities from water by
passing It through a porous medium. A combination of physical and biolOgical processes ore
involved, including straining, sedimentation, adsorption, electrostatic binding, and microbial
activity. There ore three forms of water filtration: I) slow sand filtration; 2) rapId granular
filtration, including sand and dual or mixed media filtration; and 3) dlatomoceous earth filtration.
The removal of bacteria by slow sand filtration is well documented. Hazen (1913) reported
98.31 to 99.41 percent removal of the applied bacteria at the Lawrence, Massachusetts slow sand
filtration plant. Research described by Hazen at the Lawrence plant identified Important
operating parameters. The use of finer sand, deeper beds, and lower loading rates than normally
used improved effluent bacterial quality. Tests on filters reported by 1-luismon and Wood (1974)
showed removal of 999 to 99.99 percent for total bacteria while that for was 99 to
99.9 percent. In a pilot filter, operating at a flow rate of 0.12 rn/br, using clear, gravel pit water
to which residential wastewoter W03 added, Logsdon and Fox (1961) observed collform reductions of
-------
25
greater than 99 percent when the coliform densities in the applied water were as high as I ,000 to
10,000 per 100 ml. Burman (1962) researched the effects of temperature on slow sand filter
efficiency. During cold weather, removal of fecal and total coliforms amounts to only 41 and
88 percent, respectively.
Several workers have studies the removal of viruses by slow sand filtration. Depending on the
flow rate, Robeck, et al., (1962) obtained between 22 to 96 percent removal of pollovirus using
clean sand. Poynter and Slade (1977) obtained 98.25 percent removal of pollovirus at 5°C and a
flow rate of 0.5 rn/hr and a removal of 99.999 percent at I 1°C and a filtration rate of 0.2 rn/hr.
Taylor (1970), using a slow sand filter operating at a flow rate of 0.2 m/hr, observed greater than
99.9 percent removal of potiovirus. When a filter with sterilized sand was used, virus removal was
reduced, indicating the importance of using an established filter with its associated biological
growth. Taylor (1973) also found that increasing the filter rate from 0.2 to 0.4 rn/hr reduced virus
removal from 99.8 to 91 percent.
Although specific information seems to be lacking, It Is generally accepted that a properly
operated slow sand filter will effectively remove protozoan cysts and helminth ova.
Slow sand filtration traditionally has been a unit process used alone, or in combination with
terminal disinfection. Under most conditions, the process appears effective In terms of removing
microorganisms. The mechanisms of removal of organisms by slow sand filtration ore not
completely understood, although the presence of a biological growth appears to Improve removals.
Viruses appear to be more susceptible to removal than bacteria. The slow sand filtration process is
adversely affected by low temperature, Inadequate sand depths, immaturity of the filter, and
excessive flow rates. More definitive Information Is required if slow sand filtration is to be
employed with confidence for the removal of microorganisms. For example, information Is needed
on the microbial load that a slow sand filter can handle under different operational conditions. The
possibility of endotoxin formation by the microbial growth associated with slow sand filters should
also be evaluated.
In comparison to slow sand filtration, rapid granular filtration has riot been found to be
particularly effective In removing microorganisms when used without preconditloning of the feed
water by coogulaflon-flocculation. Engelbrecht, et al., (1979) studIed the removal of yeasts, acid-
fast organisms, and total coliforms using dechlorinated tap water with a I percent raw municipal
-------
26
wastewater inoculum, but without prior coagulation-flocculation. While yeasts showed a consistent
removal of 87 to 97 percent throughout two 18 hour filter runs, acid-fast organisms and total
coliforms showed initial removals of 55 to 70 percent which decreased to only 40 to 45 percent by
the end of the experiments. Rapid sand filtration of water, without prior coagulation-flocculation,
is not considered an effective process for the consistent removal of bacteria, It is possible,
however, to achieve a significant removal of bacteria by filtration provided the water is pretreated
by coagulation-flocculation and controlled for optimum turbidity removal (1980).
Robeck, et al., (1962) showed that uncoagulated waters applied to a rapid granular filter, sand
alone or sand plus anthracite, resulted In only I to 40 percent removal of pollovirus, while
coagulated water applied to the filter showed 90 to 99 percent removal. Removals exceeded 99
percent when the water was coagulated, settled, and filtered. In other virus studies with
uncoogulated waters applied to a filter, removals of bacterial and enteric viruses were between 0
to 98 percent (Berg, et al., 1968; Guy, et at., 1977). Other Investigators have shown that the
removal of viruses by filtration is improved if the filter is impregnated with alum floc (Carlson, et
at., 1942; Gllcreas and Kelly, 1955) or if coagulation precedes filtration (Gllcreas and Kelley, 1955;
Kempf, et at., 1942).
Logsdon, et at., (1981) showed that with proper coagulation and aluminum sulfate, prior to
filtration, Giardia muris cysts could be removed in excess of 95 percent while removals were
erratic when the raw water was applied directly to the filter. In a pilot study, Baylls, etol, (1936)
reported that greater than 99.99 percent removals of E. histolytica cysts could be achieved by
filtration if the water was pretreated by proper coagulation-flocculation.
It Is evident that in terms of removing microorganisms from water, coagulation-flocculation,
with or without sedimentation, followed by rapid granular filtration should be treated as a single
process. Without coupling these two processes, the removal of microorganisms by filtration Is
Ineffective and/or erratic. With proper pretreatment of the Influent water by coagulation-
flocculation, it is possible to achieve removal of microorganisms In excess of 90 percent. The
importance of the overall process of coagulation-flocculation plus filtration cannot be overstated.
In view of the resistance of the cysts of Giardia and some of the enteric viruses to chlorination,
this combined process takes on increased Importance as the final barrier to the water transmission
of these organisms. -
-------
27
The applicability of diatomoceous earth filtration for removing microorganisms from water
appears to depend greatly on the size of the microorganisms and the pore size of the filter media.
Hunter, et al., (l96 ) studied the removal of coliforms using different types of filter media. The
most permeable filter aid gave 90 percent removal while that with the finest grade media was in
excess of 99.8 percent. Using a commercial grade diatomoceous earth having a mean particle
diameter of 49 im and a water to which the bacterial virus MS2 was added, Amirhor and
Engelbrecht (1975, 1975) found that there was little, if any, virus removal. However, greater than
95 percent virus removal from water by diafomaceous earth filtration was achieved if the filter
medium was precoated with a cationic polyelectrolyte or if the polyelectrolyte was added to the
water prior to filtration. Logsdon, et al., (1981) reviewed the experimental work performed by the
U.S. War Department (l9M). From these studies, it was concluded that it was possible to remove
essentially all cysts of E. histolytica applied to a diatomaceous earth filter, provided the filter was
properly precoated. Reporting on their own work with Giardia muris cysts, diatomaceous earth
filtration was found to consistently remove greater than 99.5 percent of the cysts when the filter
was operated properly. Proper operation required a precoaf of filter aid (1.0 kg/rn 3 ) and a
continuous adequate body feed of filter aid, regardless of the type of media used.
From these results, it may be concluded that even without pretreatment of the water with a
coagulant, diatomoceous earth filtration is capable of removing 90 percent or more of the
coliforms as well as Giardia and Entomoeba cysts in water. However, adequate removal of viruses
cannot be assumed without proper pretreatment.
Actlvat.d Cuban
Activated carbon is used In water treatment for the control of soluble organic matter. The
usual placement of the activated carbon treatment process is after filtration and prior to
chlorination when granular carbon (GAC) is used. When powdered activated carbon (PAC) is used,
it is applied prior to filtration.
Of particular Interest In considering the microbial quality of a water is the ability of
activated carbon to adsorb and later desorb bacteria and viruses, to provide a surface for bacterial
growth, and to concentrate organic materials which may be available as substrate for subsequent
bacterial growth. Thus, it Is possible that activated carbon may act as a reservoir for viruses and
bacteria which may be desorbed or sioughed Into the water supply under proper conditions.
-------
28
Further, some metabolizing bacteria can produce endotoxins which may be added to the water. A
major unanswered question with GAC relates to the possible production and release of small carbon
particles which may affect the microbial quality of the effluent water which, in turn, may affect
the finished water, e.g. the sedimentation of carbon particles in the distribution system which may
facilitate microbial colonization.
Klotz (1979), Klotz, et al., (1975, 1976, 1976), and Werner (1978, 1979) have studied the
population and metabolic activity of the microorganisms in GAC filters used in treating potable
water. Samples of activated carbon have been reported to support a range of 10 to 108 bacteria
per gram wet weight of carbon, depending on the treatment processes preceding the GAC filter and
the degradability of organic substances present in the water. Colony counts in GAC effluents were
in the range of l0 per ml of water. A wide variety of bacterial types, including 26 species from
12 genera, were identified In the activated carbon. Most species Identified belonged to the genus
Pseudomonas ; the next most common genus was Bacillus . Fungi and yeasts were recovered only
rarely. It was concluded that bacterial activity occurs in the filter at the expense of the organic
substances adsorbed to the carbon. The most readily degradable substances are used by the
bacteria with the result that the useful life of the carbon Is extended. Further, it was postulated
that the capacity of the treated water to support microbial growth in the distribution system would
be reduced because of the removal of organic substances by the activated carbon.
The capability of GAC to remove viruses from water by adsorption has been demonstrated
(Cookson, 1969; Cookson and North, 1967; and Olivierl, et al., 1977). It has also been observed that
viruses adsorbed on carbon may be later desorbed as the adsorption sites for organic matter on the
carbon become exhausted (Sproul, et al., 1967). Gerba, et al., (1975) found that the removal of
poliovirus type I by GAC can be improved by lowering the p 1 -i to 3.5-4.5 and reducing the
concentration of organic matter in the water by pretreatment with lime coagulation. Cl Iver (1971)
found that the retention of virus on activated carbon could be as high as 90 percent Initially, but
removal decreased as operation time increased. Relatively little desorption of viruses seemed to
occur when water was treated rather than wastewater. In studying GAC as an advanced
wastewater treatment process, Parkhurst (1977) observed 80 to 90 percent removal of viruses.
After reviewing the available literature, Engelbrecht (1976) concluded that activated carbon
adsorption, when used as an advanced wostewater treatment process, had variable ability for virus
-------
29
removal ranging between 10-99 percent. A report to the U.S. Department of the Interior concluded
that activated carbon should not be relied upon for virus removal; however, it had Important
properties for removal of organic matter, necessary for effective disinfection (Cuip, et at., 1979).
Disinfection
The goal of disinfection is to assure a water supply free of pathogenic microorganisms. Many
of the water treatment unit operations previously discussed have the capability of removing
microorganisms but it Is unlikely that they will produce a water that Is consistently free of
microorganisms, including enteric pathogens. For this reason, the last treatment barrier to the
transmission of pathogenic microorganisms Is terminal disinfection, it should be recognized,
however, that the treatment units employed prior to disinfection do serve a purpose. in addition to
possibly reducing the density of microorganisms in the water prior to disinfection, the unit
treatment operations previously discussed are designed to remove substances which might
otherwise interfere with effective disinfection. Further, there Is evidence that the cysts of some
protozoa, the ova of certain helmlnths and some enteric viruses are resistant to disinfection as
normally practiced with municipal water supplies. The elimination of these microorganisms is
largely dependent upon The other unit operations making up a properly designed and operated water
treatment system.
Many different disinfectants are available for use in treating water supplies. Chlorine,
ozone, and chlorine dioxide are the major dislnfectonts In use today and will, therefore, be
discussed in detail. However, because iodine and UV light ore being considered as disinfectants for
small water supply systems, they also will be discussed. Much of the disinfection information
presented is discussed In greater detail In two review publications, Miller, ., (1978) and
National Academy of Sciences (1980). More recently, Hoff and Geldreich (1981), in comparing the
biocidol efficiency of free chlorine, ozone, chloride dioxide and chioramines, pointed out the
problem associated In attempting to quantitatively assess the efficiency of different disinfectants
and the Influence of various factors and the difficulties encountered in applying laboratory results
an disinfectants to field use.
The water treatment ptents using chlorine as a disinfectant number In the thousands. In 1971,
it was determined that ozone was being used somewhere in the water treatment system In five
-------
30
facilities in the U.S. and in 1,039 facilities worldwide. While most applications for ozone are for
disinfection, ozone may also be used to assist in the control of organic matter, iron and manganese,
color, taste and odor, algae growth, and suspended solids. Chlorine dioxide was reportedly used in
the U.S. in 84 water treatment facilities of which only one facility used chlorine dioxide alone as a
terminal disinfectant. The other facilities supplemented chlorine dioxide disinfection with chlorine
to maintain a chlorine residual. In Europe, 495 facilities were identified as using chlorine dioxide.
Other common uses of chlorine dioxide include Iron and manganese removal and control of taste
and odor (MIller, etal., 1978).
The effectiveness of chemical disinfection is a function of the nature of the organism to be
inactivated, the quality of the water, the type and concentration of the disinfectant, the exposure
or contact time, and the temperature of the water. Because of the importance of disinfectant
concentration and contact time, the product of these two variables is often used to compare the
biocidal efficiency of different disinfectants. Using batch laboratory data, disinfectant concentra-
tion (c) multiplied by contact time (t) yields a c•t value for a predetermined percentage
inactivation of the test organism (see Table 3 as an example of c et values required for 99 percent
inactivation of E. coU and poliovirus I by chlorine).
An issue has been raised with respect to ‘the potential existence of microorganism strains
with disinfectant resistance different from the laboratory strains used to develop information on
disinfection efficiency. For example, chlorine-resistant strains of poliovirus I, produced by
repeated exposure to free chlorine and subculture, have been reported by Bates, et al., (1977). The
isolation of a poliovirus with increased resistance to chlorine from finished drinking water has also
been reported (Shaffer, 1980). However, attempts to confirm the results of Bates (1977) by
Engelbrecht, ., (1978) were unsuccessful. The use of similar techniques with E. coIl also did
not result in the development of a strain with increased resistance to chlorine (Hoas and Morrison,
1981).
Chlorine. Depending on the pH, temperature and the presence of ammonia and organic
amines, chlorine will react with water to form hypochlorous acid (HOd), hypochlorite (OCI),
monochloramine (NH 2 CI), dichloramine (NHCI 2 ), and organic chloramines (R-NHCI). These
chemical species exert different biocidal effects. The following order of disinfection efficiency of
the different forms of chlorine, based upon the inactivation of coliforms, has been generally
accepted (Feng, 1966 and Morris, 1966):
-------
31
HOd> OCI>NHC1 2 >NH 2 CI >R-NHCI
This order of disinfectant efficiency is in general agreement with the available data. Butterfield,
et at., (1943), Clarke, et al., (1956), and Weidenkopf, (1958), showed that free chlorine residuals
were more effective at low pH than at high pH. The effect of ammonia (Butterfield, (948; Chcing,
1971; Lothrop and Sproul, 1969; and Stringer, 1975) and organic amines (Feng, 1966 and Kjellander
and Lund, 1965) on disinfection efficiency shows that the dichloramine form is more effective than
monochioramine, which in turn, is more effective than organic chloramine. Estimates of the
disinfection efficiency of the different forms of chlorine vary, depending upon the test organisms.
However, from this review of the disinfection literature, Morris ((966) estimated that 0C1 was 50
to 100 times less effective than HOCI and NH 2 CI was 2 to 200 times less effective than HOCI.
Esposito (1974) and Esposito, j., (1974) showed dichioramine was 35 times more bocteriocidal
than monchioramine. A summary of disinfection data for f,. !! and poliovirus type I by different
chemical forms of chlorine is presented In Table 3 ((980).
The presence of turbidity or particulate matter, depending upon its nature, can also adversely
affect disinfection. Chang, et c ii., ((960) demonstrated that enterlc viruses and bacteria ingested
by aquatic nematodes can be completely protected against chlorination. Tracey, j., (1966)
attributed the survival of coliforms in a chlorinated water supply to ingestion by crustacea present
in the water. Hoff (1978) cIted a report that showed that the survlval of coliform organisms in a
water supply was due to their being embedded in particulate matter. Reporting on his own
research, Hoff found a much higher degree of protection provided to poliovirus type I when
associated with cell material than In the presence of bentonite or aluminum phosphate. It was
found In the same study that a freely suspended laboratory-grown culture of E. was inactivated
much more rapidly by chlorine than was the same culture when associated with primary wastewater
effluent solids that had been washed free of chlorine demand. Hejkal, et al., (1979) also showed
that pollovirus In the presence of fecal material Is provided substantial protection against
inactivation by chlorine. More recently, Hejka I, et al., (1981) reported that over 90 percent of the
enteroviruses In primary wastewater effluent were either freely suspended or associated with
particles smaller than 0.3 m i. They also observed that the percentage of infectious solids-
associated viruses increased as a result of chlorination, indicating that there was protection due to
-------
32
Table 3. Dosages of Various Chlorine Species Requiredjor
99% inactivation of Escherlchia cofl and Poliovlrus 1
Microorganism
Disinfecting
Agent
Concen-
tration
mg/I
Contact
Time,
mm
b
c.t
p1-I
Tempe a-
ture, C
E.coI i
Hypochlorous
acid (HOCI)
Hypochlorite
Ion (0CI)
Monochioramine
(NH 2 CI)
0.1
1.0
1.0
1.0
0.4
0.92
175.0
64
0.04
0.92
175.0
64
6.0
10.0
9.0
9.0
5
5
5
15
1.2
33.5
40.2
9.0
25
Poliovirus I
Dich loramine
(NHCI 2 )
Hypoch lorous
acid (HOCI)
1.0
1.0
0.5
5.5
1.0
2.1
5.5
1.0
1.05
4.5
6.0
6.0
15
0
5
1.0
2.1
2.1
6.0
5
1.0
1.0
.0
6.0
15
Hypochiorite
ion (0Cl)
0.5
1.0
21
3.5
10.5
3.5
10.0
10.0
5
15
Monoch loramine
(NH 2 CI)
.
10
10
90
32
900
320
9.0
9.0
15
25
Dichioramine
(NHCI 2
tOO
100
140
50
14,000
5,000
4.5
4.5
5
IS
°From National Academy of Sciences
bCor. fration of compound times contact time
-------
33
association with solids. There is some indication that inorganic solids do not provide protection. A
decrease In efficiency was not observed when koolinite and aluminum oxide were tested with
coliphage Ti and poliovirus (Broordman and Sproul, 1977). However, addition of bentonite clay to
cultures of MS2 coliphage resulted in increased survival (Stogg, et al., (977).
A residual chlorine concentration of 4 mg/I at p 1-I 7.2 to 73 and 25°C was found by Chang
(1978) to inactIvate 999 percent of pathogenic Naegleria cysts in 10 mm. Sproul, et al., (1981),
using free residual chlorine, showed that 99 percent inactivation of Noeglerla gruberl cysts could
be obtained with an average (c t) value of 12. I mg mini) at pl-I 7.0 and 25°C. StrInger, et at.,
(1975) achIeved about 99 percent Inactivation of E. histolytica in 10 mm at pH 6.0 wIth 2.0 mg/I
residual chlorine at 27°C. Jarrol I, ., (1981) reported that exposure to 1.5 mg/I chlorine for
(0 minutes at 25°C killed all Glardla lomblia cysts at pH 6, 7, and 8, but exposure to I mg/I
chlorine for 60 minutes at 5°C faIled to kill all the cysts at these three pH values. At 5°C, a
chlorine concentration of 4 mg/I killed all the cysts at all three pH values after 60 minutes, but not
after 30 mInutes. Rice, et al., (1982) reported similar results and also showed that G. lamblia and
muris were inactivated at similar rates, Indicating that muris may be a suitable model for G.
lamblia in chemical disinfection studies.
Caution must be exercised In relying too heavily on extrapolation of laboratory data to
expectation of the process in the field. A recent set of papers points out Just how tentative the
existing knowledge of dlsinfectlçn really Is. Scarpino, et al., (1972) reported the seemingly
anomalous result that hypochlorite Ion was more effective than hypochiorous acid In Inactivating
Pollovirus type I • These authors Indicated that the borate-KCI-NaOH buffer system used with
their laboratory experiments may hove caused the unusual results. A more recent study has shown
that the presence of sodium Ions con Increase the rate of inactivation of ll by chlorine at high
p1-I, with the rate being similar to that of HOCI at low pH (Haas, 1981 and Haas and Zapkin, (98)).
It was postulated that this effect was due to the formation of a NoOCI Ion pair. A similar
explanation for the observation of Scarpino, et al., ((972) has been proposed by Jensen, et al.,
(1980). LIu, et a)., (1971) studIed the response of 20 strains of human enteric viruses to chlorine
and Concluded from extrapolated laboratory data that “as groups, the reoviruses are definitely the
least resistant to chlorine treatment; both odenoviruses and ochoviruses are relatively less
-------
34
resistant; and the polioviruses and coxsackieviruses are the most resistant.” This research was
performed with alum coagulated/sand filtered Potomac River water at pH 7.8, 2°C, and with
0.5 mg/I free available chlorine. Because of the importance of the results reported by LIu, el al. ,
(1971) and Scarpino, el al., (1972), an independent study was performed to show the effect of virus
type, pH, and potassium chloride on the kinetics of virus inactivation (Engelbrechf, el al., 1980).
Based upon carefully controlled laboratory experiments using six different enteric viruses, several
observations were of particular interest. First, it was found that the different virus types
demonstrated a wide range of susceptibility to chlorine disinfection. With 0.5 mg/I free available
chorine, the rate of inactivation was greater at pH 6 than at pH 10. On the other hand, poliovirus I
was inactivated at a higher rate at pH 7.8 than at 6. Further, echovirus 5 was the least resistant of
the six viruses at pH 6; while at pH 10, it was the second most resistant. Thus, the relative
susceptibilities of the different viruses were affected differently by a change in pH, suggesting
that the pH influenced both the form of chlorine present and the susceptibility of the different
viruses to chlorine. Finally, It was shown that the addition of KCI to experiments with poliovirus I
decreased the time required for 99 percent inactivation by sixfold at p1-f 6 and 50 fold at pH 10, as
opposed to experiments without KCI.
Ozone. Ozone has been found to be a more powerful disinfectant than chlorine (Fluegge, et
al., 1979 and Munger, flj., 1977) and iodine (Kruse, et al., 1970) and at least as powerful as
chlorine dioxide (1-fol luta and Unger, 1 95L 1 ) .
The pH of a water affects the rate of ozone decomposition, such that at a higher p1-f,
decomposition of ozone to secondary oxidants increases. This results in a reduced disinfection
efficiency for a given ozone dose. Farooq, [ ., (1977) adJusted the ozone dose to provide a
constant disinfection efficiency. Therefore, the effect of pH seems to be on the stability of the
residual rather than, as with chlorine, changing the form of the disinfectant. Increases In water
temperature appear either to have little Influence on disinfection efficiency (Kinman, 1975) or to
Increase efficiency (Forooq, 1977).
Studies on the effects of turbidity on ozone disinfection have produced variable results.
Nomie (1976), after reviewing the literature on ozone disinfection of secondary wostewater
effluents, concluded that the greatest factor influencing the efficiency of ozonatlon was the
suspended solids content of the water. Block (1977) found that kaoHnlte clay turbidity protected
-------
35
pollovirus I from Inactivation by ozone. However, Cerkinsky and Trahtman (1972) and Gevaudan
(1971) concluded that turbidity had no effect on ozone disinfection. Walsh, et ci., (1980) and
Boyce, et ci., (1981) found that the inactivation of , poliovirus I, and coxsackievirus A9
adsorbed to or embedded within hydrated aluminum oxide or bentonite clay did not vary
significantly from that obtained with the organisms in a suspended state at turbidity levels of
5 NTtJ or less. Foster, !t si. , (1980) found that poliovirus I, porcine plcornavirus type 3, and
coliforms, when associated with fecol matter, were reduced to less than detectable levels using an
initial ozone concentration of 0.01 and 0.22 mg/I and a 30 sec contact time. Emerson, et at., (1982)
found that significant protection was given to poliovirus I and coxsackievirus A9 when associated
with HEp2 cells. Apparently, the effects of turbidity on disinfection with ozone are as complex as
with chlorine.
Some types of soluble organic matter in water reduce the ozone residual, thereby decreasing
disinfection efficiency. Dosage to a point in excess of the demand exerted by organic matter is
important in obtaining efficient disinfection. Inactivation of microorganisms occurs rapidly once
an ozone residual Is established (Farooq, et ci., 1978).
The method of contacting ozone with water also affects the disinfection efficiency. it has
been observed that disinfection occurs when ozone bubbles are present in water even in the absence
of a measurable ozone residual (Farooq, et at., 1977) and that the rate of inactivation under these
conditions Is proportional to the concentration of ozone In the bubble (Block, 1977).
Data on the Inactivation of pure cultures of microorganisms Indicate that the rate of
Inactivation Is very rapid (Notional Academy of ScIences, 1977). A review of the data on the
Inactivation of E. coil by ozone showed that for 99 percent inactivation, the C • t values range from
0.001 to 0.004mg mm/I at 12°C, to 0.006 to 0.022 at 1°C (Table 4). For poUovlrus I, c’ t values
ranged from less than 0.04 to 5°C, to 0.005 to 0.42 mg mm/i at 20°-25°C for 99 percent
inactivation. In studying the Inactivation of six different enteric viruses by ozone, Evison ($977)
found that susceptibility varied in that different viruses required different amounts of ozone to
achieve the same degree of Inactivation (Table 5). Newton and Jones (1949) demonstrated 98 to
99 percent inactivation of Entamoeba histolytica in 5 minutes with 0.3 mg/i ozone. No work on the
effectiveness of ozone as an ontihelminthc agent appears to have been reported (National
Academy of ScIences, 1977).
-------
36
Table k. Concentration of Ozone and Contact Time Necess 8 ry for 99%
Inactivation of Escherichia coli and Polio I Virus
Test
Microorganis
Ozone,
ms mg/I
Contact
Time, mm
b
c.t
p1-I
Temperature,
C
E.coli
0.07
0.065
O.k
0.01
0.01
0.0006
0.0023
0.0125
0.083
0.33
0.50
0.275
0.35
.7
1.03
0.33
0.006
0.022
0.2
0.027
0.035
0.001
0.002
0.004
7.2
7.2
7.2
6.0
6.0
7.0
7.0
7.0
I
I
I
II
II
(2
12
12
PolIo!
<0.3
0.245
0.042
<0.03
0.13
0.50
(0
0.16
<0.04
0.12
0.42
< 0.005
7.2
7.0
7.0
7.0
5
24
25
20
°From National Academy of Sciences
bConcentration of ozone times contact time in mg”min/I
-------
37
Table 5. Ozone Required for Inactivat1 n 8 f Viruses in 10 Minutes
at pH 7.0 and 25 C
Ozone Required, mq/l
99.9%
99%
Virus
inactivation
Inactivation
CoxsockieB3
Polio 3
0.6
0.22
0.095
0.082
Polio I
0.095
0.042
Echo I
0.086
o.o’ ’e
CoxsackieB5
Polio 2
0.076
0.052
0.053
0.039
°From National Academy of Sciences
-------
38
The best conditions for effective disinfection with ozone, i.e., a water free from particulates
and organic matter, can assure inactivation of bacteria and viruses in excess of 99 percent.
Because of limited or nonexistent data, the response of protozoan cysts and helminth ova to ozone
disinfection is uncertain. A major problem with ozone as a water disinfectant is that a residual
cannot be maintained in the distribution system.
Chlorine Dioxide. As a disinfectant, chlorine dioxide provides two advantages over chlorine.
Chlorine dioxide does not react with ammonia, nor does it react with organic matter to form
trihalornethane. However, like ozone, chlorine dioxide is unstable and must be generated on site.
As early as 1940, studies on the effectiveness of chlorine dioxide as a water disinfectant
showed it to be superior to chlorine. While these observations give good qualitative evidence, it
was not until the mid I 960’s that detection and generation procedures for chlorine dioxide became
reliable (National Academy of Sciences, 1980). More recent work has substantiated the earlier
claims that chlorine dioxide is more effective than chlorine (Baylis, et ci., 1936; Moffa, et at.,
1975) in inactivating E. coh in wastewater effluents. At pH 7, chlorine dioxide has been found to
have a similar viricidal effectiveness to that of free available chlorine; however, as the pH was
increased, chlorine dioxide proved to be more effective than chlorine (Warriner, 1967). The
observation that increased pH increases chlorine dioxide effectiveness has been confirmed by
Cronier, etal., (1978).
In a study designed to compare the bactericidal effectiveness of chlorine, Benarde, et ci.,
(1965) observed greater than 99 percent inactivation of 2 jj in 15 sec with 0.25 mg/I chlorine
dioxide, while under the same conditions, chlorine required almost 5 mm to accomplish the same
degree of inactivation. Chlorine dioxide was also found to be significantly more effective than
chlorine when high levels of organic and nitrogenous materials were present in the water. The
efficiency of chlorine dioxide as a disinfectant is also affected by water temperature (Benarde, et
ci., 1976). For 99 percent inactivation of E. coIl with 0.25 mg/I chlorine dioxide, the contact times
were 190, 74, 14, and 16 sec for water temperature of 0, 100, 200, and 30°C, respectively. The
high degree of effectiveness of chlorine dioxide in inactivating was also demonstrated by
Cronier (1977). Table 6 gives a portion of the data collected by Bernarde, el ci i., (1967) and
Cronier (1977), as presented by the National Academy Sciences (National Academy of Sciences,
1980).
Chlorine dioxide is also a rapid viricide (Table 7). The data collected by Warriner (1967)
Indicate that the rate of inactivation of poliovirus 3 increased as the pH of the water increased
-------
39
Table 6. Concentrations of Chlorine Dioxide and Contact Times Necessary
for 99% Inactivation of Escherichia coli°
Test
Microorganism
Chi rine
Dioxide,
mg/I
Contact
Time,
mm
b
c.t
pH
Tempera-
ture
C
E.coli
1fi hIy
0.25
0.50
1.8
0.83
0.45
0.41
6.5
6.5
5
5
isolated
0.75
0.50
0.38
6.5
5
from feces)
0.25
0.50
0.75
0.25
0.50
0.75
0.25
0.50
0.75
1.2
0.47
0.3
0.68
0.35
0.25
0.27
0.22
0.15
0.30
0.24
0.23
0.17
0.18
0.19
0.07
0.11
0.11
6.5
6.5
6.5
6.5
6,5
6.5
6.5
6.5
6.5
10
10
10
20
20
20
32
32
32
E. coli
1AT C 11229)
0.30
0.50
0.80
0.30
0.50
0.80
0.30
0.50
0.80
1.8
0.98
0.58
1.3
0.75
0.47
0.98
0.55
0.35
0.54
0.49
0.41
0.39
0.38
0.38
0.29
0.28
0.28
7.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
5
5
5
15
15
15
25
25
25
°From National Academy of Sciences
bConcentration of chlorine dioxide times contact time in mg.min/l
-------
40
Table 7. Concentrations of Chlorine Dioxide and Contact Times
Necessary for 99% Inactivation of Viruses
Test
Microorganism
Chlorine
Dioxide,
mg/I
Contact
Time,
mm
b
c’t
pH
Tempera-
t re,
C
Po liovirus3
0.5
0.5
0.5
1.6
1.6
1.6
5.0
5.0
0.25
5.0
1.0
0.25
2.5
2.5
0.125
8.00
(.60
0.40
5.6
7.2
8.5
5.6
7.1
8.0
•
20
20
20
20
20
20
Poliovirus I
0.3
0.5
0.8
0.3
0.5
0.8
0.3
0.5
0.8
(6.6
(2.0
6.8
4.2
2.5
1.7
3.6
2.0
(.5
5.0
6.0
5.4
1.3
(.25
1.4
.08
1.0
.2
7.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
5
5
5
15
(5
(5
25
25
25
Coxsacklavlrus A9
0.3
0.5
0.8
1 .2
0.05
0.25
0.4
0.34
0.20
7.0
7.0
7.0
(5
(5
15
°From Natlonol Academy of Sciences
bConcentratlon of chlorine dioxid times contact time In mg .mInfl
-------
41
from 5.6 to 8.5. Cronier (1977) found both poliovirus I and coxsoclcievirus A9 to be more resistont
to chlorine dioxide inactivation than !J. However, all three organisms were sensitive to
chlorine dioxide at a concentration less than I mg/I. Chlorine dioxide was also observed to be more
effective as a viriclde at higher pH values. In comparing efficacies of different disinfectants,
Cronler (1977) concluded that on a weight basis, chlorine dioxide was similar to hypochlorous acid
and better than hypochiorite ion, monochloramine, and dlchloramjne.
In an investigation using chlorine dioxide at pH 7.0 and 25°C, Sproul, et al., (1981) found that
99 percent inactivation of N. gruberl was obtained with a C • t value of 6.1 mg mm/I. No data are
available on the inactivation of helminth ova with chlorine dioxide. Additional data on the
inactivation of these agents with chlorine dioxide are needed.
Iodine. Because the equipment is easy to Install and the cost of operation and maintenance Is
low, iodine disinfection is attractive for small water supply systems. However, for public health
reasons, It Is believed that lodinated water should not be consumed over a long period of time.
Therefore, iodine disinfection seems best suited for non-resident situation, e.g., parks, summer
camps, etc.
In the pH range 5 to 8, the active chemical species of iodine in the disinfection process have
been identified as diatomic iodine (12), hypolodus acid (HOl), and hydrated iodine cation (H 2 OF )
(Cramer, et ci., 1976 and WhIte, 1972). At pH above about 8, Iodine species of lesser disinfection
efficiency are formed (hypolodite (10 ), and lodate (103 )).
The free available Iodine species at or near neutral pH are effective bacterlocides.
Chambers, et al., (1952) studied the Inactivation of Enterobactor aerogenes , E. coIl, Streptococcus
foecalls , and various species of Salmonella and Shigella . Total Inactivation of l0 organisms/mi
after one minute exposure required 0.6 mg/I residual iodine under the best conditions (pH 6.5, 20°
to 26°C) and 4.3 mg/I under the worst conditions (pH 9.15, 2° to 5°C). All tests were made with
Iodine-demand free, buffered water. Using tap water at 25°C and pH 8.1 to 8.5, Chang and Morris
(1953) found that 2 to 5mg/I iodine required 10 mm to inactIvate 6 logs of Qll. Further, S.
typhosa , S. schottmuilerl, Shigella dysenterlae and mixed cultures of coliforms were as sensitive to
Iodine as was j. VIbrIo cholerae was more sensitive than E. coil. Comparing the efficiency of
chlorine to Iodine, Berg (1966) found that Iodine was approximately one-fifth that of hypochiorous
acid but ten times as effective as hypoch lorlte Ion. In one field test on a small system serving
-------
42
700 persons, I mg/I of free iodine was found to be effective in rendering the supply free from
coliforms (Black, etal., l9 5).
The viricidal efficiency of iodine appears to be highly sensitive to the type of virus tested.
At 25°C, approximately 0 mm were required to inactivate 5 logs of coxsackievirus Bl and I mg/I
residual iodine. iodafe and iodite were ineffective as viricides (Chang, 1958). Clarke, eta!., (l9 4)
found that it required 18 hours to inactivate 99 percent of a culture of coxsackievirus A9, while it
required only I mm to inactivate 99 percent of a culture of E. coil. Thus, the relative efficiency of
iodine vs. chlorine is also organism dependent. Zoeteman (1972) concluded that while iodine is less
effective than free chlorine, it is more effective than combined chlorine as a viricide. However,
Mcihnel (1977) concluded that iodine was slightly more efficient than free chlorine at pH 5 to 8.5
against a variety of viruses.
Iodine is effective at high concentrations against E. histolytica cysts. About 8 mg/I of iodine
at pH 8 or less can inactivate 5 logs of the cysts in 10 mi i i. Iodine is more effective than HOl
against the cysts (Chang, 1958 and Morris, ef at., 1953).
From a review of the available Information, it was concluded that iodine is an effective
bacteriocide over a relatively wide pH range and that residuals of 0.5 to 1.0 mg/I of free iodine was
sufficient to produce a safe drinking water with apparently no adverse effects on human health
(National Academy of Sciences, 1980). However, it was pointed out that there was a need for
additional studies on the possible health effects of Increased iodine intake with susceptible
individuals in the population. Because of this concern, on EPA policy (20 February 1973),
recommending against the use of iodine for the disinfection of public water supplies serving
permanent populations, is still in effect. Iodine may be used, however, in emergency situations or
under conditions where the period of consumption of lodinated water Is brief.
Ultraviolet Light. Because ultraviolet (UV) light Is a physical agent rather than a chemical
agent, its effectiveness as a disinfectant is little affected by the presence of organk matter or
ammonia, or the pH of the water. This Is unlike the chemical disinfectants, such as ozone and
chlorIne, which are highly subject to the lnflue,ces of these water propertIes. The prediction of
the effectiveness of UV is based only on dose rate and contact time. To determine the effects of
water quality on disinfection, only the absorbance at 254 nm and the reactor geometry are
required.
-------
43
Much of the comparative work on the Inactivation of waterborne microorganisms by UV light
has been performed in continuous flow reactors of proprietary design. Unfortunately, because of
the complexity of the disinfection reaction in non-uniform UV intensify fields and unknown mixing
patterns in the reactors studied, and due to lack of water quality data presented by many early
researchers, it Is difficult enough to estimate the dose response in a given study, much less
compare the work of the various researchers. Only recently have attempts been made to correct
for the effects of water quality, as measured by WV absorbance, on disinfection of wastewater
effluents (Rober and Hoot, 1975 and Severin, 1980). However, even when dye tracer studies were
performed to evaluate the effects of mixing in the reactors, the dual effect of non-uniform
intensity fields and mixing on disinfection precluded adequate modeling which could be extra-
polated to reactors of different design (Johnson, ci., 1979 and Klotz, et ci., 1975). Unfortun-
ately, a major study on the dose response of various waterborne microorganisms to WV did not
Include the effectó of WV light reflection within the test system. Consequently, the doses reported
are probably low Table 8 gIves the data compiled by Morris (1972) for the Lethal dose ( w
sec/cm 2 ) to inaclivate a variety of microorganisms.
In the report prepared by the National Academy of Sciences (1977), It was concluded that
current WV light technology limits its use to small water supply systems and that equipment
maintenance remains a problem. It was also pointed out that WV radiation produces no residual,
further limiting Its use. On the other hand, WV radiation was Judged to be an effective Inactivating
agent for bacteria and viruses; no conclusion was reached as to its effectiveness against protozoan
cysts because of the lack of information. In respect to the latter, Rice and Hoff (1981) recently
reported that cysts of lambilo are extremely resistant to Inactivation by WV light. The UV dose
for Inactivation far exceeded the minimum dosage of 16,000 31W sec/cm 2 recommended by the U.S.
Public Health Service for water disinfection as well as the maximum design dose range of
25,000-34,000 w sec/cm 2 for most commercial WV treatment units. Situations In which C.
tamblia cysts may occur in the source water, e.g., high quality, low turbidity surface waters in
remote upland areas with a substantial wildlife population, have been considered as suitable for WV
disinfection. However, the apparently poor cysticidal capabilities of WV light would seem to
preclude its use in these situations.
-------
k4
Table 8. Ultraviolet Energy Neces ,ary to Inactivate
Various Organisms
Test Lethal Dos !,
Microorganism ( w .s/cm )
Escherichia coil 360
Staphylococcus aureus 210
Serrotia marcescens 290
Sorcina lutea I ,250
Bacillus globiggii spores I ,300
T3coliphage 160
Poliovirus 780
Vacciriia virus 30
Semi Ike Forrest virus 470
EMC virus 650
°From Naflonal Academy of Sciences
-------
45
Ii is obvious that much remains to be learned about the efficiency of UV disinfection,
including its engineering application to water supplies. In the case of the latter, insufficient
information exists to predict the effects of mixing in non-uniform light intensity fields on
disinfection efficiency. Further, much of the early Information on the efficacy of UV light against
certain organisms did not account for UV light attenuation within the test waters. When the
complex effects of mixing within non-uniform intensity fields on the dose response of micro-
organisms is understood, UV light may become a strong alternative to chlorine for small water
systems.
By-Product Formation. In considering the use of chlorine as the disinfectant of a water
supply, it is Important to recognize that free chlorine may result in the formation of trihalo-
methanes (THM). The U.S. EPA has established a maximum contaminant level (MCL) for THM and
the use of chlorine, as traditionally practiced, may cause some utilities to violate this MCL. On
the other hand, free chlorine which is known to maximize THM formation is the most effective
form of chlorine for disinfection. Because It Is economical and practical, chlorination is the most
commonly used method of disinfection, although alternative methods are available.
In order to meet the MCL for THM, water utilities may decide to change their disinfection
practice. It is Important to recognize, however, that the Infectious agents potentially present in a
water supply represent an immediate health hazard as opposed to the potentially chronic health
effects related to THM. Therefore, although the presence of THM Is Important, the Immediate
threat of woterborne disease transmission should be given the highest priority. Any change in
disinfection practice should be mode without compromising the microbiological quality of the
finished water.
THM production may be reduced by removing precursor material before applying free
chlorine or by using a disinfectant other than free chlorine. If the application of chlorine to a
water supply Is moved to a point later in the treatment system, the microbial load on the
coagulation-flocculatiOn, sedimentation, and filtration processes will Increase. In this situation, It
is Important to achieve a low turbidity In the filtered water, i.e., less than I NTU, because
chlorination after filtration would be relied upon to accomplish a greater degree of microbial
inactivation. The available options for reducing THM formation and still maintaining microbial
quality have been discussed by Symons, (1981)
-------
The production of potentially hazardous by-products in a water supply is not limited only to
chlorine but may also occur with ozone, chlorine dioxide and other disinfectants. For example,
concern has been expressed about the production of chlorite when chlorine dioxide is used.
Information about by-product production with ozone is lacking.
In summary, it is felt that disinfection, because of its value in controlling waterborne disease
transmission, is the most important and necessary process in the treatment of a water supply. As
such, the process must be designed and operated to maximize microbial inactivation while cit the
same time to minimize the formation of potentially hazardous by-products.
Operation of Water Treatment Facilities
The information presented above on the removal or inactivation of microorganisms by the
various water treatment processes reflects that achievable in a well-operated treatment facility.
Obviously, treatment plants that are not operated well, or which do not have the necessary
treatment processes, will not produce a microbiologicatly acceptable finished water. All elements
necessary for the satisfactory operation of a water treatment facility must be present if a
• microbiologically safe water supply is to be assured continuously. Operators must be well trained,
motivated and dedicated to meet their responsibilities. Their salaries must be commensurate with
the responsibilities associated with their position. Adequate inspection and monitoring of plant
operation by State authorities must be practiced so as to assure compliance with operoflonal
standards.
Sixnmcry ond Recommendations
A variety of unit processes and operations for the treatment of water, SO as to render it
acceptable for public use and consumption, have evolved over the years. Those unit processes and
operations most commonly used Include: raw water, storage, aeration, prechlorlnation, coagulation-
flocculation and sedimentation, hardness reduction, filtration, activated carbon treatment, and
disinfection. Of these, only disinfection is practiced specifically to meet microbial quality
objectives, i.e., absence of any detectable pathogenic microorganisms from water in addition to
achieving their pdmary objectives or render the water more susceptible to effective disinfection.
A water treatment system, consisting of an appropriate series of unit processes and operations,
-------
47
must be deliberately planned and properly designed, constructed and operated if it is to meet its
objectives, including the removal and/or inactivation of microorganisms. Based upon available
information, today’s technology is capable of providing the degree of treatment required to meet
existing microbial standards.
In general, the unit processes and operations included in water treatment systems to achieve
the desired quality of the finished water depend upon the nature and quality of the row source
water. Surface waters, as a rule, are more susceptible to microbial contamination than
groundwaters and as a result, require more extensive treatment. With few exceptions, e.g.,
groundwaters on a site-specific basis, all potable water supplies should be disinfectecj. Further, a
surface water supply normally requires a water treatment system which would include coagulation-
flocculation and sedimentation, filtration, and disinfection.
The most widely used water disinfectant today is chlorine, primarily because of its
effectiveness, ease of handling, low cost, and long experience with its use. It is now recognized,
however, that the use of free chlorine as a disinfectant can lead to the formation of trihalo-
methane (THM) with waters containing certain forms of organic matter. THM production may be
eliminated or at least reduced, without compromising the microbial quality of the finished water,
by altering the point of chlorine addition, by removing precursor organic matter before applying
chlorine, or by using a disinfectant other than free chlorine. The point of applying chlorine might
be altered by eliminating prechlorination and adding chlorine only after coagulation-flocculation,
sedimentation and filtration. The removal of THM precursors prior to the addition of chlorine may
be achieved through activated carbon treatment. Of the other possible water disinfectants,
chloramine, chlorine dioxide, ozone, Iodine, and ultraviolet light, only the first three have received
major attention as alternatives to free chlorine so as to avoid the formation of THM.
Based upon the effectiveness of the various unit processes and operations available for
treating water so as to remove and/or Inactivate microorganisms and thereby meet the microbial
quality standards for potable water supplies, It Is recommended that:
I. All water supplies be dislnfected unless if can be shown that microbial contamination Is
not a problem. In the absence of effective disinfection, it is difficult to assure the
microbial safety of a potable water supply. It Is normally assumed that disinfection will
be practiced unless It can be demonstrated that It Is unnecessary for a local, site-
-------
48
specific situation. With certain groundwater supplies, It may be possible to prove that
disinfection is not needed. In such cases, a documented sanitary survey should be
prepared for review by the responsible State authority. The survey should address, as a
minimum, the following: a) assurance of the integrity of the aquifer, b) acceptable
microbial quality of the water at the wellhead and throughout the distribution system,
and c) absence of any waterborne infection and/or disease in the population served.
2. All water supplies using a surface water source be pretreated and filtered unless It can
be shown that this degree of treatment Is not required. Surface waters are sUbject to
microbial contamination by humans and other animals and as a result, should receive
treatment in addition to disinfection. In certain cases, It may be possible to
demonstrate that, based upon raw water quality data, morbidity data covering water-
borne infections/diseases, and a documented sanitary survey, treatment of a surface
water supply beyond disinfection Is not warranted. Attention should be given, however,
to the opportunity for animals to shed microorganisms which are pathogenic to man and
which may gain access to the wateri the presence of pathogenic microorganisms from
this source may not be accompanied by elevated Indicator bacteria levels or by
Increased turbidity in the raw water.
3. All water supplies using a surface water source be monitored continuously for
disinfectant residual at a representative entry point Into the distribution systems for
small systems, compliance with this regulation may not be justified for economical
reasons. The presence of an appropriate disinfectant residual, following an adequate
contact time, would indicate effective disinfection of the water. This continuous
measurement of residual disinfectant should be part of a control system, providing
feedback information to maintain a present residual level. As part of the control
system, consideration should be given to bypassing water which has been Inadequately
disinfected so that It Is diverted bock through the disinfection process before entry Into
the distribution system.
4. The microbial quality of the finished water not be compromised In meeting the MCI.
standard establIshed for THM. If the microbial quality of the water served to the
consumer is adversely affected by control measures taken to meet the MCI.. for THM,
-------
49
the microbial standard should take precedence over the MCL established for THM. The
presence of pathogenic microorganisms in a water intended for consumption represents
a potential acute health risk while the health effect associated with THM is chronic,
depending upon a long-term exposure to THM. it is not suggested that compliance with
the established MCL and THM be abandoned. Rather, It is believed that priority should
be given to meeting the microbial quality requirements on an interim basis while THM
control measures are being put in place. in the long term, both the MCL and THM and
the microbial quality standards should be met simultaneously as compatible control
technology becomes available.
Reserch Needs
The significance of the various mechanisms of natural die-off (or growth) of micro-
organisms in raw water storage facilities should be evaluated. Some water treatment
facilities consist solely of raw water storage and disinfection. Limited information is
available upon which to predict the Influence of storage on the microbial quality of
water.
2. Conditions associated with certain unit processes and operations employed in trading
water supplies which promote the colonization and growth of pathogenic micro-
organisms should be identified and investigated. For example, it is unknown if it is
possible for Legionella p umophilla to grow In oerat on systems or if activated carbon
columns can provide a suitable habitat for the growth of opportunistic microbial
pathogens.
3. The effects of dissolved Inorganic substances and various types of particulate matter,
often found in water, on disinfection kinetics and efficiency should be established.
Recent studies have indicated that the presence of certain Inorganic salts in water can
enhance the rate of pollovirus Inactivation by free available chlorine, The Influence of
the dissolved Inorganic chemical constituents of a water on the disinfection process
needs to be fully identified. Other studies have Indicated increased resistance of
microorganisms to disinfection when associated with certain types of particulate
matter. Additional research on the type of particles in water that can confer protection
from disinfection and the nature of the protective effect may provide a more
-------
50
effective measure of the amenability of a water to disinfection than the currently used
turbidity measurement,
L , Evaluate the potential for microorganisms developing an inherited resistance to a
particular disinfectant. Recent reports have indicated that repeated exposure of
certain viruses and bacteria to chlorine can alter the resistance of their progeny to
chlorine. Because of the potential impact of this possibility, studies are needed for
corroboration and, if positive, to provide additional information.
5. Establish the effectiveness of unit processes and operations used In treating water
supplies in removing/inactivating specific pathogenic microorganisms. Information of
this type is needed for Glordlo lomblia, Entomoebo histolytica , hepatitis A and the
gostroenteri tis-causi ng viruses, Le lonel I a pneumophi I ia, Campylobocter Ie]unl I and
Yersinia enterocolitica . These specific pathogenic microorganisms either have been
only recently Implicated In waterborne disease outbreaks or have been so recognized for
some time but, because of the lack of appropriate culturing methodology, have not been
studied. Consequently, little Information Is available on their susceptlbHlty to removal
or Inactivation by water treatment systems.
6. Evaluate the potential water quality problems associated with the use of free chlorine,
other than the formation of Ti-IM, and the alternative dislnfectonts such as chloramines,
chlorine dioxide and ozone. Disinfectants, other than free chlorine, are being used by
several water utilities and ore being considered by others due to the known formation of
TI-IM with free chlorine. It Is unknown, however, whether the use of these alternative
water disinfectants can lead to the formation of undesirable by-product formation.
Further, It has been suggested that free chlorine may produce other undesirable
substances besides THM.
FINISHED WATER STORAGE AND DISTRIBUTION
A water distribution system, In addition to consisting of a network of interconnecting mains
or pipes, normally Includes storage facilities, valves, fire hydrants, service connections to users,
and pumping facilities. For the purpose of this paper, the distribution system Is considered to
Include cU facilities and hardware between the last treatment unit up to and Including the
corporation stopcock. Hardware or appurtenances associated with the consumers’ plumbing are not
-------
51
considered as part of this definition. A distribution system may be identified as either gravity flow
or pressure (pumping) systems. Very often, however, public water supply systems are a
combination of these two types. With gravity systems, water is impounded at strategic locations
sufficiently elevated to permit the water to flow into the distribution system by gravity. When
elevated impoundment or storage is impractical, the required working pressure is provided by
pumps within the system, i.e., pressure system. These pumps are nornially located at the
treatment plant and perhaps within the distribution system. These pumps are normally located at
the treatment plant and perhaps within the distribution system. In combined systems, facilities for
water storage are often provided along with provision for pumping. Typically, water Is pumped
directly into the distribution system with that quantity of water In excess of the demand going
automatically to a storage facility or reservoir. A system may also be designed so that pumps
supply the water storage facillty(Ies) directly; the water, in turn, might flow into the distribution
system by gravity.
Reservoirs, following water treatment or within the distribution system, may be used to
provide service storage. Distribution reservoirs may be classified as underground, ground level,
elevated, or standpipe. An underground reservoir or basin, either open or covered, may be at or below
grade level and formed by excavation or embankment, It is customary to line such reservoirs with
concrete, gunite, asphalt or an asphalt membrane, or butyl rubber. A stondpipe consists of a
cylindrical shell, having a flat bottom and resting on a foundation at ground level, while an elevated
tank is a reservoir supported on a tower. Steel and wood have been used In the construction of
standplpes and elevated tanks which are normally covered.
In terms of presenting barriers to disease transmission, the storage and distribution of a
finished water perhaps constitute a weak link In the entire system. Pathogenic microorganisms
may enter a distribution system through Improperly built storage reservoirs, Inadequate repair of
pipe systems, infiltration into pipe systems, or bock slphoriage and cross connections. Further, the
general microbial quality of a water may deteriorate because of the growth and regrowth of
organisms In storage reservoirs, e.g. tanks constructed out of redwood, In the distributjon syst.m
and through the use of home attachment devices such as activated carbon filters attached to the
consumer’s taps.
-------
52
Disinfectant Reslck,oI in Distribution Systems
Maintaining a disinfectant residual in the distribution system is desirable. A residual in the
water following the disinfection process is an indication that the water treatment system is
operating effectively. While the presence of a residual at this point is not an assurance that the
water is free of any detectable pathogenic microorganisms, it does indicate that a high probability
exists that the water is pathogen free. Minor plant upsets can occur and be undetected for a short
duration. During this time, Improperly disinfected water might be distributed. Disinfectants which
produce a residual In the distribution system reduce the likelihood of pathogen survival. Dis-
infected water usually contains microorganisms considered to be harmless when these organisms
are not numerous. In the absence of a disinfectant residual, these organisms can colonize and grow
within the system. A residual provides some degree of protection against microbial colonization of
the distribution system. Contamination of water in the distribution system through back siphonage,
pipe breakage or cross-connections, is always a potential hazard. The presence of a residual at the
consumer’s tap is an Indication that major Influxes of contamination have not occurred. In cases
where low level contamination has occurred, the residual may adequately control pathogens. While
It is doubtful that the level of residual normally found In distribution system is adequate to control
pathogens in the presence of massive Influxes of contamination, the residual also provides some
degree of control over the colonization of the distribution system by microorganisms entering It
through sources of contamination.
Open and Closed Reservoirs
Of the two classes of storage reservoirs, open and closed, closed reservoirs are preferred. It
is generally accepted that water in open reservoirs is subject to contamination from dust, rainfall,
mammals, birds, insects, algal growth, and violation of reservoir security (Bailey and Llppy, 1978).
It may be necessary with open reservoirs to control algal and bacterial slime growths through the
addition of copper sulfate or chlorine. Finally, in order to maintain a chlorine residual throughout
the distribution system, it may be necessary to rechlorinate waters in an open reservoir prior to
distribution. Waters In closed reservoirs are better protected from windborne contamination and
rainfall than are open reservoirs. However, proper screening of vent areas is required to protect
closed reservoirs from Intrusion by mammals and birds.
-------
53
It has been recommended that, “All new finished water storage structures shall have suitable
watertight roofs or covers which exclude birds, animals, insects, and excessive dust” (Great Lakes -
Upper Mississippi River Board of State Engineers, 1972). Further, the U.S. EPA has proposed that
new reservoirs should be covered, existing open reservoirs should be covered where economically
feasible, and where this is not economically feasible, adequate precautions should be implemented
to protect water quality in open reservoirs. These precautions include the installation of
rechlorination facilities for terminal disinfection, water quality monitoring for biological and
chemical contamination, security and control measures against human, water fowl, and other
animal intrusions, control of weeds and algae, and a by-pass system so that contaminated water
will not enter the other parts of the distribution system (Bailey and Llppy, 1978).
Distribution System Piping ond Appurtanonces
The installation of new mains, the repair of existing mains, and the physical failure of old
mains can result in a direct influx of pathogenic and other microorganisms Into a water distribution
system. For example, old piping with its often severe corrosion con Increase the turbidity of water
in a distribution system and In turn, have a major impact on the resulting microbial quality; this
problem deserves attention through research.
An existing standard (American Water Works AssociatIon, 1968) for the disinfection of water
mains identifies six areas of concern with Installing new pipes protection of new pipe sections at
the construction’ site, restriction on the type of joint packing, flushing of the pipe sections,
disinfection, final flushing and bacteriological testing. Buelow, j flj., (1916) reviewed the
literature on the disinfection of new water mains and presented their own results. They concluded
that In pipes free of extraneous debris, free available chlorine, potassium permonganate, or copper
sulfate could oil be used at low dose rates to meót collform requirements. However, only freó
available chlorine was able to eliminate large numbers of standard plate count organisms, The
major reason far new and repaired pipe falling to meet the coliform standard appeared to be
because of inadequate cleaning or flushing of pipe sections. Collform organisms residing in packing
materiCi and pip. Joints have also been found to be a major factor in th. failure of new pipe
systems to meet bacterIological standards (Hutchinson, I 97k).
Old water supply distribution systems are In part composed of cost Iran pipes that may have
been installed 80-130 years ago. In these portions of the piping system, chemically and biologically
-------
54
mediated corrosion have taken their toll causing line breaks, water discoloration, encrustation that
limits water flow, turbidity increases and sites for microbial colonization. All of these
deteriorating conditions can impact water quality and acceptability.
Microbial Regrowth and Transient Microorganisms
In order for a microorganism to colonize a water distribution system, it must: I) enter the
system, 2) become associated with a suitable habitat, 3) be able to directly or indirectly obtain
nutrients from the water, and 4) reproduce itself. Transient organisms In a water distribution
system are those that pass through the system in a viable condition as opposed to organisms which
arise from growth within the system. Entrance to a system may be gained as a result of survival
through the various treatment processes; influx by infiltration; cross-connections or back siphon-
age; or being associated with materials in new piping in the system. Not all areas in a system
provide a suitable habitat for growth. In smooth pipes, there is little opportunity for organisms to
attach to a suitable surface, primarily because of the shear forces of the water flow (Van der Kooli
and Zoefeman, 1978). However, prime sites for colonization can be found in fubercle development
and scale formation areas, slow-flow areas, dead ends, and pipe connections (Geldreich, 1980).
Settled particulate matter can suitably protect microorganisms from the effects of a chlorine
residual If the particulate matter is of organic rather than inorganic nature (Hoff, 1978 and Trocey,
., 1966). Direct sources of nutrients in the distribution system are available from unoxidized
soluble and particulate organic matter that was not removed in treatment and In the form of
certain oxidizable minerals such as Iron. indirect sources of nutrients are available in the form of
excreta and biomass products of other microorganisms Inhabiting the system (Postgate and Hunter,
1962). Electron mlcrographs of organisms inhabiting tubercies in pipe have shown them to be
associated with the upper surface of the tubercie material (Allen, et ol., 1980).
The microorganisms most commonly found to Inhabit distribution systems are represented by
the coliforms, Klebslella and Enterobacter , and organisms of the genera Pseudomonas, Flavo-
bacterium, Acinetobacter, Actinomycetes , and Arthrobacter (Geldreich, et 01., 1977 and Ptak, et
al., 1973). Some members of these genera have been Identified as opportunistic human pathogens
while others have been identified as coliform antagonists. Still others are commonly associated
with taste and odor and turbidity problems.
-------
55
It appears that while microorganisms able to inhabit distribution systems are protected
against a chlorine residual present in the water, transient organisms in the system are readily
inactivated (Snead, et al., 1980). A review of the effects of turbidity on chlorine disinfection would
indicate that the association of transient microbes with certain types of turbidity may enhance
their survival.
Bock Siphonage ond Cross-Connection
The conditions leading to back siphonage and cross-connections In a distribution system are
particularly significant in that large quantities of contaminated waters may be directly injected
into a wafer supply. Back siphonage occurs when a service line is in contact with a contaminated
water and a drop in pressure in the supply system allows contaminated water to be drawn into the
water supply. Cross-connections occur when a piped water supply becomes connected to a pipe
carrying any material other than water from the supply. While low levels of microbial
contamination may be controlled by having a chlorine residual in the water supply (Snead, et cii.,
1980), it is doubtful that there is adequate protection against a major Influx of contaminated water
(Geldreich, 1980).
Operation of Distrlbuflcn Systems
The distribution system is considered to be the final and most fragile barrier to disease
transmission in providing consumers with a microbiologically safe water. In fact, a major fraction
of the reported cases of waterborne disease outbreaks have been attributed to failures In the
integrity of the distribution sytem (Craun, 1979 and McCabe, j flj., 1970). it Is, therefore, of
great importance that a program of constant vigilance be established and enforced over the
operation, maintenance and repair of the distribution system. Such a program would Include
surveillance for signs of breakage, cross-connection, and general deterioration through natural
corrosion and biological inhabitation of the distribution system. The program should Include a
schedule for monitoring of pertinent indicators of water quality, such as pH, dIsinfectant residual,
and turbidity; a schedule of routine maintenance, such as main flushing; and a set of standards for
pipe repair.
-------
56
Siinm y ond Recommendations
After treatment, a water supply is distributed to consumers through a series of mains, pipes,
pumps and storage tanks which might range from a single source water and a simple interconnec-
tion of laterals to a complex arrangement which might involve different pressure zones and
multiple sources. The distribution system provides may opportunities for microbial degradation and
contamination of the water through such incidents as main breaks, back siphonoge, regrowth of
microorganisms within the system, introduction of microorganisms through open reservoirs, etc.
The statistics covering waterborne disease outbreaks indicate that 34 percent of the Incidents
are caused by deficiencies within the distribution system. This statistic Is hardly startling when it
is considered that the vast majority of a water supply system (in terms of bulk, equipment, and
investment) is associated with the distribution of the water and that only a minor amount of
surveillance, relative to the treatment system, is possible within the distribution network. The
sheer vastness and logistics of the distribution system prevent detailed, comprehensive monitoring
of its operation. in the past, the major concern with this element of the water supply system was
related to physical characteristics such as pressure, tank levels, flow direction, line breaks, etc.
Concern has always been expressed about the microbial quality of the water In the distribution
system but regulations have avoided any type of specific requirements. Regulatory action is
necessary in several areas (and to varying degrees) to stimulate action and surveillance at the State
and/or local levels. The dlstrlbu’tian system can no longer be Ignored with respect to providing
protection of the water that Is delivered to the consumers.
To minimize the Impact of distribution systems on the microbial quality of water supplies, It
is recommended thats
I. Th. water in all portions of a distr1but on system contain a disinfectant rósidual. Water
supply systems serving more than 10,000 people should be monitored’ for disinfectant
residual at th. same frequency as that required for bacteriological sampling. Insuf-
ficient evIdence exists to specify disinfecting species (free available chlorine,
chioramine, etc.) or the level of residual desired (detectable, 0.2 mg/i, ótc.). The
10,000 population was selected to coincide with the 1982 THM requirement.
2. All new finished water reservoirs be constructed with a cover. Similarly, the placement
of covers on existing open reservoirs should be encouraged. The hazards of microbial
contamination dictate that all distribution reservoirs be covered.
-------
57
3. All new mains, pipes and appurtenances be disinfected before being placed in service,
utilizing the AWWA standard or an equivalent procedure. Disinfection of repaired
mains, pipes, etc. should also follow AWWA disinfection procedures where applicable.
it is an accepted fact that microbial contamination of the water in the distribution
system can occur as a result of inadequate precautions taken during the installation and
repair of mains, pipes, etc.
£ . An active and effective cross-connection control program be implemented for every
public water supply, taking into account the specific character of the distribution
system. The existence of a cross-connection has been Implicated in numerous
waterborne disease outbreaks. Most agencies have a cross-connection regulation in
place, but as a general rule, it is not effectively applied or enforced.
Research Needs
Define conditions which permit microorganisms to colonize distribution systems. Water
treatment plants may suffer undetected minor upsets that inoculate the distribution
system with microorganisms, including pothogens. Additionally, microorganisms may
proliferate Into massive growths which in turn, cause degradation of the quality of the
water. Knowing the conditions which faster colonization of microorganisms in
distribution systems, It should be possible to identify routine and emergency pre-
ventative measures.
2. Establish the relationship which exists among turbidity, physical and chemical
characteristics of suspended material and the occurrence of microorganisms and their
density. The turbidity of a water is an indirect measure of insoluble contaminants
originating from different sources. On the other hand, turbidity measurements do not
reflect the physical or chemical characteristics of these contaminants. It is known that
such suspended matedal can reduce disinfectant efficacy, with InorganIc materials
more likely having a lesser effect than organic matter. Microbial growths are
undersirable In distribution systems because they cause increased disinfectant demand,
provide potential corrosion sites where damage to pipes can occur, and reduce the
aesthetic qualIty of the water.. A more exacting measure of particle size, number and
chemical nature may lead to an understanding of how microbial colonization occurs.
-------
58
LITERATURE CITED
Allen, M. J., and E. E. Geldreich. 1975. BacterIological criteria for groundwater quality. Ground
Water 3:45-52.
Allen, M. J., R. H. Taylor, and E. E. Getdreich. 1980. The occurrence of microorganisms in water
main encrustations. J. Am. Water Works Assoc. 72:614-625.
American Water Works Association. 1925. Water works practice. Wilkins and Wilkins Co.,
Baltimore MD.
American Water Works Assocation. 966. Standard for deep wells. AWWA Standard A 100-66.
American Water Works Association, Denver, CO.
American Water Works Assocation. 1968. Standard for disinfecting water mains. AWWA Standard
C60 1 -68, American Water Works Association, Denver, CO.
Amirhor, P., and R. S. Engelbrecht. 1975. Virus removal by polyelectrolyte aided filtration. J.
Am. Water Works Assoc. 67:187-192.
Amirhor, P., and R. S. Engeibrecht. 1975. interaction of virus and polyelectrolyte in wostewater
effluent. Prog. In Water Tech. 7:677-686.
Andre, D. E., H. H. Welser, and G. W. Maloney. 1967. Survival of bacterial enteric pathogens In
farm pond water. J. Am. Water Works Assoc. 2:503-508.
Arozorena, M. M. 1977. Removal of Giardia muris cysts by granular media filtration. MS. thesis,
University of Cincinnati, Cincinnati, OF-I.
BoDy, S. W., and E. C. Lippy. 1978. Should all finished water reservoirs be covered? Public Works
Mag,, April 1978, pp. 66-70.
Bates, R. C., P. T. B. Shoffer, and S. M. Sutterlond. 1977. Development of pollovirus having
increased resistance to chlorine Inactivation. AppI. Environ. Microblol. 849-853.
Baylis, J. R, 0. Gullans, and B. c<. Spector. 1936. The efficiency of rapid sand filters In removing
the cysts of amoebic dysentery organisms from water. U.S. Public Health Reports 51:1567.
Benarde, M. A., B. M. israel, V. P. Oiivieri, and M. L. Granstorm. 1965. Efficiency of chlorine
dioxide as a bacteriocide. Appi. Microbiol. 13:776-780.
Benarde, M. A., W. B. Snow, and V. P. Ollvleri. 1967. ChlorIne dioxide disinfection temperature
effects. J. AppI. Bacterlol. Q:l59 -I69.
Berg, G. 1966. VIrus transmission by the water vehicle, Ill. Removal of viruses by water
treatment procedures. Health Lab. Sc :l70-l8l.
Berg, C. 1977. MIcrobiology - detection, occurrence, and removal of viruses. J. Water Poll.
Control Fed. 49:1290-1298.
Berg, G., R. E. Dean, and D. R. Dahiing. 1968. Removal of pollovirus I from secondary effluent by
lime flocculation and rapid sand filtration. J. Am. Water Works Assoc. Q :l93-l98.
Black, A. P., R. N. Kinman, W. C. Thomas, Jr., G. Freund, and E. D. Bird. 1965. Use of iodine for
disinfection. J. Am. Water Works Assoc. 57:1401-1421.
Block, J. C. 1977. Contribution a I’etude de Ia desinfectlon des easux d’ alimentation par l’ozone.
Ph.D. Thesis, d’Etat es-Sciences Pharmacentiques, Naney, France.
-------
59
Boardman, G. D, and 0. J. Sproul. 1977. Protection of viruses during disinfection by adsorption to
particulate matter. J. Water Poll. Control Fed. 49:1857-1861.
Boring, J. R., Ill, W. 1., Martin, and L. M. EllIot. 1971. Isolation of Salmonella typhlmurlum from
municipal water, Riverside, CA. Am. J. Epidemiol. 93:49-54.
Boyce, D. S., 0. J. Sproul, and C. E. Buck. 1981. The effect of bentonite clay on ozone
disinfection of bacteria and viruses in water. Water Res. 15:759-765.
Braza, A. 1964. On the survival of Salmonella In the sludges of a fertilizing irrigation plant. I.
Giene Mod. (ltd.) 57:635-647.
Buelow, R. W., R. H. Taylor, E. E. Geldreich, A. Goodenkauf, L. Wllwerdlng, F. Holdren, M.
Hutchinson, and I. H. Nelson. 1976. Disinfection of new water mains. J. Am. Water Works
Assoc. 68:283.
Burman, N. P. 1962. Bacteriological control of slow sand filtration. Effl. and Water Treat. J.
2:674-677.
Burman, N. P., J. K. Stevens, and A. W. Evans. 1978. Streptococci as indicators In water supplies,
p. 335-347. In: Streptococci, Soc. AppI. Bact. Symp. Ser. No. 7, F. A. Skinner and L. B.
Quesnel (eds.rAcademlc Press, London.
Butterfield, C. T. 1948. Bactericidal properties of chloramines and free chlorine In water. Public
Health Reports 63:9340-9349.
Butterfield, C. 1., E. Wattle, S. Mergreglan, and C. W. Chambers. 1943. Influence of pI-I and
temperature on the survival of coliforms and enteric pathogens when exposed to free
chlorine, Public Health Reports 58:1837-1866.
Carlson, J. H., G. M. Ridenour, and C. F. McKhann. 1942. Efficacy of standard purification
methods in removing poliomyelitls virus from water. Am. J. Public Health 32:1256.
Cerkinsky, S. N., and N. Trahtman. 1972. The present status of research on the disinfection of
drinking water in the USSR. Bull. World Health Organ. 46:277-283.
Chambers, C. W., P. W. Kabler, G. Maloney, and A. Bryant. 1952. IodIne as a bactericide. Soap
and Sanit. Chem. 28 : 1 49-I 6S.
Chang, S. L. 1958. The use of active Iodine as a water disinfectant. J. Am. Pharm. Assoc. :417-
423.
Chang, S. L. 1971. Modern concept of disinfection. J. San. Eng. Dlv., Am. Soc. Civil Engr.,
SA S:689-707.
Chang, S. L. 1978. ResIstance of pathogenic Noeglerla to some common physical and chemical
agents. AppI. Environ. Microblol. 35:368-375.
Chang, S. L., C. Berg., N. A. Clarke, and P. W. Kabler. 1960. SurvIval and proteátlon against
chlorination of human enteric pothogens In free-living nematode isolates from water supplies.
Am. J. Trop. Med. Hyg. 2:136-142.
Chang, S. L., and J. C. Morrjs. 1953. Elemental iodine as a disinfectant for drinking water. lr,d.
and Eng. Chem. 45:1009-1012.
Chang, S. L., R. E. Stevenson, A. R. Bryant, R. L. Woodward, and P. W. Kabler. 1958. Removal of
coxsockle and bacterial viruses in water by flocculation, I. Removal of coxsackie and
bacterial viruses In water of known chemical content by flocculation with aluminum sulfate
or ferric chloride under various testing conditions. Am. J. Public Health 48:51-61.
-------
60
Chang, S. L., R. E. Stevenson, A. R. Bryant, R. L. Woodword, and P. W. Kabler. 1958. Removal of
coxsackie and bacterial viruses in water by flocculation, H. Removal of coxsackie and
bacterial viruses and the native bacteria in raw Ohio River water by flocculation with
aluminum sulfate and ferric chloride. Am. J. Public Health 48:159-169.
Clarke, N. A., G. Berg, P. W. Kabler, and S. L. Chang. 1964. Human enteric viruses in water:
source, survival and removability. Adv. Water Poll. Res. 2:523_Sill.
Clarke, N. A., R. E. Stevenson, and P. W. Kabler. 1956. The inactivation of purified type 3
adenovirus in water by chlorine. Am. J. Hyg. 64:314-319.
Cliver, D. 0. 197. Viruses in water and wastewater - effects of some treatment methods,
p. 149-157. In: Virus and water quality: occurrence and control, Proc. Thirteenth Water
Quality Conf.7V. Snoeyink and V. Griffin (eds.), University of Illinois, Urbana, IL.
Cohen, B. 1922. The effects of temperature and hydrogen ion concentration upon the viability of
Bact . coil and Bact. typhosa in water. J. Bact. 7:183-230.
Commission of the European Communities. 1975. Council directive of May 16, 1975, concerning
the quality of surface waters Intended for the abstraction of drinking water in the member
states. Official J. European Communities 194:26-31.
Commission of the European Communities. 1978. Proposal to a council directive concerning
measuring methods and frequency for sampling and analysis of parameters as a quality of
source water Intended for abstraction of drinking water. Official J. European Communities
208:1-li.
Coolicson, J. T., Jr. 1969. Mechanism of virus adsorption on activated carbon. J. Am. Water Works
Assoc. 61:52-59.
Cookson, J. T., Jr., and W. J. North. 1967. Adsorption of viruses on activated carbon-equilibria
and kinetics of the attachment of Escherlchla coil bacteriophage T4 on activated carbon.
Environ. Sd & Technol. 1:46-50.
Cram, E. 8. (943. The effect of various treatment processes on the survival of he$mlnth ova and
protozoan cysts in sewage. Sew. Works J. j : I I I 9-I 138.
Cramer, W. M., K. Kawata, and C. W. Kruse. 1976. Chlorination and lodination of pollovirus and
f2. J. Water Poll. Control Fed. 48:61-78.
Craun, G. F. $979. Waterborne dIsease - a status report emphasizing outbreaks In ground water
systems. Ground Water 17:183-191.
Craun, G. F., and L. J. McCabe. 1973. RevIew of the causes of waterborne disease outbreaks. i
Am. Water Works Assoc. 65:74-84.
Cronier, S. D. 1977. Destruction by chlorine dioxide of viruses and bacteria In water. M.S. thesis,
University of Cincinnati, Cincinnati, OH.
Cronier, S., P. V. Scarpino, and M. L. Zinc.. 1978. Chlorine dioxide destruction of viruses and
bacteria in water, p. 651-657. j.!j: Water chlorination environmental Impact and health
effects, vol. 2, R. L. Jolley, H. Gorchev, and D. H. HamIlton, Jr. (eds.). Ann Arbor Sd, Ann
Arbor, MI.
Culp/Wesner/CuIp and M. V. Hughes, Jr, 1979. Water reuse and recyclIng, 1. Evaluation of needs
and potential. Office of Water Res. and Technol., U.S. Dept. of the tnt., OWRT/RU-791l.
CuIp/Wesner/Cuip. 1979. Water reuse and recycling, II. Evaluation of treatment technology.
Office of Water Res. and Technol., U.S. Dept. of the mt., OWRT/RU-79/2.
-------
‘I
Cummins, B. B., and H. D. Nash. 1978. Microbial implications of alternative treatment.
Presented at 6th Annual Water Quality Technol. Conf., American Water Works Assac.,
Denver, Co.
Emerson, M. A., 0. J. Sprout, and C. E. Buck. 1982. Ozone inactivation of cell-associated viruses.
In press, Appi. Environ. Microblol.
Engelbrecht, R. S. 1976. Removal and inactivation of enteric viruses by wastewater and water
treatment processes. j : Advanced wastewater treatment, Japan Research Group of Water
Pollution Seminar, Tokyo, Japan.
Engelbrecht, R. S., C. N. Hoos, J. A. Shulor, 0. L. Dunn, D. Roy, A. Laichandani, B. F. Severin,
and S. Farooq. 1979. Acid-4ost bacteria and yeasts as Indicators of disinfection efficiency.
MERL, Environmental Protection Agency, Cincinnati, OH, EPA 600/2-79-091.
Engetbrecht, R. S., M. J. Weber, C. A. Schmidt, and B. L. Salter. 1978. VIrus sensitivity to
chlorine disinfection of water supplies. MERL, Environmental Protection Agency, Cincinnati,
OH, EPA 600/2-78-123.
Englebrecht, R. S., M. J. Weber, B. L. Salter, and C. A. Schmidt. 1980. Comparative inactivation
of viruses by chlorine. Appl. Environ. Microbiol. 40:249-256.
Enslow, L. H. 1928. Progress in chlorination of water, 1927- 1928. J. Am. Water Works Assoc.
20 :8 19-846.
Environmental Protection Agency. 1971. Report on Missouri River water quality studies. Office
of Water Quality, Region VII, Kansas City, MO,
Environmental Protection Agency. 1973. Use of Iodine for disinfection of drinking water. Policy
letter from J. H. McDermott, Director, Water Supply Division, to Directors of Regional Air
and Water Programs DivisIons, February 20, 1973.
Environmental Protection Agency. 1980. Protocol developments criteria and standards for
potable reuse and feasible alternatives. In: Microbiological criteria and standards for
potable reuse of water. Office of Drinking ?Thter, Environmental Protection Agency.
Esposito, P. V. Scorpino, S. L. Chang, and G. Berg. 1974. Destruction by dichioramine of viruses
and bacteria In water. Abstract No. G99. Abstracts of Annual Meeting, Am. Soc. for
Microbiol., Washington, D.C.
Evison, L. M. 1977. Disinfection of water with ozone; comparative studies with enteroviruses,
phczge, and bacteria. Presented at 3rd Congress of Intern. Ozone Inst., Paris, France.
Farooq, S., E. S. K. Chion, and R. S. Engeibrecht. 1977. Basic concepts In disinfection with ozone.
J. Water Poll. Control Fed. 49 l8I8-l831.
Farooq, S., R. S. Engelbrecht, and E. S. K. Chian. (977. The effect of ozone bubbles on
disinfection. Prog. Water Tech. ! 157-166.
Farooq, S., R. S. Engelbrecht, and E. S. K. Chion. 1977. Influence of temperature and UV light on
disinfection with ozone. Water Res. fl:7 3 7- 74 I.
Farooq, S., R. S. Engelbrecht, and E. S. K. Chian. 1978. Process considerations in design of ozone
contactor for disinfection. J. Environ. Eng. Div., Am. Soc. Civil Eng. 104:835-847.
Feng, 1. H. 1966. Behovk r of organic chicromines in disinfection. J. Water Poll. Control Fed.
38 *614-628.
Fennell,’H., 0. B. James, and J. Morris. 1974. Pollution of a storage reservoir by roasting gulls. J.
Soc. Water Treat. Exam. 23:5-24.
-------
62
Fluegge, R. A., T. G. Metcalf, and C. Walles. 1979. Virus inactivation in wastewater effluents by
chlorine, ozone, and ultraviolet light. In: Prog. in wastewater disinfection technology,
MERL, Environmental Protection Agency, Cincinnati, OH, EPA/9-79-0 18.
Foliguet, J. M., and F. Doncoeur. 1975. Elimination des enterovirus and cours du traitement des
eaux d’alimentation par coagulation-flocculation filtration. Water Res. 9:953- 6l.
Foster, D. M., M. A. Emerson, C. E. Buck, D. S. Walsh, and 0. J. Sproul. 1980. Ozone inactivation
of cell and fecal associated viruses and bacteria. J. Water Poll. Control Fed. 52:2174-2184.
Gaufin, A. L. 1974. A study of natural water systems in western Montana. Joint Water Resources
Research Center, Bozemari, MI, OWRT-A-0 Ii.
Geldreich, E. E. 1966. Sanitary significance of fecal coliforms in the environment. U.S. Dept. of
Interior, Water Poll. Control Res. Ser. WP 20-3.
Geldreich, E. E. 1970. Applying bacteriological parameters to recreational water quality. J. Am.
Water Works Assoc. 62: I 13-120.
Geidreich, E. E. 1972. Buffalo Lake recreational water quality. A study in bacteriological data
interpretation. Water Res. 6:913.
Geidreich, E. E. 1972. Water-borne pathogens. In: Water pollution microbiology, R. Mitchell
(ed). Wiley lnterscfence, New York.
Geidreich, E. E. 1980. Microbiological processes in water supply distribution. Presented at Am.
Soc. for Microbiol. Seminar, Miami Beach, FL, May lLe, 1980.
Geidreich, E. E. 1981. Emerging microbiological issues. Presented at Northwest Reg. Water
Supply Short Course, University of Seattle, Seattle, WA.
Geldreich, E. E., H. D. Nash, and D. Spiro. 1977. Characterizing bacterial populations in treated
water supplies: A progress report. Proc. Am. Water Works Quality Technol. Conf., Kansas
City, MO.
Geidreich, E. E., and R. H. Bordner. 1971. Fecal contamination of fruits and vegetables during
cultivation and processing for market. A review. J. Milk Food Technol. 3k(4):I8Li I85.
Gerba, C. P., C. Wallis, and J. L. Melnick. 975. Fate of wastewater bacteria and viruses in soil.
J. Irrigation and Drainage Div., Am. Soc. Civil Eng. I0l(IR3) l45-l55.
Gerba, C. P., M. D. Sobsey, C. Wallis, and J. L. Melnick. 1975. Adsorption of poliovirus onto
activated carbon in wastewater. Environ. Sci. & Technol. 9:727-732.
Gevaudan, P., G. Bossy, C. Gulian, and Y. Sanchez. 1971. Action de l’ozone sur le virus de Ia
poliomyelite present dons Peau. Terres et Eaux. 67:25-28.
Gilcreas, F. W., and S. M. Kelly. 1955. Relation of coliform-organism test to enteric-virus
pollution. J. Am. Water Works Assoc. 47:683-694.
Great Lakes-Upper Mississippi River Board of State Engineers. 1972. Recommended Standards for
Water Works. 1968 ed. with 1972 revisions. Health Education Services, Albany, NY.
Guy, M. D., J. D. Mclver, and M. J. Lewis. 1977. The removal of virus by a pilot treatment plant.
Water Res. 11:421-428.
Hoas, C. N. 1981. Sodium alteration of chlorine equilibria: Quantitative description. Environ.
Sci. & Technol. 15:1243.
I-laos, C. N., and E. C. Morrison. 1981. Repeated exposure to chlorine; Production of strains
possessing altered sensitivity. Water, Air and Soil. Poll. 16:233-234.
-------
63
Haas, C. N. and M. Zapkin. 1981. Enhancement of chlorine inactivation of E. coli by sodium ions.
In: Water chlorination: Environmental impact and health effects, vol. 4, F T. Jolley (ed.). In
press.
Hattingh, W. H. J. 1978. Health aspects, p. 259. In: Manual for water renovation and
reclamation. Council for Sci. and md. Res., Pretoria, outh Africa, Tech. Guide K42.
Hazen, A. 1913. The filtration of public water supplies. 3rd ed., John Wiley and Sons, New York.
Hejkal, 1. W,, F. M. Wellings, P. A. LaRock, and A. L. Lewis. 1979. Survival of poliovirus within
organic solids during chlorination. Appi. Environ. Microbiol. 38:114-118.
Hejkal, 1. W., F. M. Wellings, A. L. Lewis, and P. A. LaRock. 1981. Distribution of viruses
associated with particles in wastewater. Appl. Environ. Microbiol. 41:628-634.
Hirshleifer, J., J. DeHaven, and J. Milliman. 1960. Water supply economics, technology, and
policy. University of Chicago Press, Chicago, IL.
Hoekstra, A. C. 1978. Disinfectie van Haags drinkwater en zuiverende werking van de Iangzome
zandfilter, periode 1966-1978. Report of DWL’s-Gravenhage, Drinkwaterleiding van’s-
Gravenhage, ‘s-Gravenhage, The Netherlands.
Hoff, J. C., 1978. The relationships of turbidity to disinfection of potable water. In: Evaluation of
the microbiology standards for drinking wafer, C. W. Hendricks (ed.). Environmental
Protection Agency, Washington, D.C., EPA-570/9-78-OOC.
Hoff, J. C., and E. E. Geldreich. 1981. Comparison of the biocidal efficiency of alternative
disinfectants. J. Am. Water Works Assoc. 73:40-44.
Holluta, J., and U. Unger. 1954. Die keimototung von bact. E. coli durch chlordioxyd und ozon.
Vom Wasser 21:129-142.
Houston, A. C. 1908-1911. First through seventh report on research work. Metropolitan Water
Board, Thames Water Authority, London.
Huisman, L., and W. E. Wood. 1974. Slow sand filtration. World Health Organization, Geneva.
Hunter, J. V., G. R. Bell, and C. N. Henderson. 1966. Coliform organisms removal by diatomite
filtration. J. Am. Water Works Assoc. 58:1160-1169.
Hutchinson, M. 1974. WRA medlube: An aid to mains disinfection. Proc. Soc. for Water
Treatment and Exam. 23(2): 174.
Jarroll, E. L., A. K. Bingham, and E. A. Meyer. 1981. Effect of chlorine on Giardia lomblia cyst
viability. Appl. Environ. Microbiol. 41:483-487.
Jensen, H., K. Thomas, and D. G. Sharp. 1980. Inactivation of coxsackie-virus B3 and B5 in water
by chlorine. AppI. Environ. Microbiol. 40:633-640.
Johnson, J. D., K. Aldrich, D. E. Francisco, T. Wolff, and M. Elliot. 1979. UV disinfection of
secondary effluent, p. 108. In: Progress in Wastewater Disinfection Technology, Environ-
mental Protection Agency, Cincinnati, OH EPA-600/9-79-0l8.
Kempf, J. E., M. G. Wilson, M. E. Pierce, and M. H. Soule. 1942. Effect of aluminum hydroxide
sedimentation, sand filtration and chlorination on the virus poliomyelitis. Am. J. Publ.
Health 32:1366-1372.
Keswick, H. B., and C. P. Gerba. 1980. Viruses in groundwater. Environ. Sd. & Technol. j :l 290-
1297.
-------
64
Kinmcsi, R. N. 1975. Water and wastewater disinfection with ozone: A crItical review. Critc
Rev. Environ. Cont. 5:141-152.
Kjellander, J., and E. Lund. 1965. Sensitivity of Escherichia coli and poliovirus to different forms
of combined chlorine. J. Am. Water Works Assoc. 57:893 b.
Klotz, M. 1979. Microbioiogische untersuchungen zur trinkwcisseraufbereitung mit aktivkohle.
Quantitative Erfassung und Bewertung. Diplom-Arbeit, Universi tat des Saarlandes,
Saarbrucken, FRG.
Klotz, M., P. Werner, and R. Schweisfurth. 1975. Untersuchungen zur biologie in aktivkohlefiltern,
p. 270-282. In: Heft 9:7. Vortragsreihe mit Erfahrungsaustousch uber spezielle Fragen der
Wassertechnologie-Adsorption. Engler-Bunte instifut der Universitat Karisruhe, Korisruhe,
FRG.
Klotz, M., P. Werner, and R. Schweisfurth. 1976. Mikrobiologische untersuchungen der
aktivkohlefiltration zur trinkwcsseraufbereitung, p. 73-82. In: Forschung und entwicklung in
der wosserpraxis (wissenschaftliche berichte unber untersuchungen und planungen der
stadtwerke Wiesbaden AG), Band 3, Stadtwerke Wiesbaden AG, Wiesbaden, FRG.
Klotz, M., P. Werner, and R. Schweisfurth. 1976. Investigations concerning the microbiology of
activated carbon filters, p. 312-330. In: Translation of report on special problems of water
technology. vol. 9. Adsorption. nference held in Karisruhe, FRG, 975. MERL,
Environmental Protection Agency, Cincinnati, OH, EPA-600/9-760-030.
Kool, H. J. 1977. Het gedrag van heterotrofe bocterien in relatle tot het fytoplankton In bet
sparrbekken De Grate Rug en drie model reservoirs, p. 1-16. NatI. Inst. for Water Supply,
Leidschendam, The Netherlands, RID CBH 78/09.
Kool, H. J. 1978. Treatment processes in publIc water supply for the removal of microorganisms,
p. 17- I - 17-31. In: Biological Indicators of water quality A. James and L. E. Evlson (eds.).
John Wiley and Sons, Ltd., Chichester, UK.
Kruse, C. W., V. C. Hsu, A. C. Griffiths, and R. Stringer. 1970. Halogen action on bacteria,
viruses and protozoa, p. 113- I 36. In: Proc. Nat. Specialty Conf. on Disinfection, Am. Soc.
Civil Eng., New York.
Liu, 0. C,, H. R. Seroichekas, E. W. Akin, D. A. Brashear, E. L. Katz, and W. J. Hill, Jr. 1971.
Relative resistance of twenty human enteric viruses to free chlorine in Potomac water,
p. 171 - 195. In: Virus and water quality: Occurrence and control. Proc. 13th Water Quality
Conf., V. Snoeyink (ed.), University of Illinois at U-C, Urbana, IL, Bull. 69.
Logsdon, G. S., J. M. Symons, R. L. Hoge, Jr., and M. M. Arozarena. 1981. Alternative filtration
methods for removal of Giardia cysts and cyst models. J. Am. Water Works Assoc. Z. :I I I -
ill.
Logsdon, G. S., and K. R. Fox. 1981. Getting your money’s worth from filtration. Presented at
Pacific Northwest Section Annual Meeting, Am. Water Works Assoc.
Lothrop, T. L., and 0. J. Sproul. 1969. High level inactivation of viruses In wastewater by
chlorination. J. Water Poll. Control Fed. 41:567-575.
Mahnel, H. 977. Untersuchengen zur virusinaktivierung in trink und oberfiochenwasser. Em
beitrog zur dekontamination von trlnkbaren wassern mit feldangleichen methoden. AbI.
Bact. Hyg.l, Abt. Orig. B, 165:527-538.
Manwaring, J. F., M. Chaudhurl, and R. S. Engelbrecht. 1971. Removal of viruses by coagulation
and flocculation. J. Am. Water Works Assoc. 63:298-300.
McCabe, L. J., J. M. Symons, R. D. Lee, and G. G. Robeck. 1970. Survey of community water
supply systems. J. Am. Water Works Assoc. 62:670-687.
-------
65
McCarty, P. L., M. Reinhard, C. Dolce, H. Nguyen, and D. G. Argo. 1978. Water factory 21:
Reclaimed water, volatile organics, virus, and treatment performance. MERL, Environ-
mental Protection Agency, Cincinnati, OH, 600/2-78-076.
McCarty, P. L., M. Reinhard, J. Graydon, J. Schreiner, K. Southerland, T. Everhart, and D. G.
Argo. 1980. Advanced treatment for wastewater reclamation at water factory 2 I. Dept. of
Civil Eng., Stanford Univ., Stanford, CA, Tech. report 236.
McCormick, R. F., and P. H. King. 1980. Direct filtration of Virginia surface waters: Feasibility
and cost. Virginia Water Resources Res. Center, Polytechnic Inst. and State Univ.,
Blacksburg, VA, Bull. 129.
McFeters, G. A., G. K. Bissonnette, J. J. Jezeski, C. A. Thomas, and D. G. Stuart. 974.
Comparative survival of indicator bacteria and enteric pathogens in well water. Appl.
Microbiol. 27:823-829.
Metropolitan Water Board. 197 1-1973. Reports on research work - bacteriological chemical, and
biological examination of London water. Metropolitan Water Board, Thames Water
Authority, London 45:147.
Miller, G. W., R. G. Rice, C. M. Robson, R. C. Scullin, W. Kuhn, and H. Wolf. 1978. An
assessment of ozone and chlorine dioxide technologies for treatment of municipal water
supplies. MERL, Environmental Protection Agency, Cincinnati, OH, EPA-600/2-78- 147.
Moffa, P. E., E. C. Tifft, S. L. Richardson, and J. E. Smith. 1975. Bench scale high rate
disinfection of combined sewer overflows with chlorine and chlorine dioxide. U. S.
Environmental Protection Agency Final Report, EPA-670/2-75-02 I.
Morris, E. J. 1972. The practical uses of ultraviolet radiation for disinfecting purposes. Med. Lab.
Tech. 29:41-47.
Morris, J. C. 1966. The future of chlorination. J. Am. Water Works Assoc. : 1475- 1482.
Morris, J. C., S. L. Chang., C. M. Fair, and C. I-I. Conant, Jr. 1953. Disinfection of drinking water
under field conditions. md. and Eng. Chem. 45:1012-1015.
Munger, S. F., A. A. Heyward, and R. G. Swartz. 1977. Inactivation of virus by chlorination or
ozonation ii four wastewater treatment system effluents. Abstract Paper Q47, Annual
Meeting Am. Soc. Microbiol., New Orleans, LA.
Namie, H. 1976. Ozone, an alternative to chlorine disinfection: A cost estimation. Proc. Forum
on Ozone Disinfection, Intern. Ozone Inst., Chicago, IL.
National Academy of Sciences - National Academy of Engineering. 1973. Water quality criteria
1972. Environmental Protection Agency, Washington D.C., EPA-R3-73-033-3/73.
National Academy of Sciences. 1977. Drinking water and health, vol. I. Safe Drinking Water
Comm., National Academy Press, Washington, D.C.
National Academy of Sciences. 1980. Drinking water and health, vol. 2. Safe Drinking Water
Comm., National Academy Press, Washington D.C.
Newton, W. L., and M. F. Jones. 1949. Effect of ozone in water on cysts of Endamoeba
histolytica . Am. J. Trop. Medicine 29:669-681.
Ohio River Valley Water Sanitation CommIssion. 980. Water treatment process modification for
trihalomethane control and organic substances in the Ohio River. Ohio River Valley Water
Sanitation Comm., Cincinnati, OH, EPA-600/2-8-028.
Olivieri, V. P., C. W. Kruse, and K. Kawata. 977. Microorganisms in stormwater. MERL,
Environmental Protection Agency, Cincinnati, OH, EPA 600/2-77-087.
-------
66
Parkhursf, J. D. 1977. Pomona virus study - final report. California SWRCB and U.S.
Environmental Protection Agency.
Pillai, S. C., M. I. Gurbaxani, and K. P. Menon. 1952. Influence of activated sludge on certain
pathogenic bacteria. Indian Med. Gaz. 87:117-I 19.
Postgate, J. R., and J. R. Hunter. 1962. The survival of starved bacteria. J. Gen. Microbiol.
29:233.
Poynter, S. F. B., and J. K. Stevens. 1975. The effects of storage on the bacteria of hygienic
significance. In: Proc. of Symposium, The Effects of Storage on Water Quality, Water
Research Centre, Medmenham Laboratory, Medmenham, UK.
Poynter, S. F. B., and J. S. Slade. 1977. The removal of viruses by slow sand filtration. Prog.
Water Technol. 9:75-78.
Price, A. 1., and L. R. G. Valadon. 1970. A bacteriological study of a new Thames Valley
reservoir, Microbios 22:269-271.
Ptak, D. J., W. Ginsburg, and B. F. Willey. 1973. Identification and incidence of Klebsiella in
chlorinated water supplies. J. Am. Water Works Assoc. 65:604-607.
Rice, E. W., and J. C. Hoff. 1981. Inactivation of Giardia lamblia cysts by ultraviolet irradiation.
AppI. Environ. Microbiol. 52:546-547.
Rice, E. W., J. C. Hoff, and F. W. Schaefer. 981. Inactivation of Giardia cysts by chlorine. AppI.
Environ. Microbiol. 43:250-251.
Robeck, G. G., K. A. Dostal, and W. L. Woodward. 1964. Studies of modifications in water
filtration, J. Am. Wafer Works Assoc. 56:1981-1990.
Robeck, G. G., N. A. Clarke, and K. A. Dastal. 1962. Effectiveness of water treatment in virus
removal. J. Am. Water Works Assoc. 54:1275-1292.
Rober, J. A., and F. M. Hoot. 1975. Ultraviolet disinfection of activated sludge effluent
discharging to shellfish waters. MERL, Environmental Protection Agency, Cincinnati, OH,
EPA 600/2-75-060.
Scarpino, P. V., G. Berg, S. L. Chang, D. Dahling, and M. Lucas. 1972. A comparative study of the
inactivation of viruses in water by chlorine. Water Res. :959-965.
Schoub, S. A., and B. P. Sagik, 1975. Association of enteroviruses with natural and artificially
introduced colloidal solids in water and infectivity of solids-associated virions. AppI.
Microbiol. 30:212-222.
Severin, B. F. 1980. Disinfection of municipal wastewater effluents with ultraviolet light. J.
Water Poll. Control Fed. 52:2007-2018.
Shaffer, P. 1. B., 1. G. Metcalf, and 0. J. Sproul. 1980. Chlorine resistance of poliovirus isolants
recovered from drinking water. Appi. Environ. Microbiol. 40:1115-1121.
Slode, J. S. 1977. Enteroviruses in partially purified water. J. Soc. Water Treat. Exam. 31:219-
224.
Snead, M. C., V. P. Olivieri, C. W. Kruse, and K. Kawatta. 1980. Benefits of maintaining a
chlorine residual in water supply systesm. Water Res. 14:403-408.
Spector, B. K., J. R. Boylis, and 0. Gullans. 1934. Effectiveness of filtration in removing from
water, and in chlorine in killing, the causative organism for amoebia dysentery. U.S. Pub.
Health Reports 49:786-800.
-------
Sproul, 0. J. 1971. Recent research results on virus inactivation by water treatment processes.
In: Virus and water quality: occurrence and control. Proc. 13th Water Quality Conf., V.
Sinoeyink and V. Griffin (eds.). University of Illinois at U-C, Urbana, IL, Bull. 69.
Sproul, 0. J. 1980. Critical review of virus removal by coagulation arid pH modification. U.S.
Environmental Protection Agency, Cincinnati, OH. EPA-600/2-80-004.
Sprout, 0. J., L. R. Larochelle, D. F. Wentworth, and R. 1. Thorup. 1967. Virus removal in water
reuse treating processes. Chem. Eng. Prog. Symp. Ser. 63:130.
Sproul, 0. J., Y. S. R. Chen, J. P. Engel, and A. J. Rubin. 1981. Cyst inactivation with chlorine
and chlorine dioxide. Presented at First State-of-the-Art Research Seminar in Environmental
Engineering, U.S. Environmental Protection Agency, Cincinnati, OH.
Stagg, C. H., C. Wallis, and C. H. Ward. 1977. Inactivation of clay associated bacteriophage MS2
by chlorine. Appl. Environ. Microbial. 33:385-391.
Streeter, H. W. 1927. Studies of the efficiency of water purification processes. U.S. Public
Health Service, Washington, D.C., Bull. 172.
Streeter, H. W. 1929. Studies of the efficiency of water purifIcation processes. U.S. Public
Health Service, Washington, D.C., Bull. 193.
Streeter, H. W., and C. T. Wright. 1930. Experimental studies of water purification. V.
Prechlorination in relation to the efficiency of water filtration processes. Public Health
Reports 45:3105-3128.
Stringer, R. P., W. M. Cramer, and C. W. Kruse. 1975. Comparison of bromine, chlorine, and
iodine as disinfectants for amoebic cysts, p. 193-210. In: Disinfection, water and
wastewater, J. D. Johnson (ed). Ann Arbor Science, Ann Arbor, Ml.
Swayne, M. D., G. H. Boone, B. D. Bauer, and J. S. Lee. 1980. Wastewater in receiving waters at
water supply abstraction points. MERL, Environmental Protection Agency, Cincinnati, OH,
EPA-600/2-80-044.
Symons, J. M., A. A. Stevens, R. M. Clark, E. E. Geldreich, 0. 1. Love, and J. DeMarco. 1981.
Treatment techniques for controlling trihalomethanes in drinking water. MERL, Environ-
mental Protection Agency, lncinnati, OH, EPA-600/2-8l- 156.
Taylor, E. W. 1970. Forty-fourth report on the result of the bacteriological, chemical and
biological examination of London waters for the years 1969-1970. Metropolitan Water Board,
London.
Taylor, E. W. 1973. Forty-f Ifth report on the results of the bacteriological, chemical and
biological examination of London waters for the years 1971-1973. Metropolitan Water Board,
London.
Thames Water Statistics. 1976. Thames Water Authority, London 2:E.2. I.
Thayer, S. E., and 0. J. Sproul. 1966. Virus Inactivation in water-softening precipitation
processes. J. Am. Water Works Assoc. 58:1063.
Thorup, R. T., F. P. Nixon, D. F. Wentworth, and 0. J. Sprout. 1980. Virus removal by coagulation
with polyelectrolytes. J. Am. Wafer Works Assoc. 62:97-101.
Tracey, H. W., V. M. Camorena, and F. Wing. 1966. Cot Iform persistence in highly chlorinated
waters. J. Am. Water Assoc. 58:1151-1159.
U.S. Government Printing Office. 1980. Environmental Quality - 1980. Eleventh Annual Report of
the Council on Environmental Quality, USGPO, Washington, D.C.
-------
U.S. Public Health Service. 1946. Drinking water standards - 1946. J. Am. Water Works Assoc.
38:361 and Public Health Reports 61:371.
U.S. Public Health Service. 962. Drinking water standards - 1962. Public Health Service PubI.
956.
U.S. Public Health Service. 1969. Manual for evaluating public drinking water supplies. Public
Health Service PubI. 1820.
U.S. War Department. 944. Efficiency of standard army purification equipment and diatomite
filters in removing cysts of Endamoeba histolytica from water. Submitted to the Engineering
Board, Ft. Belvoir, VA, and the Chief of Engineers, U.S. Army, Washington, D.C. by the
Water Supply Equip. Branch Tech. Div. III, the Engineer Board, Ft. Belvoir, VA, and the Nat.
Inst. Health, U.S. Public Health Service, Report 834.
van der Kooij, D., and B. C. J. Zoefeman. 1978. Water quality in distribution system. mt. Water
Supply Assoc. Congress, Kyoto, Japan.
Vogt, J. E. 196 I. Infectious hepatitis outbreak in Posen, Michigan, p. 87-91. In: Ground Wafer
Contamination, Proc. 1961 Symp., Div. of Water Supply and Poll. Control, 0S. Dept. Health,
Education and Welfare, Tech. Report W6 1-5.
Walsh, D. S., C. E. Buck, and 0. J. Sproul. 1980. Ozone inactivation of floc associated viruses and
bacteria. J. Environ. Eng. Div., Am. Soc. Civil Eng. 106:711-716.
Walton, G. 1961. Effectiveness of water treatment processes as measured by coliform reduction.
Public Health Service PubI. 98.
Warriner, 1. R. 967. lnaktvering cv poliovirus med klordioxid. Vatten 67:284-290.
Weidenkopf, S. J. 1958. Inactivation of type I poliomyelitis virus with chlorine. Virology 5:56-67.
Wellings, F. M., A. L. Lewis, C. W. Mountain, and L. V. Pierce. 1975. Demonstration of virus in
groundwater after effluent discharge onto soil. AppI. Microbiol. 29:751-757.
Wentworth, D. F., R. T. Thorup, and 0. J. Sproul. 1968. Poliovirus inactivation in water-softening
precipitation processes. J. Am. Water Works Assoc. 60:939.
Werner, P. 1978. Mikrobiologische untersuchungen des wasser von trinkwasser -- versorgung-
sanlagen unter besondere berucksichtigung der aktivkohlefiltration. Diplom - Arbeit,
Universitat des Saarlandes, Saarbrucken, FRG.
Werner P. 1979. Okologisch-mikrobiologische untersuchungen an aktivkohlefiltern in zusammen-
hang mit trinkwasser - aufbereitung. Dissertationsarbeit, Universitat des Saarlandes,
Saarbrucken, FRG.
White, G. C. 1972. Handbook of chlorination. Van Nostrand Reinhold Co., New York.
Woodward, L. 1961. Goundwater contamination in the Minneapolis and St. Paul suburbs, p. 66-71.
In: Ground Water Contamination Proc. 196 I Symp., Div. of Water Supply and Poll. Control,
13.S. Dept. Health, Education and Welfare, Tech. Report W-6 1-5.
Yeciger, J. G., and R. 1. O’Brien. 977. Enterovirus and bacteriophage inactivation in groundwater
and translocation in soil. Abstract Paper 57, Annual Meeting Am. Soc. Microbiol., New
Orleans, LA.
Zoeteman, B. C. J. 1972. The suitability of iodine and iodine compounds as disinfectants for small
water supplies. Tech. Paper No. 2, World Health Organization Intnl. Ref. Centre for
Community Water Supply. WHO File Ref. W2/l8/2.
-------
COMPLIANCE AND POLICY ISSUES
Arnold E. Greenberg
Panel Members: Darrel Brock, F. Donald Maddox, Ira M. Markwood, Leland zJ. McCabe, S. M.
Morrison, Nell Roberts, Peggy Wallen, Jerry Williams
ABSTRACT 2
INTRODUCTION 2
CURRENT AND SUGGESTED REVISED NATIONAL DRINKING WATER REGULATIONS 3
Maximum Microbiological Contaminant Levels 3
Existing Regulations 3
Suggested RevIsions 5
Microbial Contaminants Sampling and Analytical Requirements 5
Existing Regulations 5
Suggested Revisions 8
SAMPLING AND ANALYTICAL TECHNIQUES 9
Frequency of Sampling 9
Recommendations 10
Sampling from Distribution System Taps 10
Recommendations I I
Sample Preservation and Transit Time I I
Recommendations 12
Membrane Filter Test, Confluent Growth or Too Numerous to Count 12
(TNTC)
Recommendaflons 13
ANALYSIS OF COLIFORM TEST RESULTS 13
Check Samples 13
Recommendations 14
Use of the Most Probable Number 14
Recommendcit ions 16
Use of Average Collform Numbers 16
Recommendat fans 17
ACTION RESPONSE 17
GENERAL REGULATIONS 18
Standard Plate Count 18
Recommendations 20
Chlorine Substitution 20
Recommendat ions 2 I
ChlorIne Residual DependabIlity 21
Recommendations 22
Presence-Absence Test for Coliform Bacteria 22
Recommendations 22
Turbidity 23
Recommendations 23
Minimum Treatment for Surface Water Sources 23
Recommendations 24
Minimum Treatment for Small Water Systems 25
Recommendations 25
CONCLUDING REMARKS 25
LITERATURE CITED 26
-------
2
ABSTRACT
To proceed from interim to final national drinking water regulations, the Environmental
Protection Agency has undertaken a review of the regulations and operating experiences with
them. This report suggests that simplified but not relaxed regulations with respect to micro-.
biological quality will permit more effective enforcement. It provides for specific revisions of the
maximum contaminant level, and proposes clarifying details with respect to bacteriological
sampling, analysis, and the utilization of bacteriological data.
INTRODUCTION
Pursuant to the Public Health Service Act and the Safe Drinking Water Act (Public Law 93-
523, 40 FR 11990), the Environmental Protection Agency (EPA) adopted National Interim Primary
Drinking Water Regulations (40 FR 59566, December 24, 1975 and 4! FR 28402, July 9, 1976).
These interim regulations have been amended several times. In an effort to proceed from interim
to permanent regulations as directed by the Safe Drinking Water Act, it is necessary for EPA to
review the interim regulations and the experiences of the federal and state regulatory agencies as
well as of the regulated industry.
At the onset of this review it is appropriate to indicate that microbiological standards of
water quality are not intended to define a water which will guarantee safety, that Is, an
impossibility of transmission of waterborne infectious disease. Instead, the standards are designed
to produce water which will hav a low probability, actually a vanishingly small probability, of
transmitting human disease. A second important observation is that when water meets bacterial
standards, such as those first established by the Public Health Service in 1914, the public health has
been well served and waterborne disease outbreaks have been rare and exceptional. Stated
differently, experience accumulated over more than sixty years shows that water meeting the
standards seldom has been involved with disease. This pragmatic conclusion regarding standards
and water safety is critical in any examination of standards because it indicates clearly that the
standards themselves are beneficial although their details and enforcement may be subject to
discussion and dispute.
When the standards were first adopted they were perceived as standards of desirability and
attainability. The desire or goal was to serve consumers a safe water, that is, a water free from
-------
3
fecal contamination as evidenced by the absence of coliform bacteria. Because zero is difficult to
measure (actually because of the statistical nature of the multiple tube dilution technique for
coliform bacteria, it is impossible to measure) the standards were set in terms equivalent to small
numbers of coliform bacteria, say for example, a most probable number (MPN) of I organism per
100 ml of water. The c.ncept of attainability refers to the practical observations that treatment
of surface water could meet the goal at reasonable expense with available technology. On the
whole, these standards were applied to large or relatively large water systems and they worked.
Serious questions about the standard arise when they are applied to small community systems, or
more especially to noncommunity water systems in which conventional technology often is not
applied.
In the development of this paper it became clear that a major revision of the National
Interim Primary Drinking Water Regulations (NIPDWR) was needed. To simplify this presentation,
the current regulations and suggested revisions of Section 141.114 and 141.21 based on the recom-
mendat ions of an Ad Hoc Committee on Microbiological Drinking Water Standards (1977) are given
first. These are followed by a discussion of various operational regulatory concerns, for which if
the recommendations are adopted, further revisions of the regulations would be required. These
issues will be presented under individual headings using the following general format:
i. The pertinent regulatory item, if one exists, or the key issue of regulatory need.
ii. A discussion of the issue including information on the pros and cons.
lii. A recommendation for revised regulations or interpretation of the regulations.
CURRENT AND SUGGESTED REVISED NATIONAL DRINKING
WATER REGULATIONS
Maximum Microbiological Contaminant Levels
Existing Regulations. The existing regulations regarding maximum microbiological contam-
inant levels are included in Section 141.14 of the NIPDWR which reads:
“The maximum contaminant levels for coliform bacteria, applicable to community
water systems and non-community water systems, are as follows:
(a) When the membrane filter technique pursuant to § 141.21(a) is used, the number of
coliform bacteria shall not exceed any of the following:
(I) One per 100 milliliters as the arithmetic mean of all samples examined per
compliance period pursuant to § llel.2 1(b) or (c), except that, at the primacy Agency’s
discretion systems required to take 10 or fewer samples per month may be authorized to
exclude one positive routine sample per month from the monthly calculation if:
(i) as approved on a case-by-case basis the State determines and indicates in writing to
the public water system that no unreasonable risk to health existed under the conditions of
-------
4
this modification. This determination should be based upon a number of factors not limited
to the following: (A) the system provided and had maintained an active disinfectant residual
in the distribution system, (B) the potential for contamination as indicated by a sanitary
survey, and (C) the history of the water quality at the public wafer system (e.g., MCL, or
monitoring violations); (ii) the supplier initiates a check sample on each of two consecutive
days from the same sampling point within 24 hours after notification that the routine sample
is positive, and each of these check samples is negative; and (iii) the original positive routine
sample is reported and recorded by the supplier pursuant to § 141.31(a) and § 141.33(a). The
supplier shall report to the State its compliance with the conditions specified in this
paragraph and a summary of the corrective action taken to resolve the prior positive sample
result. If a positive routine sample is not used for the monthly calculation, another routine
sample must be analyzed for compliance purposes. This provision may be used only once
during two consecutive compliance periods.
(2) Four per 100 milliliters in more than one sample when less than 20 are examined
per month; or
(3) Four per 100 milliliters in more than five percent of the samples when 20 or more
are examined per month.
(b)(I) When the fermentation tube method and 10 milliliter standard portions pursuant
to § 141.2 1(a) are used, coliform bacteria shall not be present in any of the following:
(i) More than 10 percent of the portions (tubes) in any one month pursuant to 141.21(b)
or (c) except that, at the State’s discretion, systems required to take 10 or fewer samples per
month may be authorized to exclude one positive routine sample resulting in one or more
positive tubes per month from the monthly calculation if: (A) as approved on a case-by-case
basis the State determines and indicates in writing to the public water system that no
unreasonable risk to health existed under the conditions of this modification. This
determination should be based upon a number of factors not limited to the following: (I) the
system provided and had maintained an active disinfectant residual in the distribution system,
(2) the potential for contamination as indicated by a sanItary survey, and (3) the history of
the water quality at the public water system (e.g., MCL or monitoring violations); (B) the
supplier Initiates a check sample on each of two consecutive days from the sampling point
within 24 hours after notification that the routine sample is positive, and each of these check
samples is negative; and (C) the original posItive routine sample is reported and recorded by
the supplier pursuant to § 14 1.31(a) and 14 1.33(a). The supplier shall report to the State its
compliance with the conditions specified In this paragraph and report the action taken to
resolve the prior positive sample result. If a positive routine sample Is not used for the
monthly calculation, another routine sample must be analyzed for compliance purposes. This
provision may be used only once during two consecutive compliance periods.
(II) three or more portions In more than one sample when less than 20 samples are
examined per month; or
(iii) three or more portions in more than five percent of the samples when 20 or more
samples are examined per month.
(2) When the fermentation tube method and 100 milliliter standard portions pursuant to
§ 141.21(a) are used, coliform bacteria shall not be present in any of the following:
(i) More than 60 percent of the portions (tubes) in any month pursuant to § 141.2 1(b) or
(c), except that, at State discretion, systems required to take 10 or fewer samples per month
may be authorized to exclude one positive routine sample resulting in one or more positive
tubes per month from the monthly calculation if: (A) as approved on a case-by-case basis the
State determines and indicates in writing to the public water system that no unreasonable
risk to health existed under the conditions of this modification. This determination should be
based upon a number of factors not limited to the following: (I) the system provided and
maintained an active disinfectant residual in the distribution system, (2) the potential for
contamination as indicated by a sanitary survey, and (iii) the history of the water quality at
the public water system (e.g., MCL or monitoring violations); (B) the supplier Initiates two
consecutive daily check samples from the same sampling point within 24 hours after
notification that the routine sample is positive, and each of these check samples is negative;
and (C) the original positive routine sample is reported and recorded by the supplier pursuant
to § 14 1.31(a) and 141.33(a). The supplier shall report to the State its compliance with the
conditions specified in this paragraph and a summary of the corrective action taken to
resolve the prior positive sample result. If a posItive routine sample Is not used for the
monthly calculation, another routine sample must be analyzed for compliance purposes. This
provision may be used only once during two consecutive compliance p&iods.
-------
5
(ii) five portions in more than one sample when less than five samples are examined per
month; or
(iii) five portions in more than 20 percent of the samples when five or more samples are
examined per month.
(c) For community or non-community systems that are required to sample at a rate of
less than 4 per month, compliance with paragraphs (a), (bXl), or (bX2) of this section shall be
based upon sampling during a 3 month period, except that, at the discretion of the State,
compliance may be based upon sampling during a one-month period.
(d) If an average MCL violation is caused by a single sample MCL violation, then the
case shall be treated as one violation with respect to the public notification requirements of
§ 141.32.”
&tggested Revisions. The Ad-Hoc Committee on Microbiological Drinking Water Standards
proposed the following as a replacement of the entire Section 141.14 in the Interim Regulations:
“The following are the maximum contaminant levels for coliform bacteria applicable to
community water systems and noncommunity water systems. Compliance with maximum
contaminant levels for coliform bacteria is determined pursuant to Section 141.21.
(a) When the membrane filter technique pursuant to Section 141.21 (a) is used, the
number of coliform bacteria shall be zero per 100 milliliters in 95 percent of all samples
collected and evaluated pursuant to 141.21 (b) or (c).
(b) When the fermentation tube method and either ten milliliter standard portions or
100 milliliter standard portions pursuant to Section 141.21 (a) are used, coliform bacteria
shall not be present in 95 percent of all samples collected and evaluated pursuant to Section
141.21 (b) or (c).”
Microbial Contaminants Sampling and Analytical Requirements
Existing Regulations. The existing regulations regarding sampling and analysis are included in
Section 141.21 of the NIPDWR which reads:
“(a) Suppliers of water for community and non-community water systems shall analyze
or use the services of an approved laboratory for coliform bacteria to determine compliance
with § 141.14. Analyses shall be conducted in accordance with the analytical recom-
mendations set forth in “Standard Methods for the Examination of Water and Wastewater,”
American Public Health Association, 14th Edition. Method 908A. Paragraphs I, 2, and 3-pp.
916-918; Method 908D, Table 908:l-p.923; Method 909A, pp. 928-935, or “Microbiological
Methods for Monitoring the Environment, Water and Wastes,” U.S. EPA, Environmental
Monitoring and Support Laboratory, Cincinnati, Ohio 45268-EPA-60018/78-017, December,
1978. Available from ORD Publications, CERI, U.S. EPA, Cincinnati, Ohio 45268. Part Ill.
Section B 1.0 through 2.6.2., pp. 108—112; 2.7 through 2.7.2(c). pp. 112-113; Part Ill, Section
B 4.0 through 4.6.4(c), pp. 114-118, except that a standard sample size shall be employed.
The standard sample used In the membrane filter procedure shall be 100 milliliters. The
standard sample used in the 5 tube most probable number (MPN) procedure (fermentation
tube method) shall be 5 times the standard portion. The standard portion is either 10
milliliters or 100 milliliters as described in § 141.14 (b) or (c). The samples shall be taken at
points which are representative of the conditions within the distribution system.
(b) The supplier of water for a community water system shall take coliform density
samples at regular time intervals, and in number proportionate to the population served by
the system. In no event shall the frequency be less than as set forth below:
Minimum number of
Population served: samples per month
25to 1,000
1,001 to 2,500 2
2,501 to 3,300 3
3,301 to 4,100 4
4,101 to 4,900 5
4,901 to 5,800 6
5,801 to 6,700 7
6,701 to 7,600 8
-------
6
7,601 to 8,500 9
8,501 to 9,400 10
9,401 to 10,300 II
10,301 to 11,100 12
11,101 to 12,000 13
12,001 to 12,900 14
12,901 to 13,700 IS
13,701 to 14,600 16
14,601 to 15,500 Il
15,501 to 16,300 18
16,301 to 17,200 19
17,201 to 18,100 20
18,101 to 18,900 21
18,901 to 19,800 22
19,801 to 20,700 23
20,701 to 21,500 24
21,501 to 22,300 25
22,301 to 23,200 26
23,201 to 24,000 27
24,001 to 24,900 28
24,901 to 25,000 29
25,001 to 28,000 30
28,001 to 33,000 35
33,001 to 37,000 40
37,001 to 41,000 45
41,001 to 46,000 50
46,001 to 50,000 55
50,001 to 54,000 60
54,001 to 59,000 65
59,001 to 64,000 70
64,001 to 70,000 75
70,001 to 76,000 80
76,001 to 83,000 85
83,001 to 90,000 90
90,001 to 96,000 95
96,001 to 111,000 100
111,001 to 130,000 110
130,001 to 160,000 120
160,001 to 190,000 130
190,001 to 220,000 140
220,001 to 250,000 150
250,001 to 290,000 160
290,001 to 320,000 170
320,001 to 360,000 180
360,001 to 410,000 190
410,001 to 450,000 200
450,001 to 500,000 210
500,001 to 550,000 220
550,001 to 600,000 230
600,001 to 660,000 240
660,001 to 720,000 250
720,001 to 780,000 260
780,001 to 840,000 270
840,001 to 910,000 280
910,001 to 970,000 290
970,001 to I ,050,000 300
1,050,001 to 1,140,000 310
I, I 40,001 to I ,230,000 320
i,230,00I to 1,320,000 330
1,320,001 to 1,420,000 340
-------
7
1,420,001 to 1,520,000 350
1,520,001 to 1,630,000 360
1,630,001 to 1,730,000 370
1,730,001 to 1,850,000 380
1,850,001 to 1,970,000 390
1,970,001 to 2,060,000 400
2,060,001 to 2,270,000 410
2,270,001 to 2,510,000 420
2,510,001 to 2,750,000 430
2,750,001 to 3,020,000 440
3,020,001 to 3,320,000 450
3,320,001 to 3,620,000 460
3,620,001 to 3,960,000 470
3,960,001 to 4,310,000 480
4,310,001 to 4,690,000 490
4,690,001 or more 500
Based on a history of no coliform bacterial contamination and on a sanitary survey by
the State showing the water system to be supplied solely by a protected ground water source
and free of sanitary defects, a community water system serving 25 to 1,000 persons, with
written permission from the State, may reduce this sampling frequency except that in no case
shall It be reduced to less than one per quarter.
(c) The supplier of water for a non-community water system shall be responsible for
sampling coliform bacteria in each calendar quarter that the system provides wafer to the
public. Such sampling shall begin within two years after promulgation. The State can adjust
the monitoring frequency on the basis of a sanitary survey, the existence of additional
safeguards such as a protective and enforced well code, or accumulated analytical data. Such
frequency shall be confirmed or modified on the basis of subsequent surveys or data. The
frequency shall not be reduced until the non-community wafer system has performed at least
one coliform analysis of Its drinking water and shown to be in compliance with § 141.14.
(d)(l) When the coliform bacteria in a single sample exceed four per 100 millilIters
( 141.14(a)), at least two consecutive daily check samples shall be collected and examined
from the same sampling point. Additional check samples shall be collected daily, or at a
frequency established by the State, until the results obtained from at least two consecutive
check samples show less than one coliform bacterium per 100 milliliters.
(2) When coliform baçterf a occur in three or more 10 ml portions of a single sample
( l4l.l4(b)(l)), at least two consecutive daily check samples shall be coliected and examined
from the same sampling point. Additional check samples shall be collected daily, or at a
frequency established by the State, until the results obtained from at least two consecutive
check samples show no positive tubes.
(3) When coliform bacteria occur in all five of the 100 ml portions of a single sample
( l41.14(bX2)), at least two daily check samples shall be collected and examined from the
same sampling point. Additional check samples shall be collected daily, or cit a frequenc
established by the State, until the results obtained from at least two consecutive chec
samples show no positive tubes.
(4) The location at which the check samples were taken pursuant to paragraphs (d) (I),
(2), (3) of this section shall not be eliminated from future sampling without approval of the
State. The results from all coliform bacterial analyses performed pursuant to this subpart,
except those obtained from check samples and special purpose samples, shall be used to
determine compliance with the maximum contaminant level for coliform bacteria as
established in § 141.14. Check samples shall not be included in calculating the total number
of samples taken each month to determine compliance with § 141.21 (b) or (c).
(e) When the presence of coliform bacteria in water taken from a particular sampling
point has been confirmed by any check samples examined as directed in paragraphs (d) (I), (2),
or (3) of this section, the supplier of water shall report to the State within 48 hours.
(f) When a maximum contaminant level set forth in paragraphs (a), (b) or (c) or § 141. 14
is exceeded, the supplier of water shall report to the State and notify the public as prescribed
In § 141.31 and § 141.32.
-------
8
(g) Special purpose samples, such as those taken to determine whether disinfection
practices following pipe placement, replacement, or repair have been sufficient, shall not be
used to determine compliance with § 141.14 or § 141.21 (b) or (c).
(h) A supplier of water of a community water system or a non-community water system
may, with the approval of the State and based upon a sanitary survey, substitute the use of
chlorine residual monitoring for not more than 75 percent of the samples required to be taken
by paragraph (b) of this section, PROVIDED, That the supplier of water takes chlorine
residual samples at points which are representative of the conditions within the distribution
system at a frequency of at least four for each substituted microbiological sample. There
shall be at least daily determinations of chlorine residual. When the supplier of water
exercises the option provided in this paragraph (h) of this section, he shall maintain no less
than 0.2 mg/I free chlorine throughout the public water distribution system. When a
particular sampling point has been shown to have a free chlorine residual less than 0.2 mg/I,
the water at that location shall be retested as soon as practicable and in any event within one
hour. If the original analysis is confirmed, this fact shall be reported to the State within 48
hours. Also, if the analysis is confirmed, a sample for coliform bacterial analysis must be
collected from that sampling point as soon as practicable and preferably within one hour, and
the results of such analysis reported to the State within 48 hours after the results are known
to the supplier of the water. Analyses for residual chlorine shall be made in accordance with
“Standard Methods for the Examination of Water and Wastewafer,” I 3th Ed., pp. 129-132.
Compliance with the maximum contaminant levels for coliform bacteria shall be determined
on the monthly mean or quarterly mean basis specified in § 141.14, including those samples
taken as a result of failure to maintain the required chlorine residual level. The State may
withdraw its approval of the use of chlorine residual substitution at any time.
(i) The State has the authority to determine compliance or initiate enforcement action
based upon analytical results or other information compiled by their sanctioned representa-
tives and agencies.”
Suggested Revisions. The Ad-Hoc Committee on Microbiological Drinking Water Standards
proposed the following replacements for paragraphs (d), (e), (f) of the Interim Regulations;
“(dX I) When a maximum contaminant level of Section 141. 14 (a) or (b) is exceeded, or
when the coliform count per 100 ml in any routine sample exceeds 4 when using the
membrane filter technique or when three or more tubes are positive when using the
fermentation tube method and 10 ml standard portions (or 5 tubes are positive when using
100 ml standard portions), one check sample shall be collected from the same sampling point
and examined. If a subsequent sample has already been collected from the sampling point, it
shall be considered the check sample.
(2) When the examination of the check sample required in Section 141.21 (d) (I) shows
the presence of coliform organisms, the supplier of water shall:
(i) Report to the State within 48 hours, and
(ii) Immediately initiate an investigation by an individual and in a manner satisfactory
to the State, including the collection and examination of a check sample from the same point
and other sampling points in the area, and
(iii) Simultaneously institute a review of treatment practices, Including examination
and cleaning of the distribution system, and all other possible sources of contamination, and
(iiii) Institute appropriate corrective action.
(3) When the examination of the check sample required In Section 141.21 (d) (I) shows
the absence of coliform organisms, the supplier shall:
(i) Maintain the original sample data in the plant records and forward this information
to the State.
(ii) Meet the State reporting requirements to the State for the specif led period through
inclusion of check sample results.
(4) Following the investigation required in (herein suggested) Section 141.21 (d) (2) (II),
and based upon the findings of that investigation, the supplier of wafer shall notify the public
in the area affected by the indicated contamination as prescribed In Section 141.32, or the
State public notification requirements.
(5) The State may, at its discretion, require check samples be collected at a specifIed
frequency from the same sampling point and other sampling points In the area and examined
to identify and eliminate suspected health hazards when a sample exceeds a maximum
-------
9
contaminant level pursuant to Section 141.14 (a) or (b) even if the check sample required in
Section 141.21 (d) (I) does not indicate the presence of coliform bacteria.
(6) After the cause of the indicated contamination has been determined and corrected,
additional check samples shall be collected at a frequency directed by the State. The
requirement for notification of the public shall end when corrective action, satisfactory to
the State, has been taken and additional check samples do not indicate the presence of
coliform bacteria.
(7) The location at which the check samples were taken pursuant to Paragraph (I) of
this section shall not be eliminated from future sampling without approval of the State.
(e) The State may determine that unreliable examination results for a sample collected
in a discrete monitoring period pursuant to Sections 141.21 (b) or (c) were caused by factors
beyond the control of the water supplier. Such factors do not include sampling errors but
could be excessive transit time between collection and examination of the sample, samples
being broken in transit, or interference in test results when the membrane filter technique is
used, If this is the case, and the supplier of water does not learn of these results until the
following monitoring period, another sample collected immediately thereupon may be
attributed to the previous monitoring period in determining compliance with Section 141.21
(b) or (c). However, a single sample may not be attributed to more than one monitoring
period.
(f) Samples with unreliable examination results, and special purpose samples, such as
those taken to determine whether disinfection practices following water main placement,
replacement, or repair have been sufficient, shall not be used to determine compliance with
Section 141.21(b) or (c).”
SAMPLING AND ANALYTICAL TECHNIQUES
Frequency of SampIInQ
Section 141.21 (b) of NIPDWR specifies the minimum frequency of sampling for coliform
bacteria in community water systems as a function of the population served. The minimum number
of samples per month ranges from I (for a population of 25 to 1000) to 500 (for a population of
4,690,001 or more). SectIon 141.21 (c) specifies that for non-community systems sampling shall be
once per each calendar quarter in which the system serves water to the public.
California (California Administrative Code, Title 22, Section 64421 (e)) specifies minimum
sampling requirements for community systems on the basis of service connections (excluding fire
hydrants). The requirement for noncommunity systems is essentially identical to the federal one.
Other schemes for defining frequency of sampling depend on type of source, i.e., surface vs.
groundwater or pumping rate. With the exception of the source type, all of the methods of defining
sampling frequency provide essentially equivalent information. There may be minor variations in
the minimum sampling frequency required, but no method appears to have any decided advantages
although service connections or pumping rate may be more easily defined by the water utility.
For non-community water systems the required minimum number of samples may be as few
as one per year and leaves much to be desired. In sampling water systems using groundwater it
-------
lO
may be assumed that water quality is relatively stable; for surface sources this assumption cannot
be made. Section 141.21 (c) permits a state to adjust sampling frequency on the basis of a sanitary
survey or additional safeguards. This suggests that bacteriological monitoring is less than ideal for
such small systems (and possibly even larger ones) and gives indirect weight to the significance of
the sanitary survey. The World Health Organization (1982) states this more emphatically: “No
bacteriological or chemical analysis of samples, however carefully it is made, is a substitute for
complete knowledge of conditions at the source and within the distribution system .. . Samples
represent a single point in time, and are reported after the fact. Contamination is often random
and intermittent and is not revealed by occasional sampling.”
It is widely argued that a single sample per month provides inadequate assurance of
bacteriological water quality. To provide statistically significant data considerably more samples
(not less than 5) would be required. Because of the costs of sampling and sample analysis, it is
suggested that a compromise be made and that no community water system should be sampled less
frequently than twice per month. For noncommunity water systems it is proposed that two
categories, large and small, be identified. A large non-community water system is a system
serving 500 or more persons per day when the system is in operation; a small non-community water
system serves fewer than 500 persons. A large non-community system should be sampled as though
it were a community system. A small non-community water system should be sampled in
accordance with Section 141.21 (c).
Recommendations. With regard to the frequency of sampling, the following are
recommended:
I. There is no reason to change from the existing frequency table of Section 141.21.
2. For community water systems, and large non-community water systems, the minimum
number of samples per month should be two (2).
3. Emphasis should be given in the regulations to the significance of a sanitary survey in
addition to regular bacteriological monitoring.
Smipling from Distribution System Tqs
The NIPDWR are silent on how or where bacteriological samples shall be collected but cites
APHA (l97 ) and EPA (1978) for specific analytical methods. In sampling from a distribution
system tap without attachments, the former states:
-------
II
“. . . it should be ascertained that the tap chosen is supplying water from a service pipe
directly connected with the main, and is not, for example, served from a cistern or storage
tank. The top should be opened fully and the water allowed to run to waste for 2 or 3 mm, or
for a time sufficient to permit clearing of the service line. The flow from the tap should
then be restricted to one that will permit filling the bottle without splashing. Leaking taps
that allow water to flow over the outside of the tap must be avoided as sampling points.”
The latter reference (page 14) is more specific with respect to the tap:
“Make certain that samples are not collected from spigots that leak around their stems,
or from spigots that contain aeration devices or screens within the faucet.”
The current, 15th edition of APHA (1981) echoes the EPA manual by saying:
“In sampling from a mixing faucet remove faucet attachments such as sckeen or splash
guard . . .“
The AWWA (1978) states:
“Do not sample from drinking fountains or taps which have aerators, strainers, or swivel
faucets, or taps off of individual homeowner treatment units.”
By contrast, Marquart (1980) examined 1,080 samples and concluded that:
restrictive requirements on tap selection and procedure in collecting routine
bacteriological water samples are not justified. The requirements on tap selection and
procedure for collecting bacteriological water samples quoted from Standard Methods for the
Examination of Water and Wastewater, 1975, (14th Ed.) are adequate to provide reliable
bacteriological water sample results.”
Recommendations. In light of the conflicting views relative to tap sampling procedures the
following are recommended:
I. Conduct further studies on the need for removing faucet attachments before sampling.
2. Unless there is confirmatory evidence showing that there is no need to remove faucet
attachments, specify that attachments shall be removed.
3. Modify Section 141.21 to include specific requirements regarding sampling taps or
spigots consistent with the research findings.
Sample Preservation and Transit Time
The NIPDWR are silent on how or if bacteriological samples are to be preserved and the
permissible elapsed time between sample collection and analysis. The references cited above (in
Section 141.21 (a)) agree that samples should be held at a low temperature (less than 10°C or
1-4°C) for as short a time as possible and that “the time elapsing between collection and
examination should in no case exceed 30 hr.”
The technical literature contains a voluminous amount of material concerning the effect of
storage time and temperature on coliform bacteria counts in water (see APHA 1976, page 907 for
-------
12
typical references). Individual studies could be cited to show that coliform numbers in stored
water samples increase, decrease, or remain constant during storage cit refrigerator temperature,
room temperature, or a temperature below the ambient water temperature when samples are held
for 6, 24, 30, or 48 hours. From reviewing this literature it is reasonable to conclude that sample
storage, under any conditions, Is undesirable. Ideally, samples should be collected and analyzed
promptly, certainly on the same day as collection. Practically, however, this is not always possible
because water laboratory services may not be available at the site of a water treatment and distri-
bution facility, especially for non-community water systems which additionally often are geo-
graphically isolated. In special studies a field laboratory may be used. For routine sampling the
deiayed-incthation membrane filter technique ( Standard Methods , page 935) may be used but
practical difficulties with untrained personnel at non-community systems may preclude this option.
Samples delivered by a common carrier to the water laboratory may be the only solution possible at
a reasonable cost.
A recent comprehensive study (Nash and Geldreich, In press) showed that when potable water
samples (rather than more highly contaminated samples) containing low levels of coliform bacteria
were held up to 54 or 72 hours, significant changes in coliform counts did not occur if the samples
were refrigerated.
Reoomrnendatf y . it is recommended that Section 141.21 (a) be modified to specify that
samples should be analyzed as soon as possible, and that samples shall be analyzed no later than 30
hours after collection if not refrigerated from the time of collection or analyzed within 54 hours of
collection if refrigerated from the time of collection.
Membrcme Filter Test, Confluent Growth or Too N ierous to Count (TNTC)
In specifying analytical methods for coliform bacteria, the regulations (Section i4i.2l (a))
simply cite references (APHA 1976 or EPA 1978) wIthout providing details. This approach Is ideal
but it may lead to Interpretative difficulties.
Membrane filters showing confluent growth (“growth covering the entire filtration area of the
membrane with no discrete colonies”) or too numerous to count (TNTC) (which occurs when “the
total number of bacterial colonies, coliforms plus non-coliforms, exceeds 200 per membrane”) are
reported as such but essentially the sample result is rejected and a new sample Is to be requested.
Such an action, in fact, may reject useful information. Using current regulations which depend on
-------
13
the quantitative analysis for coliform bocteria, no other practical conclusion is possible. With the
suggested revised MCL, the showing of any coliforms is meaningful and such samples need not be
rejected although resampling is desirable. Membranes with confluent growth or TNTC, provided
typical sheen is present even though discrete sheen producing colonies are not, may be considered
as demonstrating the presence of coliform bacteria in the sample.
Recommendations, if by the process of verification a membrane filter with confluent growth
or TNTC shows evidence of any coliform bacteria, it shall be considered that the sample contained
at least I coliform bacterium per 100 milliliter; another sample should be collected immediately.
If there is no evidence of coliform bacteria on such membranes, the sample shall be rejected and
another sample shall be collected immediately for compliance purposes.
ANALYSIS OF COLIFORM TEST RESULTS
check Sc iptes
Section 141.21 (d) of the Interim Regulations requires that when the coliform limits of 141.14
are exceeded, consecutive daily check samples shall be collected and analyzed. Further, check
samples shall not be included in determining compliance with Section 141.14.
Implicit in the requirement for repeat sampling from the same sample point following a so-
called bad sample (one which exceeds the Maximum Contaminant Level), is the assumption that
such samples occur occasionally because of improper sampling techniques, laboratory errors, or
other, possibly unexplained, sampling or laboratory causes, If such causes are involved rather than
an actual deterioration of bacteriological water quality, repeat sampling probably would produce
negative results. As a consequence it would be appropriate to include check samples in deter-
mining compliance so as to dilute the total effect of the occasional spurious result in calculating
the monthly average as specified in the interim Regulations. Alternatively, a bad sample, if
followed by negative check samples, could be excluded from the calculation (141.14 (a)). Such a
course of action may eliminate unnecessary public notification.
On the other hand, the requirements for check samples somehow Implies that water quality
can be tested into the product. Section 141. 14 deals with the bad result, check samples, and paper
manipulations regarding reporting results. It fails totally to deal with the real problems of water
quality. Thus, if by the use of check samples, it should be demonstrated that the original bad
-------
‘4
imple was indeed unrelated to water quality, then the matter could be dismissed and the bad
sample disregarded. But if the bad sample was shown to reflect bad water quality, the regulations
should impose a requirement for corrective action, not merely more sampling and statistical
manipulations. The imposition of exact and strict numerical standards of water quality fails to
recognize the inherent variability of bacterial tests and bacterial behavior and, even more
important, fails to recognize that the MCL is not an absolute dividing line between safe and unsafe
wafer. There is a wide but ill-defined gray area between the safe and unsafe waters. Dealing with
this questionable area requires the exercise of professional Judgment rather than the blind
adherence to numerical standards.
Recommendations. Adopt the suggested revised Section 141.21 which deals with check
samples and corrective actions.
Use of the Most Probthle Nunther
Section 14 I. 14 (2) specIfies that when the fermentation tube method Is used for coliform
bacteria, the Maximum Contaminant Levet is set in terms of the number of portions (tubes)
positive by the confirmed test.
The present practice was initiated with the adoption of drinking water standards In 1914 and
has never been changed. It has been suggested, however, that a calculation based on the number of
tubes positive for a single sample, the Most Probable Number (MPN), be used in place of the
number of tubes positive. If the MPN were used to report results In the required test commonly
performed (5 tubes each Inoculated with 10 ml of the same sample) the possible MPN results would
be as indicated in Table I.
It is argued that manipulating results, calculating averages, displaying results, etc., would be
simplified and that when public notification would be required It would be possible to convey more
information through using the MPN than through percent tubes positive. ExaminOtion of Table I
leads to serious challenges to these contentions. While the use of the small numbers obtained for
the MPN does indeed simplify reporting, using the MPN may convey an impression of accuracy of
measurement not appropriate to the test. The 95% confIdence limits shown in the table indicate
clearly this inexactness. Furthermore, informing the public about the presence of coliforms In
decimal numbers can be confusing, not clarifying. Coliform bacteria usually are distributed log-
normal rather than normal. This means that the geometric mean and not the arithmetic mean
-------
‘5
Table I. Possible MPN Results
No. of tubes
Giving Positive
Reaction out of
5 of 10 ml each
MPN
Index
per 100
ml
95%
Confidence
Limits
Lower
Upper
0
<2.2
0
6.0
I
2.2
0.1
12.6
2
5.1
0.5
9.2
3
9.2
1.6
29.4
4
6.0
3.3
52.9
5
> 6.0
8.0
InfinIte
-------
16
would be better suited to summarizing MPN values. Percent tubes positive, by contrast can be
averaged arithmetically with no loss of sensitivity or Information. Calculating the arithmetic
mean is simpler than determining the geometric mean.
Recommendations. Continue to report coliform results, when using the fermentation tube
method, by percent tubes confirmed test positive, not MPN.
Us. of Avera . Coflform N nbers
Section 141.14 (a) (I) provides the only reference to the use of averages and specifies that
when the membrane filter technique Is used the MCL shall be “one per 100 milliliters as the
arithmetic mean of all samples examined per compliance period. . .“. No mention Is mode of the
use of averages for the fermentation tube method although it is stated that no “more than 10
percent of the portions (tubes) in any one month . . .“ may be positive.
For community systems analyzing many samples per month, using monthly averages or
percent tubes positive adequately summarizes the accumulated data. However, If the suggested
revised MCL is adopted (in effect, that coliform bacteria shall be absent from 95 percent of the
samples tested per compliance period), averaging becomes unnecessary and the entire process
would be simplified. For small systems that are required to analyze as few as two samples per
month or one sample per quarter there are so few data points that some means of accumulating
data over time would seem to be necessary. If 10 or fewer samples are required per month the
proposed MCL would require that the number of coliform bacteria be less than 1/100 ml in all
samples. This suggests that it would be helpful to use averaging, or rather the accumulating over
time, of enough samples so that a single positive result does not invariably constitute a violation.
This number would be 20 samples. Thus if 20 samples were accumulated, the proposed MCL would
require that in 19 samples (95%) collform bacteria should be less than 1/100 ml (membrane filter
technique) or absent (fermentation tube method); this Is a modified moving average based on
sample number, not time. To avoid a situation In which a water system analyzing two samples per
month would be In violation of the MCL for 10 months (20 samples) when two samples are positive,
the suggested revised regulation (Section 141.21(d) (6)) states that the need for public notification
ceases when correction action has been taken and demonstrated as successful. Successful
correction action would permit dropping the Investigated positive sample from the moving average.
-------
17
For non-community small water systems the calendar-quarterly sample clearly is Inadequate
and data accumulating or averaging is meaningless. Alternate reliance on the sanitary survey, as
discussed above In Frequency of Sampling and currently detailed in proposed Section 141.21 (c),
provides the basis for true protection.
Recommendations. With regard to average coliform numbers, the following are recommended:
I. Adopt the suggested revised MCL (Section 141.14), thereby making averaging Irrelevant.
2. When fewer than 20 samples are required to be analyzed per month, assess compliance
on the basis of the most recent 20 samples.
3. Adopt the suggested revised Section 141 .21(d) and drop the positive sample from the
accumulated samples when the investigation has Identified the source of contamination
and the situation has been corrected.
ACTION RESPONSE
The NIPDWR prescribe Maximum Contaminant Levels (Subpart B) and monitoring and
analytical requirements (Subpart C) but with the exception of reporting and public notification they
are silent as to the need for corrective action (although the need ii implied) In the event that an
MCL is exceeded.
From the failure to Indicate that corrective action Is required, the inference may be drawn
that by collecting and analyzing check samples and notifying the public of the deficiency the
situation will be improved; this is clearly unlikely. Because the causes for the appearance of
coliform bacteria are various It may be impossible to specify the action response but this does not
obviate the need for such a response. California, at least, requires (California Administrative
Code, Title 22, Section 64461) that when there is a significant rise in the bacterial count the water
supplier shall “Furnish information on the current status of physical works and operating procedures
which may hove caused the elevated bacteriological findings, or any Information on community
Illness suspected of being woterborne.” This requirement indicates to the water utility the need
an Investigation even though one is not specifically required or defined. Thus, in this sense, the
California regulations are better than NIPDWR, but neither is good. The above suggested revised
regulations specify that If coliform bacteria are present, the water purveyor will “immediately
iris tate on investigation.. . and simultaneously Institute a review of treatment practices, Including
-------
18
examination and cleaning of the distribution system . . . and institute appropriate corrective
action.” This language is not restrictive by being overly specific and is general enough to leave
details of the investigation and remedial actions to the judgment of the purveyor but it does clearly
specify that action is required to eliminate the cause of bad water.
A related matter is timing for the investigation and corrective action. Because of the
potential for waterborne infection, it seems appropriate and prudent to begin the investigation as
soon as a result that exceeds the MCL is reported from the laboratory. In this way there will be no
delay in determining the reason for the result. Simultaneous with the initiation of the
investigation, the process of collecting check samples should begin. If a check sample is negative,
no corrective action will be necessary while, on the other hand, if a check sample is positive the
assessment of the problem will have been completed, or at least it would be underway, and
corrective action could be expedited. Because bacteriological results at best constitute a
historical record of water quality, delay in correcting water systems deficiencies is unacceptable.
Recwn,nendat ions. The recommendations relative to action response are:
I. Adopt the suggested revised MCL including greater emphasis on corrective action.
2. The sanitary survey should be stressed as an adjunct to sampling and a preliminary to
corrective action.
GENERAL REGULATIONS
Standm d Plate Count
The N1PDWR are silent with respect to the need for or desirability of an MCL based on the
total microbial count or a standard plate count (SPC). Proposed National Secondary Drinking
Water Regulations discussed the need for an MCL but when they were adopted (1 L. FR ‘2 195, July
19, 1979), the regulations did not include an MCL for the Standard Plate Count.
In discussing the need for an MCL for standard plate count it is useful to define the intended
test and the results to be obtained. APHA (1976) includes such a test and states “the Standard
Plate Count procedure provides a standardized means of determining the density of aerobic and
facultotive anaerobic heterotrophic bacteria in water. This Is an empirical measurement . .
Alternative terms for the test include colony count, total microbial count, viable count, water
plate count, total bacterial count, and aerobic mesophilic viable bacterial count. Irrespective of
-------
19
the term applied and the specifics of the analytical method, the procedure provides a quantitative
estimate of the number of bacteria producing colonies under defined conditions of culture medium,
incubation time and temperature, etc. Results obviously will be affected by the test technique so
that it is essential to use a standardized analytical method, Included among the organisms
enumerated will be members of such bacterial genera as Pseudomonas, Flavobacterium,
Aeromonas, Proteus, K lebsiel Ia, Bacillus, Serratia, Corynebacterium, Spi ri I lum, Clostridium,
Arthrobocter, Gollionella, Leptothrix , and Achromobacter .
EPA (1976) has produced a good review of information on the general bacterial population in
water. The NATO (in press) report and the WHO (in press) report likewise include useful reviews.
The NAS (1977) concluded that
“The standard plate count is a valuable procedure for evaluating the bacterial quality of
drinking water. Ideally, standard plate counts should be performed on samples taken
throughout systems. The SPC has major health significance for surface-water systems that
do not use flocculation, sedimentation, filtration, and chlorination, and for those groundwater
systems that do no chlorination. When it is used, the sampling frequency should be at least
10% of the frequency of the cotiform analysis, except that at least one sample should be
collected and analyzed each month.”
The NATO report (in press) cited above was less positive with respect to the need for a standard
based on the colony count (“. . .should n 9 t be considered essential for assessing the safety of
potable water supplies . . .“) but did consider it useful in evaluating apparent pollution problems.
(“The main value of colony counts lies in comparison of the results obtained from regular samples
from the same supply so that any significant change from the normal range in a particular loGation
can be detected.”)
Available data (see EPA-570/9-76-003) indicate that increased health risks to specific
populations, as in hospitals or nursing homes, are associated with high plate counts even in the
absence of coliform bacteria. Pseudomonas , for example, is an opportunistic pathogen. It would be
enumerated by the plate count and is a particular problem in hospitals where it causes a significant
number of nosocomial infections (Hoadley, 1976).
Non-coliform bacterial populations may suppress the development of coliform bacteria to
yield false negative or low coliform counts. LeChevallier, et al., (1980) found that when more than
20 percent of the plate count population were coliform antagonists, the coliform population was
significantly suppressed. Similarly, Geldreich, et al., (1972) found that when the plate count
exceeded 500 colonies per milliliter there was coliform suppression. LeChevallier, et al., (1980)
-------
20
also found a positive correlation between increased plate count numbers and wafer turbidity. High
quality water characteristically has a low plate count; after adequate treatment it may approach
zero.
From the foregoing it appears then that there is significant technical justification for
establishing a total count MCL.
Arguments in opposition to such a standard include the increased cost of water testing, the
increased need for public notification when insufficient samples are analyzed (a technical violation
of the MCL that is unrelated to bacteriological water quality but is related to water system
management), and the general difficulties associated with testing infrequently at small water
systems (see Frequency of Sampling, above). By establishing an MCL and permitting noncommunity
systems to vary from it on the basis of sanitary survey information, as in Section l1 l.2l Cc), most
objections can be dealt with adequately.
Recommendations. The total colony or standard plate count (SPC) is a useful Indication of
water quality and water treatment efficiency. Hence the following are recommended:
I • The use of SPC Is encouraged for all water systems.
2. For surface water systems riot providing coagulation, sedimentation, filtration, and
disinfection, or equivalent, and for groundwater systems not providing disinfection, the
SPC should be determined when In the test for coliform bacteria I) by the membrane
filter test, the filter shows colonies too numerous to count (TNTC) or confluent growth,
or I I) by the multiple tiibe method there Is evidence of Interference with coliform
growth.
3. The MCL of 500 colonIes per milliliter shall be applied when the SPC test Is required.
4. The test method shall be that described In APHA (1976) or APHA (1981).
5. For reasons of feasibility, sampling shall be consistent with the MCL and sampling
requirements for coliform bacteria.
6. Research should be conducted to determine equivalent bacterial removal by methods
other than conventional coagulation, sedimentation, and filtration.
Chkwlns Sib.tltutlon
Section 141.21 (h) of N IPDWR specifies that a substitution of chlorine residual monitoring
may be made for 75 percent of the required bacteriological samples based on a sanitary survey,
-------
21
a daily determination of the chlorine residual, and the maintenance of a chlorine residual
throughout the distribution system of at least 0.2 mg/I of free chlorine.
This procedure permits substitution of an indirect chemical test for a major part of the
bacteriological monitoring but, if attention is paid only to the test substitution, the inference could
be made that there is an equivalence in the two tests and the significance of their results. Such an
inference would be incorrect, and because reliance Is not placed simply on water testing, this
conclusion also is irrelevant. In practice, most systems that are likely to make use of this
substitution are small community or noncommunity water systems that do not operate their own
bacteriological laboratory and find that bacteriological testing is difficult because of the
immediate unavailability of a laboratory and the cost of testing. Furthermore, taking a possible
worst case in which only a quarterly bacteriological sample is required, it is theoretically possible
that the plant operator may not even be physically present at the plant for three months at a time.
By contrast, in using the substitution rule, the operator can make a simple test using a field kit for
which operating instructions are relatively simple and will be in attendance on possible problems in
the water system at least daily. This practical improvement in attention given the water supply
should be invaluable in protecting the public health. It has been mentioned earlier that there is
serious doubt about the value of infrequent bacteriological tests; daily tests for residual chlorine
are infinitely preferable.
Despite the apparent merits of the chlorine substitution rule, few states, if any, have adopted
it or encouraged its use. This suggests that the rule, in effect, is irrelevant. States should be
encouraged to promote use of the chlorine substitution rule which in no way represents a relaxing
of the standards.
Recommendations. Unless investigation shows that chlorine substitution is being effectively
used by a significant number of states, it should be eliminated from the regulations even though It
Is potentially valuable as a method of public health protecfion.
ChIcrine Residual Dependthlllty
Section 141.28 specifies that samples shall be analyzed in an approved laboratory except that
“measurements for turbidity, free chlorine residual, temperature and pH may be performed by any
person acceptable to the State.” No criteria for “acceptable to the State” are provided.
-------
22
Quality assurance with respect to field measurement of chlorine residuals often is absent,
that is, field kits and their reagents may not be adequately standardized and/or maintained and
operators may be insufficiently trained to correctly perform such measurements. The end result
may be seriously flawed data on chlorine residuals and a false sense of security that the water is
being satisfactorily disinfected. Although remedying this situation is a responsibility of the state,
it would be appropriate to provide as a supplement to the regulations an advisory on what
acceptability criteria should be and how they should be applied.
Recommendations. Develop an advisory for Section 141.28 of NIPDWR indicating what should
constitute acceptable criteria for individuals making field measurements for chlorine.
Presence-Absence Test far Coilform Bocteila
The suggested revised MCL (Section 141.14) specifies the absence of coliform bacteria in
95 percent of the samples tested but also imposes a single maximum coliform number (Sec-
tion 141.21 Cd) (I). This Is in keeping with the goal of distributing to consumers a water free from
coliform bacteria and makes unnecessary, for routine samples, the need for quantifying the
collform bacteria. If this MCL were to be adopted a simpler, cheaper analytical technique that
merely determines the presence or absence of coliform bacteria would be adequate. Such a
presence-absence (P-A) test has been proposed by Clark (1968). While it is not argued that Clark’s
test should be adopted, it should be agreed that his concept is valid and that efforts are needed to
develop a suItable test for presence or absence of coliform bacteria. ideally, this test would
permit the analysis of large samples (at least 100 ml), provide results no less rapidly than the
presently used membrane filter technique, and be simple and inexpensive to execute. Canadian
experience shows the merits of using a P-A test. Quantitative analyses would be required when the
presence of coliform bacteria was demonstrated.
Recommendations. The following recommendations are made with regard to the P-A test:
I. Adopt the suggested revised MCL and conduct a major study to develop an acceptable
presence-absence test for coliform bacteria.
2. To quantitatively supplement coliform data, adopt the recommended MCL for total
colony count.
3. Check samples should always be analyzed quantitatively.
-------
23
Turbidity
Section 1141.13 establishes an MCL for turbidity in surface water systems. The MCL is
I turbidity unit (TU) but it may be increased to not more than 5 TU under certain circumstances.
Two points may be made with respect to this MCL. First, the MCL is applicable only to
water systems using a surface water source. This suggests that groundwater is invariably free from
turbidity which is untrue. Turbidity in a groundwater system could result from such causes as
pumping sand or precipitation of insoluble iron compounds. What is less known is the effect that
groundwater turbidity might have on coliform bacteria recovery. LeChevallier, et al., (1980) have
shown a relationship between high turbidity and high plate counts in potable water, which suggests
turbidity interferes with adequate disinfection. For these reasons alone a turbidity MCL is
important.
Second, the MCL for surface water may be as high as 5 TU if the water purveyor can
demonstrate, among other things, that the turbidity does not “(3) interfere with microbiological
determinations.” Again, it is unclear how this interference would be demonstrated.
Recommendations. Do not change Section 141 . 13, Maximum Contaminant Levels for turbidity,
but conduct research to determine the effect of turbidity on coliform bacteria recovery.
Mlnlmtxn Treatment for Surface Water Sources
The NIPDWR prescribe Maximum Contaminant Levels (Subpart B) and monitoring and
analytical requirements (Subpart C) but are silent as to how these levels are to be achieved.
The entire thrust of the regulations is to define the quality of water that is to be served to
consumers. In no area do they even suggest how standards are to be met or what treatment
processes should be used. This is completely consistent with the philosophy of standard setting
which assumes that when the goals are established, available technology will be employed to
achieve them. If minimum treatment for surface water sources for meeting the bacteriological
MCL is to be established, consistency and logic would require comparable treatment specifications
for inorganic and organic chemicals, turbidity, and radioactivity. The consequence would be
complete reformation of the NIPDWR. However, the Safe Drinking Water Act (Section 1401)
indicates that MCL’s shall be set and that only if “...it is not economically or technologically
feasible to so ascertain the level of such contaminant...” shall a “...treatment technique known to
the Administrator [ of EPA] which leads to a reduction in the level of such contaminant sufficient to
-------
24
satisfy the requirements...” be established, It seems clear that in the case of microbiological
contaminants which can be readily measured at reasonable cost there is no authority to define
treatment methods although it should be emphasized that readily available methods measure
coliform bacteria, the indicator of fecal contamination, rather than the pathogenic organisms of
direct concern.
Leaving aside the issue of legal authority and regulatory consistency, is there a need to
specify a minimum treatment for surface water sources? Prudent sanitary engineering and public
health concerns certainly would yield the conclusion that no surface water source should be used
for a public water supply without complete treatment (coagulation, sedimentation, and filtration or
a comparable process yielding equivalent results). Recent information on the incidence of
waterborne giordiasis (see especially W. Jakubowski & J. C. Hoff, 1979) and the apparent failure of
conventional chlorination to destroy the causative organism indicate an even more acute reason for
requiring complete treatment. When these are coupled with the current inability to accurately
determine the presence of cysts of Giardia lamblia in water there may be both technical dnd legal
bases for requiring at least filtration.
If in addition, one considers the relationship of turbidity and adequacy of disinfection (see
Environmental Protection Agency, 1976), as well as the interfering effect of turbidity on the
determination of coliform bacteria, especially by the membrane filter technique, it is reasonable to
conclude that filtration should be required for all surface waters because they may be turbid and
also may contain Giardia cysts.
NAS (1977) succInctly summarized the situation with respect to water testing, water
treatment, and public health: “Good engineering and public health practices emphasize the need
for using raw water of the highest possible quality, and for providing adequate sanitary survey
information. Bacteriological testing - - or the imposed use of bacteriological standards -- are
adjuncts, not replacements for good-quality raw water, proper water treatment, and integrity of
the distribution system.”
Recommendations. The following are recommended regarding treatment requirements:
I. In an appropriate format revise the NIPDWR to include a requirement that, to produce
microbiologically safe water from a surface source, coagulation, sedimentation, and
filtration, or equivalent, and disinfection, is necessary.
-------
25
2. Conduct research on filtration and prefiltration treatment to determine the minimum
treatment for removal of cysts of Giardia lamblia and turbidity.
3. Conduct research on analytical methods for Giardia in water.
Minimum Treatment for Small Water Systems
As has been indicated already the NIPDWR are silent with respect to treatment require-
ments, irrespective of water source or water system size.
In the case of small water systems, and most notably for noncommunity water systems,
definite deficiencies exist in bacteriological monitoring and the assessment of water safety; this
has been discussed above, especially in the section on Frequency of Sampling. The recommendation
was mode that the sanitary survey should be given greater weight in addition to or in place of
regular bacteriological monitoring. An alternative approach has been suggested to meet this need,
namely, to require minimal treatment, at least disinfection, for every system for which infrequent
bacteriological analyses are mode. Thus, If adequate assurance of safety Is not provided by a
suitable monitoring program, require disinfection, or more, as a minimal treatment. Such a
requirement would be discriminatory to the small system, possibly prohibitively expensive, not
technically justified for all systems, and contrary to the tow (see section on Minimum Treatment
for Surfoce Water Sources). Although this alternative should be rejected 1 It attempts honestly to
deal with a real problem that needs resolution or d would be a simple means to improve public
health protection. The original proposal of engineering controls and the sanitary survey is more
appropriate and acceptable.
Recommendations: Do not revise the NIPDWR to require treatment for small water systems;
revise the regulations to require meaningful sanitary surveys.
CONCLUDING REMARKS
From the foregoing review It should be abundantly clear that the drInking water standards,
though they are not perfect and do need revisions, generally serve the public health well. What Is
not clear is whether slavish enforcement of numerical standards or the combination of professional
expertise by the water works Industry and the regulatory bodies and the judicious use of numerical
standards, has been responsible for this high level of protection. Water sampling and analysis, the
monitoring programs prescribed by the regulations, provide valuable Information about water
-------
26
quality. However, no monitoring program can provide the best assessment of water safety nor can
it make a bad water safe. On-the-spot examination of the water system, the sanitary survey,
should be the first level of water system evaluation. Monitoring data can support the sanitary
survey but never replace it.
Laboratory testing of water is the heart of the drinking water regulations, but intelligent
application of professional judgment is the key to providing consumers a safe water.
LITERATURE CITED
Ad Hoc Committee on Microbiological Drinking Water Standards (Advisory to EPA, Office of
Drinking Water) 1977, A. E. Greenberg, Chairman. Recommendations f or Amending Micro-
biological Standards for Drinking Water, April 15, 1977. American Water Works Association,
1978. Distribution System Bacteriological Sampling and Control Guidelines, California-
Nevada Section, American Water Works Association, 1st ed.
APHA 1976. Standard Methods for the Examination of Water and Wastewater , (4th ed.,
Washington, D.C.
APHA 1981. Standard Methods for the Examination of Water and Wastewater , 15th ed.,
Washington, D.C.
Clark, J.A. 1968. A presence-absence (P-A) test providing sensitive and inexpensive detection of
coliforms, fecal coliforrns, and fecal streptococci in municipal drinking water supplies.
Canada. J. Microbiol 14:13.
Geldreich, E.E., H.D. Nash, D.J. Reasoner, and R.H. Taylor. 1972. The necessity of controlling
bacterial populations in potable waters; community water supply. J. Amer. Water Works
Assoc. 64:596.
Hoadley, A.W. 1976. Potential health hazards associated with Pseudomonas aeruginosa in water.
In: Amer. Soc. Testing and Materials, A.W. Hoadley, and B. J. Dutka, (eds.), Tech. PubI. 635,
Philadelphia, PA.
Jakubowski, W. and J. C. Hoff (eds.) (979. Waterborne Transmission of Giardiasis,
EPA-60019-79-00l.
LeChevallier, M., R. J. Seidler, and T. M. Evans, 1980. Enumeration and characterization of
standard plate count bacteria in chlorinated and raw water supplies. Appl, Environ.
Microbiol. 40:922.
Marquart, R. 1980. Special report on tap selection and procedures for bacteriological water
sample collection from public water supplies, Lincoln-Lancaster County Health Department,
Lincoln, Nebraska.
Nash, H.D. and E. E. Geldreich, (in press). Effect of storage on coliform detection in potable
water.
National Academy of Sciences 1977. Drinking Water and Health , Washington, D.C.
NATO (in press). Pilot Study on Drinking Water Supply Problems, Committee on the Challenges of
Modern Society.
U.S. EPA 1976. National Interim Primary Drinking Water Regulations, EPA-570/9-76-003.
U.S. EPA 1978. Microbiological Methods for Monitoring the Environment, Water and Wastes, EPA-
600/8-78-017.
WHO (in press). Guidelines for Drinking Water Quality.
-------
TECHNICAL REPORT DATA
(PIe&.e read Inst,vciions on the Pevefle before completing)
f REPORTNO. 2.
EPA 570/9—83—001
3. RECIPIENrS ACCESSION NO.
4. TITLE AND SUBTITLE
Asses nent of Microbiology arid Turbidity Standards
for Drinking Water
5. REPORT DATE
July 1983
I. PERFORMING ORGANIZATION CODE
7. AUTHOR(S
Paul_S._Berger_and_Yeracl2miel_Pzgainan
I. PERFORMING ORGANIZATION REPORT NO.
PERFORMING ORGANIZATION NAME AND ADDRESS
RE Corp.
P .O. Box 40284
Nashville, TN 37204
10. PROGRAM ELEMENT NO.
11.CONTRACT/ORANTNO.
68—03—2917 % D 18
12. SPONSORING AGENCY NAME AND ADDRESS
Criteria and Standards Division
Office of Drinking Water (WH—550)
U.S. EPA, 401 M St., S.W.
Washington,_D.C.__20460
13. TYPE OF REPORT AND PERIOD COVERED
Workshop Proceedings
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
- - seen sew
15. .. .
This document represents the proceedings of a rkshop in dririkinc water microbiology
sponsored in Decenber 1981 by the USEPA Office of Drinking Water (ODW) in conjunction
with the American Society for Microbiology. The purpose was to examine options on
revising the current drinking water microbiology arid turbidity standards, and make
reccrm eridations to O [ . Participants included representatives fran local, State,
Federal, and foreign goverrinent agencies, universities, and industry. Six panels
were convened for t fo11 ing areas of drinking water microbiology: (1) Aaents of
Waterborne Disease, (2) Measur oent of Microbial Water Quality, (3) Monitoring,
(4) Methods, (5) Source, Treatment, arid DistriJ itiOfl, and (6) Canpliance and
Policy Issues. Each panel reviewed arid debated a draft issues paper prepared by the
panel’s chairman. The six issues papers in this document reflect the views,
conclusions, and recaimendations of each panel.
7. KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
C. COSATI Field/Group
Drinking Water
Sanitary Microbiology
Waterborne Disease
Potable Water
Coliforms
Turbidity
S. DI$TRISUTION STATEMENT
Open Distribution
18. SECURITY CLASS (This Report)
Non-sensitive
21. **o) OF PAGES
414
20. SECURITY CLASS (This page)
22. PRICE
EPA F.cm 2220—1 (isv. 4.77) PREVIOUS EDITtON IS OBSOLETE
U.S, GOVIRNMSN I PRINTIN3 OFFICI 1 9 8’. 1+2 1 0 82 S 3 1
------- |