United States i - Office of
Environmental Protection Drinking Water
Agency
EPA-570/9-78-OOC
Water
-------
-------
United States
Environmental Protection
Agency
Official Business
Penalty for Private Use
$300
Fourth-Class Mai!
Postage and Fees Paid
EPA
Permit No. G-35
Washington DC 20460
-------
-------
EVALUATION OF
THE MICROBIOLOGY ISTANDARDS
FOR DRINKING WATER
Edited by
Charles W. Hendricks
Criteria and Standards Division
Office of Drinking!Water
Washington, D.CJ20460
U.S. ENVIRONMENTAL PROTECTION AGENCY
Washington, D.C. 20460 :
This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22151
-------
DISCLAIMER
This manual has been reviewed by the Office of Water and Hazardous Materials,
U.S. Environmental Protection Agency, and approved for publication. The opinions
expressed in the papers are those of the individual authors and do not necessarily
represent the views of the U.S. Environmental Protectioh Agency. Mention of trade
names, commercial products or techniques does not constitute endorsement or
recommendation for use.
-------
PREFACE
Although the quality of America's drinking water is often taken for granted, safe
drinking water is a result of careful utilization of the latest technology and proper
water treatment. In the late 1800's it was discovered that microorganisms responsible
for diseases such as typhoid and cholera could pass through the intestinal tract and
find their way to the water systems. By 1900,30 people per 100,000 population died of
typhoid in the United States. The death rate from typhoid fever was substantially
reduced with the addition of filtration to water supply facilities and many cities had
this equipment on line by 1907. Chlorination was being practiced in most U.S. cities
by 1914, and during the period 1946-1970 waterborne outbreaks of typhoid fever had
dropped to a total of 53.
It is apparent that the incidence of typhoid fever and cholera has been drastically
reduced in the United States, but there is also information to indicate that
waterborne disease warrants further investigation. While there were 53 cases of
typhoid during the years of 1946-1977, there were 297 outbreaks of gastroenteritis of
undetermined etiology. It is also significant to note that 71 percent of these outbreaks
occurred from private water supplies but at least 83 percent of the illnesses were
associated with community water systems.
The above statistics indicate a risk of infectious disease associated with drinking
water, but they do not address whether the water supplied met the 1962 Public Health
Service Drinking Water Standards. It is significant to note that one of the reasons for
the passage of the Safe Drinking Water Act of 1974, was the lack of uniform
application of the 1962 Standards to all public water systems. Another largely
unknown factor is the extent to which the present microbiology standards are
adequate to protect against waterborne disease. Certainly the epidemiological data
supports the adequacy of the standards for protection against cholera and typhoid
but the prevalence of the less severe diseases such as hepatitis, giardiasis and
gastroenteritis suggests that some adjustment of the regulations may be required.
This volume focuses on the present standards for coliform bacteria, coliform bacteria
detection and control, compliance with the coliform standard, and alternative means
of determining compliance, and lays the groundwork for future regulations
development.
Sincere appreciation is acknowledged for the effort of Dr. Riley D. Housewright,
Mr. Edwin E. Geldreich and Dr. S.M. Morrison for aiding in the planning of the
symposium. Special acknowledgments are also given to Cindy Moneymaker and
Lynette Simmons for their secretarial support and to Barbara Menking for her
editorial assistance in preparing this Proceedings.
Charles W. Hendricks
August 1978
ill
-------
-------
FOREWORD
The Safe Drinking Water Act of 1974 provides a unique opportunity for the
federal government to constructively add to the traditional and varied roles of State
and local governments in providing safe drinking water for this nation's public.
Included in the Safe Drinking Water Act are requirements for the promulgation of
Primary Drinking Water Regulations which will apply to all public water supplies.
These regulations are to contain:
1. a list of contaminants that have adverse effects on human health.
2. maximum contaminant levels (MCL's).
3. criteria and procedures to "assure a supply of drinking water which
dependably complies with such maximum contaminant levels.
On December 24, 1975, the U.S. Environmental Protection Agency promulgated
National Interim Primary Drinking Water Regulations. These Regulations replaced
the Public Health Service Drinking Water Standards of 1962 and for the first time
applied maximum contaminant levels for microorganisms and selected chemicals to
all public water supplies serving 25 or more persons. It is estimated that there are at
least 50,000 community and more than 200,000 non-community supplies covered by
the Interim Primary Regulations.
The papers in this Proceedings focus on current microbiological issues and
problems associated with the coliform MCL and the means of determining
compliance. It is anticipated that these discussions, in addition to the National
Academy of Sciences review, will provide the: Environmental Protection Agency with
new insights into the technical problems associated with the MCL's for coliform
bacteria and help to establish a firm basis for their revision.
Joseph A. Cotruvo
Director
Criteria and Standards Division
Office of Drinking Water (WH-550)
U.S. Environmental Protection Agency
-------
-------
TABLE OF CONTENTS
PREFACE
FOREWORD
WELCOME Dr. Joseph A. Cotruvo, Director, Criteria and Stand-
ards Division, U.S. Environmental Protection Agency,
Washington, D.C. \
SESSION!
EVALUATION OF THE COLIFORM STANDARD
Chairman: Dr. Riley D. Housewright, National Academy of Sciences, Wash-
ington, D.C. '
Evaluating the microbial quality of potable waters 3
M.J. Allen and E.E. Geldreich
Interferences to coliform detection 13
E.E. Geldreich, M.J. Allen and R.H. Taylor ' ",
Impact of the coliform standard on the transmission of disease 21
G.F. Craun '•
Alternative indicators of water contamination and some physiological
characteristics of heterotrophic bacteria in water ., 37
G.A. McFeters, J.E. Shillinger and D.G. Stuart!
Some statistical considerations in water quality control.,..'.... 49
L.R. Meunz
Chlorine residual substitution—rationale .... 57
L.J. McCabe
Public health considerations of the microbiology of "potable" water 65
R.R. Colwell, B. Austin and L. Wan
SESSION! '••'- ;
COLIFORM DETECTION AND CONTROL
Chairman: Mr. Edwin E. Geldreich, U.S. Environmental Protection Agency,
Cincinnati, Ohio , .
Standard sample collection and preservation ..; 79
H.D. Nash !
The EPA methods manual and the microbiological standards 85
R.H. Bordner and J.A. Winter
Evaluation of standard procedures—Standard Methods 91
A.E. Greenberg i
95
Virus detection
D.O. Cliver
Disinfectant resistant organisms
101
P.T.B. Shaffer
vii
-------
The relationship of turbidity to disinfection of potable water ..... ........ 103
J.C. Hoff
Wastewater reuse and the problems of waterborne disease
H.W. Wolf
Evaluation of microbiological standards for drinking water— engineering
control practices
P.O. Haney
119
SESSION 3
COMPLIANCE WITH THE COLIFORM STANDARD
Chairman: Dr. Charles W. Hendricks, U.S. Environmental Protection Agency,
Washington, D.Q
Compliance with drinking water regulations • 145
C.W. Hendricks
Experiences with the coliform standard under the interstate program 153
F. Taylor
Experiences with the coliform standard 159
J. Harrison
Implementation of the coliform standard by a community water utility 169
R.J. Becker
Compliance with PL 93-523 coliform standards 181
H.J. Ongerth
Canadian drinking water standards •
T.P. Subrahmanyan, D.A. Schiemann, A.J. Rhodes, J. Henderson, E. Bowmer,
fc.R. Rozee and P. Payment
1 !
PANEL DISCUSSION
ALTERNATIVE MEANS OF DETERMINING COMPLIANCE
Chairman: Dr. Sumner M. Morrison, Department of Microbiology, Colorado
State University, Fort Collins, Colorado
Alternative means of determining compliance ......... ................. 197
S.M. Morrison
The coliform MCL— Can we defend it? ............................ • • • 199
W. Litsky
Alternative means of determining compliance ......... ....... • .......... 205
B.E. Gay
Alternative means of determining compliance ......... ............ ...... 211
I. Markwood
Alternative means of determining compliance .......................... 215
W.O. Pipes
Alternative means of measuring compliance with bacteriologic standards of
public water supplies ............................... ................. 219
G.W. Fuhs
VUl
-------
LIST OF ATTENDEES 229
INDEX i 233
IX
-------
-------
. . ..WELCOME . ,
Dr. Joseph A. Cotruvo, Director
Criteria and Standards Division '
Office of Drinking Water :
U.S. Environmental Protection Agency :
Washington, D.C. 20460 :.
I want to extend a sincere welcome to all of you in attendance at today's
symposium entitled "Evaluation of the Microbiology Standards for Drinking
Water". We are fortunate to have many of the nation's leading experts in water
supply who will address both the technical and regulatory aspects:of the current
regulations. I would invite all of you to participate in the discussions to the maximum
extent possible and everyone here should iind something of interest.
The symppsium is sponsored by the Criteria and Standards Division, Office of
Water Supply and is intended to gather information in support of future regulations
development. The Safe Drinking Water Act (PL-93-523) requires EPA to revise the
Interim Primary Regulations within 90 days after the receipt of the National
Academy of Science review. The presentations and discussions here will add to our
understanding of the problems and issues and, together with the National Academy
of Sciences review, will enable us to make the necessary revisions to the regulations.
Despite the advanced technology and management practices applied to water
treatment, waterborne disease outbreaks still occur and a significant portion of the
time we cannot determine what etiological agent is involved. Most of the outbreaks
and cases of illness resulted from contamination of groundwater and deficiencies in
treatment but many of the outbreaks (31 %) were due to inadequacies or interruptions
in the chlorination process. This indicates that we must constantly improve the
operation of existing treatment and monitoring processes as well as consider new
analytical methods and techniques.
The preamble of the Interim Primary Regulations addressed three specific areas
where changes in the primary regulations may be desirable. The month'ly average by
which compliance is determined, is affected by the upper limit of coliform bacteria
when the sample contains coliform bacteria too numerous to count. For example,
some laboratories assume an upper limit of 50 while other laboratories seek to count
all the colonies on the plate. There is also a problem in dealing with spurious positive
samples. All samples taken to determine compliance must be counted because a
negative check sample taken days after a positive sample does not mean that the
original positive result was in error. The third aspect to be resolved concerns relating
monthly averages of coliform bacteria to monthly percentages of positive samples.
Each routine sample taken during the month may be in compliance with the limit and
a system could still fail the monthly average maximum contaminant level. I would
anticipate, in your discussions over the next two days, each of these questions could
be discussed in detail as to provide the Criteria and Standards Division with a means
of dealing with them without compromising the integrity of the present maximum
contaminant levels.
To help facilitate answering these and other compelling questions, the symposium
is divided into three sessions. Session I-Evaluation of the Coliform Standard, will
have Dr. Riley D. Housewright from the National Academy of Sciences, Safe
Drinking Water Committee, to act as chairman. His session will focus on microbial
water quality, alternative indicators, and the problems encountered with testing
procedures and treatment. Session 2-Coliform Detection and Control, chaired by
Mr. Edwin E. Geldreich, Municipal and Environmental Research Laboratory,
Cincinnati, Ohio; will deal with the problems of the growth and detection of coliform
bacteria and other indicator organisms. In addition, viruses, turbidity and water
reuse will be discussed. Dr. Charles W. Hendricks from my office will chair Session 3-
xi
-------
Compliance with Coliform Standards. This discussion includes the compliance
aspects of the National Interim Primary Regulations and specific changes that should
be made in the primary regulations. A panel discussion to be conducted by Dr.
Sumner M. Morrisdn from Colorado State University will consider alternative
means of determining compliance with bacteriological standards.
I thank all of you for attending and hope this symposium will be beneficial for all
concerned as we work toward providing safe drinking water.
Xll
-------
Fly-Away Virus Concentrator (US Environmental Protection Agency)
SESSION 1
Evaluation of the Coliform Standard
Chairman: Dr. Riley D. Housewright, Project Director, Safe Drinking Water
Committee, National Academy of Sciences, Washington, D.C.
-------
,J ' II'
„ j; . {;
-------
MICROBIAL QUALITY OF WATER/M. J. ALLEN
Evaluating the Microblal Quality of Potable Waters
Martin J. Allen and Edwin E. Geldreich, Jr.
Water Supply Research Laboratory/ MERL
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
Bacterial Indicator Concepts
Coliform bacteria have been the bacteriological tool used to measure the
occurrence and intensity of fecal contamination in drinking waters for nearly 70 years
(11,13). During this period, a mass of microbiological data has been developed to
permit the conclusion that the absence of total coliforms in a potable water supply is
evidence of a safe supply. Although there are some differences between various
coliform strains with regard to natural survival and perhaps resistance to
disinfectants, these are minor biological variations that are more clearly demonstrat-
ed in the laboratory than in a water treatment system. The presence of any coliform
bacteria, fecal or nonfecal, in treated water should not be tolerated, and when found,
suggests either inadequate treatment or postchlorination contamination in the
distribution system (5,15). For these reasons it is not surprising that the total coliform
indicator system is part of this nation's drinking water standards.
The coliform group of bacteria was first adopted in 1914 by the U.S. Public Health
Service pursuant to the Interstate Quarantine Regulations for common carriers (21).
Since the inclusion of this bacterial quality parameter for potable waters, the
coliform concept has remained essentially unchanged in successive Public Health
Service standards of water quality, i.e. 1925, 1942, 1946. During this time there have
been a few minor refinements/clarifications in the multiple tube procedure. The
development and acceptance of the membrane filter analytical method in the 1962
Public Health Service Standards added an alternate procedure, but since then no
further advances have been made in methodology. Most recently, with the
promulgation of the National Interim Primary Drinking Water Regulations in 1975,
the total coliform group of bacteria was again selected as the indicator of sanitary
significance for evaluating the microbial quality of drinking water (20). In view of
advances in methodologies for detection of enteric viruses and bacterial pathogens,
the infallibility of the coliform standard has been occasionally 'questioned (18)
although it continues to be the basic biological test of water potability.
To enumerate all waterborne pathogenic organisms, the microbiologist would
have to perform a variety of complex, time-consuming, costly, and often tentative
procedures for each sample analyzed. A more realistic approach is to use»a bacterial
indicator system which will detect and quantify fecal pollution from all warm-
blooded animals. Insofar as bacterial pathogens are concerned, the total coliform
group of organisms, which comprise only a fraction of the number and types of
bacteria present in feces (12), is considered a reliable indicator of the adequacy of
water treatment. From a practical standpoint, the standard total coliform methods
(MPN, MF) offer the finished water laboratory a microbial test parameter which is
easy to perform, requires little time, modest equipment expenditures, is relatively
inexpensive, and provides results in 24 hours for the MF or within 96 hours when
using the MPN procedure. Both the MPN and MF protocols can be easily.learned
and performed by most water treatment plant personnel. :
While the coliform standard and its concomitant methodology are the preferred
microbial surveillance tool, there have been instances where such monitoring has not
signaled the presence of waterborne Giardia cysts. Newly reported data on non-
filtered surface water systems note that: the removal of coliforms during the
disinfection process may not be an effective measure of cyst inactivation (9). Data
-------
I I
II
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS*
10
Tl 1
COXSACKIEViRUS A2
M. FORTUITUM
COXSACKIE-
VIRUS \
A9
C. PARAPSILOSIS
ADENOVIRUS
TYPES
0.01
10 100
CONTACT TIME (MIN.)
1000
Figure 1. FREE CHLORINE RESIDUALS AND CONTACT TIMES NECESSARY FOR
99.9% INACTIVATION FOR SELECTED VIRUSES, BACTERIA, AND YEAST (2, 10).
collected to correlate virus occurrences with total coliforms in finished water support
the continued use of the coliform group as the sanitary standard.
The Relative Resistance of Bacteria and Viruses
There is a considerable amount
of information available on the relative resistance
of bacteria and viruses to chemical disinfectants. Although a number of viruses may
occur in polluted surface waters. I.e. infectious hepatitis virus, poliovirus, reovirus,
adenovirus, echovirus, coxsackievirus and bacteriophages, disinfection data are
limited principally to the enterip viral groups of polio and coxsackievirus (16).
Because chlorine is widely used as'the disinfectant of choice in most water treatment
plants, determination of the relatij/e resistance of these enteric viruses to free chlorine
and chloramines is of obvious practical importance. Ideally, it is desirable that the
general resistance of all enteric viniises to chlorination would be of the same order and
be similar to inactivation rates ofjthe indicator coliform bacteria. The literature on
studies using a variety of microorganisms and species of chlorine/chloramines show
that there are measurable differences in susceptibility to disinfectants (Figure 1).
While some enteric viruses, bactefia(Mycobacteriumfortuitum, M. phlei), and yeast
(Candida parapsilosis) are mo{re resistant that coliforms to inactivation by
conventional disinfectants (10), current recommended water treatment practices use
-------
MICROBIAL QUALITY OF WATER/M. J. ALLEN
i
i
chlorine concentrations greater than 1 mg/1 w;ith protracted contact periods to
increase the likelihood of total inactivation of viral and bacterial pathogens.
The Significance of Virus-rColiform Ratios ;
The primary objective in the bacteriological examination of water is the detection
of treatment deficiencies which can result in the eintry of pathogens into the finished
water supply. However, the occurrence and density of viral and bacterial pathogens
in polluted waters is highly variable. This variability reflects the intestinal diseases
that are prevalent at a given time in the human or <|>ther animal subpopulations which
contribute their wastes to a particular effluent or watershed, as well as the numerous
physical and chemical conditions which aifect microbial virulence and survival.
The relative number of coliforms and viruses in'feces, sewage, and polluted surface
waters have been considered by some to be important in evaluating the waterborne
disease potential and validity of the coliform standard (6). While coliforms are
discharged in the feces of nearly everyone:, viral excretion is generally limited to
children under the age of 15 years with approximately 10% of these children shedding
viruses at any given time. About 30% of the population falls into this age group, and
calculations based on a per capita daily discharge of virus and coliforms result in
estimated virus-coliform ratios of 1:65,000 foij feces, 1:92,000 for sewage, and
1:50,000 in polluted surface water (1); studies: made on Missouri River waters
indicate a ratio of 1:500,000 (7). First year resu
ts of an in-house study (Brashear,
unpublished) of polluted surface water have shown that even with improved viral
methodology, no apparent relationship exists be'
tween the number of total or fecal
coliforms in water and the number of viruses (Table 1). Viral isolates which included
poliovirus 1,2,3, coxsackievirus B2, B4, and echovirus 1, 4, 25, were obtained from
sample volumes of approximately 380 liters (1QO gallons); virus to, fecal coliform
ratios ranged between 1:23 to 1:11,000. Although these ratios are considerably lower
than previously reported, presumably due to improved recovery methodology, the
coliform bacteria obviously greatly outnumber enteric viruses in sewage and polluted
surface waters and thus still appear to be a more sensitive indicator of possible
pathogen occurrences. i
An important point which should be noted with respect to virus-coliform ratios is
relative sample volumes. Traditionally, coliform densities in waters have been
expressed as organisms per 100 ml or the approximate volume of a glass of water
(Butterfield, personal communication). The standard drinking water sample volume
for membrane filtration analysis was recently raispd from 50 ml to 100 ml. Byway of
contrast, the methodology now being developed for isolating viruses from water
(potable, surface, ground) processes from 1 to 1,9pO liters, and typical finished water
sample volumes range from 380-1,900 liters. For equivalent density comparisons
between viruses and coliforms, it is necessary to analyze comparable water volumes
of a liter or more for each microorganism.
Efficacy of Water Treatment Processes
A recent field study reported by Akin et al. (1) examined finished water quality of
six municipal treatment plants in three states. The treatment plants use surface
waters as their principal raw source water and rely on conventional treatment
practices, i.e. flocculation, sedimentation, filtration, and chlorination (Table 2).
Bacteriological quality of the source waters was variable, and at times exceeded the
recommended maximum fecal coliform density
(2000/100 ml) for surface water
sources (19). Although the primary objective of this study was to determine whether
human enteric viruses could be detected in finished waters using the equipment
/procedures currently under development, parallel bacteriological analyses
-------
o
§
w
HM
O
TABLE 1. MICROBIOLOGICAL QUALITY OF OHIO RIVER WATER FROM CINCINNATI AREA*. £
Location1
Schmidt (467)3
Southside (475)3
Schmidt
Schmidt
Public Landing (470)3
Schmidt
Public Landing
Schmidt
Public Landing
Southside
Public Landing
Total Coliforms
(per 100 ml);
11,000
49,000
85,000
38,000
3,400
105,000
35,000
99,000
48,000
135,000
7,300
Fecal Coliforms
(per 100 ml)
6,500
17,000
23,000
11,000
370
43,000
3,500
42,000
7,200
33,000
1,200
.Standard Plate
Count per ml
37,000
62,000
79,000
48,000
20,000
162,000
112,000
193,000
46,000
79,000
8,300
Virus2
• Isolates
0
3
9
1
16
17
1
31
12
59
3
* From Brashear
1 Sampling period 9/8/75 through 12/1/75
2 From 100 gallon (380 liters) water samples
3 Ohio River Mile
i
i
B
-------
TABLE 2. PHYSICAL/BACTERIOLOGICAL SOURCE WATER CHARACTERISTICS AT
WATER TREATMENT PLANTS*.
Location
Source Water(s) Capacity/mean Output Fecal Coliforms F.C.
(10,000 mV day) (MPN/lOOml) Geometric Mean
Columbus, OH
Sidney, UH
Muncie, IN
Seymour, IN
Kansas City, MO
St. Joseph, MO
Scioto River :
Tawana Creek, wells :
White River j
E. Fork White River ;
Missouri River .j
Missouri River ! '•
26.5 / 13.6
1.4 / I.I
6.0 / 4.9
0.56/ 0.45
79.5 / 41.5
11 / 5.3
38-3,000
40-2,800
50-2,000
1 1-14,000
1,700-4,300
280-7,000
280
263
210
536
3,010
1,890
s
b
69
I
*,Frpm Akin el al. (1)
o
•n
$
1
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
TABLES. TOTAL COLIFORM/FECAL COLIFORMS RESULTS OF
FINISHED WATfR SAMPLES FROM SIX TREATMENT
.PLANTS*.
Treatment Plant
Location
Columbus, OH
Sidney, OH
Muncie, IN
Seymour, IN
Kansas City, MO
St. Joseph, MO
Membrane Fill
Total
Coliforms
0/4*
1/3
2/4
3/11
0/2
1/1
5r° Modified MPN* Large Volume Sampler'
Fecal Total
Colifcjrms Coliforms
o/
I/
o/
o/
o/
o/
I 2/4
3 2/3
4 0/4
11 5/11
2 2/2
1 0/1
Fecal Total Fecal
Coliforms Coliforms Coliforms
0/4 3/4 0/4
1/3 0/3 0/0
0/4 3/4 2/4
1/11 10/11 6/11
0/0 0/0 0/0
o/i . - -
* From Akin et al. (1)
" Up to 5 liters tested; negative results indicate a value <0.02/100 ml; the largest value obtained
was 0.12 organisms/100 ml. |
* 5.5 liters tested; negative results indicz te a value <0.02/100 ml; the largest value obtained was
0.23 organisms/100 ml.
c 380 liters tested; negative results indicate a value <0.00059 organisms/100 ml; the largest
value obtained was >0.0043/100 ml.
J Occurrences/Number of Samples
involving large volume sampling protocols (380 liters) were performed to establish
coliform densities in finished waters.
A summary of the bacteriological findings is presented in Table 3. Using the
membrane filter procedure with up to 5 liters of finished water, coliforms were
detected in seven samples; one sample also contained fecal coliforms. The modified
MPN procedure (5.5 liter samples1) found coliforms in 11 samples, 2 of which also
contained fecal coliforms. Finally], the large volume sampling method (380 liters)
found coliforms in 17 samples, 8 of which also contained fecal coliforms. In all cases,
coliform densities were below the maximum contaminant level (MCL) set by
NIPDWR. More importantly, inJ56 finished water samples in which the average
volume of water analyzed was 1,100 liters, no viruses were isolated. Previous
laboratory evaluations of the sensitivity of the virus recovery procedure indicated a
detection level of 3-5 units (poliovirusjper 380 liters when 1,900 liters were sampled
(14). In contrast, the large volume sampling protocol for coliforms was able to detect
organisms using only 380 liter sample volumes.
Thus, the results of this study support the adequacy of conventional treatment
processes for removal of bacterial dnd viral pathogens from even the poorest quality
source waters (2,3,4)^ These data show that at least 99.9999% coliform reduction is
readily attainable in water treatment facilities with complete treatment; this percent
reduction is based on data in Table 3 where a negative large volume sampling result
from a highly polluted source water indicates a coliform density of < 0.00059/100
ml. Filtration, particularly when preceded by chemical floccuiation, is very effective
in removing microorganisms (17).
utilize rapid sand filtration prior
While the majority of water treatment facilities
to chlorination and distribution, a number of
utilities lack filtration capabilities and rely solely on chemical disinfection to
eliminate microbial pathogens. Such treatment deficiencies have been associated
with increased outbreaks of giardiasis (8).
-------
MICROBIAli QUALITY OF WATER/M.J. ALLEN
Epidemiologica! Findings
Twelve outbreaks of giardiasis including 4,860 cases in Rome, New York, were
documented during the period 1971-1974 (8,9). all of these outbreaks, with the
exception of Aspen, Colorado, involved municipal systems which used surface water
supplies that were only chlorinated prior to 'distribution. Giardia, a flagellated
protozoan, forms cysts which are not inactivated by chlorination at the dosages and
contact times normally employed in water treatment processes; whether cdliforms
were present or absent from systems associated with the outbreaks has not been
established. Although high chlorine concentrations and protracted contact periods
will inactivate the cysts, it is recommended that nitration be implemented as part of
the complete treatment of all surface waters. It is possible that with unfiltered surface
waters, the coliform indicator system may not b e adequate to signal the presence of
Giardia or other pathogenic cyst forming protozoans. Since the analysis for this
organism is time-intensive, complex, and imprecise, the logical approach is the
elimination of the problem through improved water treatment practices, especially
for finished water systems which have had a previous (chronic or intermittent) history
of waterborne disease outbreaks attributed to trpatment deficiencies. Regarding the
etiology of waterborne outbreaks of gastroenteritis, hepatitis-A, shigellosis, typhoid,
and salmonellosis, Craun et al. (9) concluded that treatment deficiencies, such as
inadequate disinfection, interruption of disinfection or lack of treatment of ground
waters were the primary causes.
Conclusions
With the implementation of the National Interim Primary Drinking. Water
Regulations, the coliform group of microorganisms remains the most relevant
sanitary index for finished waters. Recent findings confirm earlier work that the
coliform indicator concept is adequate in assuring bacterial and viral safety of water.
The inability of present virological methodology to isolate viruses, from finished
waters results in continuation of our reliance on and confidence in the coliform group
for monitoring finished water quality.
Although surface water quality may vary widely
through protection of receiving waterways, pres
and must be continually improved
ent water treatment practices are
capable of consistently producing pathogen-free water to the consumer. The .critical
biological safeguards in water treatment continue to be flocculation/filtration with
post-chlorination. Excluding protozoan cysts, all bacterial and viral pathogens are
readily inactivated by chlorine dosage and contact times routinely used in properly
operating water treatment facilities. The key to eliminating waterborne disease
outbreaks, however, is constant surveillance of finished waters and their distribution
network and the maximum use of good treatment processes. Treatment plants with
any deficiencies must be upgraded to remedy these inadequacies and assure
production of the best quality potable water.
REFERENCES
1. Akin, E.W., D.A. Brashear, E.G. Lippy, and N.A. Clarke. 1975. A virus-in-water study of
finished water from six communities. Health Ejects Research Laboratory, Office of
Research and Development, U.S. Environmental Protection Agency, Cincinnati, Ohio.
2. Berg, G. 1966. Virus transmission by the water vet icle: III. Removal of viruses by water
treatment processes. Health Lab. Sci. 5:170-181.
3. Chang, S.L., R.E. Stevenson, A.R. Bryant, R.L. Woodward, and P.W". Kabler. 1958.
Removal of Coxsackievirus and bacterial virus in water by flocculation: II. Removal of
Coxsackievirus and bacterial viruses and native bacteria in raw Ohio River water by
flocculation with aluminum sulfate and ferric chlorate. Am. J. Public Health 45:159-169.
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
4. Chaudhuri, M., and R.S. Engelbrecht. 1970. Removal of viruses from water by chemical
coagulation and flocculation. JAWWA 62:563-567.
5. Clark, H.F., and P.W. Kabler. 1964. Reevaluation of the significance of coliform bacteria.
JAWWA 55:931-936.
6. Clarke, N.A., G. Berg, P.W. Kabler, and S.L. Chang. 1964. Human enteric viruses m water:
source, survival, and removability. International Conference on Water Pollution Research,
Pergamon Press, Oxford.
7. Clarke, N. A., H.D. Nash. 1972. Equating bacterial standards with viral numbers in potable
waters. Proceedings Annual Meeting of the American Society for Microbiology,
Philadelphia, Pennsylvania.
8. Craun, G.F. 1976. Microbiology-waterborne outbreaks. J. Water Pollut. Control Fed.
45:1378-1397. . •
9. Craun, G.F., L.J. McCabe, and J.M. Huges. 1976. Waterborne disease outbreaks m the
U.S. (1971-1974). JAWWA 55:420-424. ,
10. Engelbrecht, R.S., D.H. Foster, E.O. Greening, and S.H. Lee. 1974. New microbial
indicators of wastewater chlorination efficiency. Environmental Protection Technology
Series EPA-670/2-73-082. U.S. Environmental Protection Agency, Washington, D.C.
11. Geldreich, E.E. 1966. Sanitary significance of fecal coliforms in the environment. Water
Pollution Research Service Publication No. WP-20-3. U.S. Department of Interior, Wash-
ington, D.C.
12. Geldreich, E.E. 1977. Bacterial populations and indicator concepts in feces, sewage,
stormwater, and solid waste. In: Berg, G. (Ed.), Indicators of viruses in water and food.
Publication, Ann Arbor, Michigan.
13. Geldreich, E.E., and N.A. Clarke. 1972. The coliform test: a criterion for viral safety of
water. Proceedings 13th Annual Sanitary Engineering Conference, Urbana, Illinois.
14. Hill, W.F., W. Jakubowski, E.W. Akin, and N.A. Clarke. 1975. Detection of viruses in
drinking water: sensitivity of the tentative standard method. Annual Meeting American
Society Microbiology Proceedings, New York.
15. Kabler, P.W. and H.F. Clark. 1960. Cqliform group and fecal coliform organisms as
indicator of pollution in drinking water? JAWWA 52:1577-1579.
16. Kabler, P.W., N.A. Clarke, G. Berg, and S.L. Chang. 1961. Viricidal efficiency of
disinfectants in water. Public Health Reports 75:565-570.
17. Robeck, G.G., N.A. Clarke, and K.A. Dostal. 1962. Effectiveness of water treatment
processes in virus removal. JAWWA 54:1275-1292.
18. Shuval, H.I. 1976. Water need and usage, pp. 12-25. In: Berg, G. et al (Eds.), Viruses in
water. American Public Health Association, Washington, D.C.
19. The Committee on Water Criteria. 1972. Water quality criteria-1972. Environmental
Studies Board, National Academy of Sciences, National Academy of Engineering, Wash-
ington, D.C.
20» U.S. Environmental Protection Agency. 1975. National Interim Primary Drinking Water
Regulations. Federal Register 40:59566-59574.
21. U.S. Public Health Service. 1914. Bacteriological standard for drinking water. Public
Health Reports 29:2959-2966.
QUESTION AND ANSWER SESSION. ' .
i '
G.W. Fuhs, Division of Laboratories & Research, New York Department of Health,
Albany, New York
Is a standard plate count on groundwater more sensitive to contamination which
may have occurred one week previous to sampling?
M.J. Allen, Water Supply Research Laboratory, U.S. Environmental Protection
Agency, Cincinnati, Ohio
Waters from undisturbed or protected aquifers, unlike those from surface sources,
have low indigenous bacterial populations. However, in shallow aquifers where
percolating water from surface sources can provide sufficient nutrients to support
microbial growth, an unusually high bacterialpopulation could indicate infiltration
of contaminated water into the aquifer. More importantly, abnormally high
10
-------
MICROBIAL QUALITY OF WATER/M J. ALLEN
microbial levels in shallow aquifers can adversely affect detection of the more specific
indicators. Thus, coliforms could be undetectable in ground water where excessive
bacterial densities occur. The failure to detect coliform bacteria results, unfortunate-
ly, in the assumption that these contaminated supplies are safe to drink. Therefore, it
could be necessary to employ additional micirobial indicator groups for ground water.
These indicator organisms may supplement or replace the coliform group when
ground water conditions are such that erroneous results could occur.
11
-------
,,|' Ji, HI,,!I illli,,!,
-------
INTERFERENCES TO COLIFORM DETECIjION/E.E. GELDREICH
Interferences to Coliform Detection i n
Potable Water Supplies
Edwin E. Geldreich, Martin J. Allen and Raymond! H. Taylor
Water Supply Research Laboratory/ MElf L !
U.S. Environmental Protection Agency)
Cincinnati, Ohio 45268
The monitoring of potable water supplies must always jitilize. bacteriological
procedures that are as sensitive for coliform detection as the stjite-of-the-art permits.
This goal is of particular importance in a. continuing surveiljance for low levels' of
coliform bacteria that signal water treatment or transmission deficiencies which
could allow the breakthrough of waterbotne pathogens. White there has been much
effort directed towards improving methodology, standardization of testing
procedures, and achieving greater levels of data reliability through quality control
and laboratory certification programs, little attention has I been given to those
peripheral water characteristics that may impact on the detection of indicator
organisms. This concern has become more critical with the change of the official MF
test volume from 50 ml to 100 ml in the Drinking Water Standards, in order to obtain
better statistical reliability at the baseline limit of one coliform per 100 ml. These
interferences relate to the nature and density of potable water turbidity or to the
composition and magnitude of the finished water microbial flora.'
Turbidity in Potable Water j ] .
Turbidity in potable water may be derived.from source Water, water treatment
practices, problems in the distribution system, or a combination of these factors.
Thus, turbidity may range from algal growth, natural siltation
industrial waste discharges to surface v/aters, chemicals hi
filtration practices, residual coagulants in water treatment,
, decaying vegetation,
ground water, poor
corrosion, microbial
growth or soil intrusions in wlter disfiribution systems. As a result.of these
conditions, a variety of materials may be represented in turbidjity, some of which can
impart a chlorine demand. Other turbidities may be sources of interference to
coliform detection, and only a few kinds of material will bje of little concern as
protective habitats for microbial survival.
of treatment barriers.
ral mineral turbidity,
\J i,\j\^l,L V W 4AC4-^/Jll.C*.l.O ± Vy » ii«.*v* •*-• v Jtwi wv--. ••*—••-•
Turbidity can be a transport vehicle for bacterial penetratior
While sand particles may be of little concern, other'natu
hydrated oxides from flocculation, and organic debris have bleen shown to provide
absorption sites for bacteria (16,17,18,22),, These conditions are of major concern in
those supplies that employ disinfection as the sole treatment rirocess for raw surface
waters (22). Without use of more extensive treatment processes, the barrier to
microbial penetration becomes inadequate because of turbidity fluctuations caused
by stormwater runoff, algal blooms, seasonal turnover of impounded waters and
water uses upstream. Where water treatment processes are poojrly controlled, natural
turbidity, alum floe particles or. cationic polymers used in coagulation may break
through filter beds into the distribution system. These conditions will produce
turbidity levels higher than 1 NTU and may bring associated bacteria into the potable
water. . '
Turbidity Interferences i
Suspended material (turbidity) in a potable water sample niay preclude use of the
membrane filter coliform procedure. Reasons for rejecting thi MF procedure relate
to the volume of samples that can be filtered, character of suspended material, and
the thickness of the suspended material that deposits on the membrane filter surface.
13
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
]
| - -, ' '
Relatively thin layers of gelatinous, finely divided, or hygroscopic materials, such as
suspended ferrous, manganese, and alum floes or algal cell populations may clog
filter pores or cause a confluent film of growth to develop during incubation. Thick
surface layers of crystalline or siliceous materials may cause little or no difficulty.
There is no practical method of removing turbidity without also removing some
unpredictable portion of the bacterial population including coliform bacteria in the
sample. Our experience has indicated that pre-filtering turbid water through a coarse
filter or decanting the supernatant from settled turbidity in a water sample results in a
20 to 80 percent loss of organisms. Turbidity limits for membrane filtration have
never been determined because the limits vary with the type of material (sand VS.
clay), chemical composition (phosphates, coagulants) and presence of various algal
masses. While the multiple tube method may be the method of choice when testing
potable water samples that contain turbidity interferences to membrane filtration,
this approach may have some undesirable consequences. Selection of the multiple
tube method automatically limits routine analyses to a 50 ml.test portion, rather than
the 100 ml test portion used with the membrane filter procedure. Thus, the theoretical
detection limit for total cqliforms in potable water becomes 2 organisms per 100 ml
and assumes that these organisms are uniformly distributed in two 30 ml test
portions. However, uniformity of bacterial suspension may not be achieved in turbid
samples. Adsorption pf bagteria to turbidity particles and settling of heavier particles
are factoFs that must be rectified since the random occurrenc.es of only three
coliforms in the 50 ml test portion is sufficient to produce an unsatisfactory result,
Coliforms trapped in turbidity particles may not produce gas in the presumptive
medium unless released by vigorous sample agitation, This condition may be the
cause of occasional differences in results from duplicate samples analyzed at the local
water plant laboratory and at the State laboratory after indeterminant intervals of
agitation during transit,
: !
: ! !
Microbial Interferences
: i
While water produced at the treatment plant may be of high bagterjqlogiqa!
quality, finished water entering a distribution network is subject to continuous
bacteriological deterioration (1,4,7,8,10,12,19,20). Some factors which contribute to
the deterioration include: 1) inadequate dosage or loss of disinfectant residual; 2)
incomplete removal of organics and/or turbidity due to treatment deficiencies; 3)
seasonal temperature fluctuations; 4) extended retention periods in reservoirs,
standpipes, and dead-end water mains; 5) sediment accumulations; and 6) line breaks
and cross-connection occurrences in the distribution network. Although the
microbiological flora in finished waters is highly variable, bacterial groups most
frequently associated with finished water deterioration include: Pseudomonas,
Flavobacterium, Achromobacter, Proteus, Klebsiella, Bacillus, Serratia, Coryne-
bacterium, Spirillum, Clostridium, Gallionella.-Arthrobacter, and Leptothrix (6),
These organisms, in general, are able to survive and many may even multiply in
finished waters. Such bacterial regrowth may result in slime deposits, tubercle
formation, development of taste and odor problems, accelerated pipe deterioration
(6,12,20,21) increased health risk (14) and may also interfere with bacteriological
monitoring procedures.
Although finished waters which have been shown to be CQljform free are generally
considered safe for human consumption, there are occasions when the absence of
detectable coliforms cannot be equated with potability. Excessive densities of non-
coliform 'organisms in finished water may desensitize assay procedures for total
coliforms (5,9,11,23). These non-coliform organisms can coexist with coliforms in
water, but when introduced into lactose broth, many of the non-coliform organisms
may multiply at a rapid rate, thus intensifying the factor of coliform inhibition (15) or
simply outgrow a smaller coliform population.
14
-------
INTERFERENCES TO COLIFORM DETECTION/E.E. GELDREICH
In laboratory tests conducted by Hutchison et al (17), suspensions of various
antagonistic organisms (Pseudomonas, Actinomyces, Sarcina, and 'Micrococcus) in
a density range of 1,000 to 2,000 per ml, when added to lactose tubes simultaneously
to a suspension containing 10 E. coli per ml, resulted in reduced coliform detection.
The loss of sensitivity ranged from 28 to 97 percent, depending upon the combination
of mixed strains in these experiments. Considering that the baseline for coliform
detection is one organism per 100 ml, any loss of test sensitivity becomes critical.
Analysis of bacteriological data from survey of distribution water samples from
969 public water plants in this nation suggested that interference with coliform
detection may be a problem (13). These data (Table 1) showed that the frequency of
detecting total and fecal coliforms by the MF procedure, increased ,as the standard
plate count increased to levels up to 500 per ml, but decreased in frequency when the
non-coliform bacterial population exceeded 1,000 organisms per ml. These findings
again suggest that high densities of non-coliform bacterial populations have an
adverse affect on detection of the coliform indicator.
While provisions for a Standard Plate Count measurement are not currently
included in the National Interim Primary Drinking Water Regulations, field studies
in several different geographical areas are now in progress to obtain additional
information on the impact of the total microbial populations on MPN and MF
methodology. One of these studies involves the Cincinnati water distribution system.
This public water supply network, like many other systems, has numerous dead-end
water mains. The water in these low flow mains is subject to abnormal deterioration,
as demonstrated by the number of customer complaints in the areas, of a number of
these low flow sections. For this reason, the Water Distribution Division-
Maintenance Section of the Cincinnati Water W.orks has maintained a routine
flushing program for at least 30 years.
Weekly, a number of troublesome dead-end water mains are flushed to clear out
accumulated sediments and bring fresher waters with free chlorine residuals into
these mains. During a two year period, with the cooperation of the Maintenance
Section personnel, samples from these, flushes were ^obtained for'bacteriological
analysis. All samples were iced immediately and processed within five hours of
collection. Bacteriological analyses of 613 water samples from 32 sampling sites
included the total coliform multiple tube and membrane filter procedures and a
TABLE 1. BACTERIAL PLATE COUNT VS. COLIFORM DETECTION IN
DISTRIBUTION WATER NETWORKS FOR 969 PUBLIC
WATER SUPPLIES.
General Bacterial Population* Total Coliform Fecal Coliform
Density Range Number of
per 1 ml Samples Occurrences Percent Occurrences Percent
<1
11
31
101
301
501
>1
- 10
-30
- 100
-300
-500
- 1,000
,000
1013
371
396
272
120
110
164
47
28
72
48
30
21
31
4.6
7.5
18.2
17.6
25.0
19.1
18.9
22
12 :
28
20
11
9
5
2.2
3.2
7.1
7.4
9.2
8.2
3.0
TOTAL 2446 277 — 107
'* Standard Plate Count (48 hrs. incubation, 35° C)
15
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
standard plate count (SPC agar pour plates, incubated at 35° C for 48 hours).
Physical/chemical parameters measured were free chlorine residual (DPD test),
turbidity, water temperature and pH.
Bacteriological analysis of water samples obtained from dead-end water mains
(Table 2) showed that 76 of 613 samples contained coliforms when examined by the
MPN procedure. In contrast, the membrane nitration technique detected coliforms
in only 19 of these samples. Thus, the MF method detected coliforms in only 25% of
the samples that were positive for coliforms by the MPN method. Closer
examination of the data shows a correlation between excessive bacterial densities and
desensitization of the membrane filter procedure.
While detection of coliforms by both the M PN and M F techniques increased as the
standard plate count increased, the MF method detected disproportionately fewer
coliforms as noted in the range of MF coliform values compared to the MPN
coliform values per 100 ml. Although it is not apparent from the data in Table 2 that
the MPN method is also subject to interferences by SPC organisms, a limited number
of water samples examined by the method displayed growth inconsistencies. Several
MPN tubes that showed no gas, but heavy growth, were inoculated into brilliant
green tubes and incubated at 35° C for 48 hours. These tubes did produce gas,
indicating that the competing non-coliform bacteria had suppressed the normal
metabolic pathway to gas production by coliforms in the presumptive medium. Such
false-negatives were also noted with the MF procedure. Membrane filter cultures
with excessive densities of colonies but no differentiated sheen colonies, were
aseptically transferred into lauryl tryptose broth for evidence of gas production
within 48 hours at 35° C then confirmed in brilliant green bile broth. This extension
of the MF procedure also revealed the presence of undetected coliforms that were
marked by high densities of the Standard Plate Count.
Controlling Interferences Through Treatment
Improved water treatment practice is the first line of defense against problems
related to turbidity in finished water. With reduction of turbidity to below 1 NTU,
TABLE 2. COLIFORM OCCURRENCES VS. STANDARD PLATE
COUNT.
' ' • I
Standard Plate Count Range (per ml)
__ < 500 501-1000 > 1000
All Samples* 502 • 35 51
Samples with Coliforms**
(MPN Method) 50 8 15
Percent of Samples 10% 22.9% 29.4%
MPN Range (per 100 ml) 2-49 5-79 , 2-920
MPN Mean (per 100 ml) 16 18 110
Samples with Coliforms* * *
(MF Method) 11 2 5
Percent of Samples 2.2% '5.7% 9.8%
MF Range (per 100 ml) < 1-18 < 1-50 < 1-90
MF Mean (per 100 ml) 2 7 13
MPN % Occurrence less
MF % Occurrence 7.8 17.2 19.6
* 588 of 613 samples had countable SPC
** 3 MPN positive samples did not have countable SPC
*** 1 MF positive samples did not have countable SPC
• !
16
-------
INTERFERENCES TO COLIFORM DETECTION/E.E. GELDREICH
TABLE'S. MPN COLIFORM OCCURENCES VS. FREE CHLORINE
RESIDUALS (DPD)*
Free Chlorine Number of Coliform SPC Average
Residual (mg/1) Occurrences (bacteria/ml)
<0 1 . . 58 3040
0.1-0.5 ,13 860
0.6-1.0 4 50
>1.0 1 , 79 .
* Typical values for finished waters were 1-2 mg I free chlorine and Standard Plate
Count < 1' ml. ... .
disinfection efficiency increases and enhances control of bacterial regrowth
throughout the distribution network. Analysis of the effectiveness of free chlorine
residuals in controlling bacterial deterioration, as demonstrated in this study (Table
3) and in the Community Water Supply Study (Table 4) prove the feasibility of this
approach for minimizing the potential non-coliform interferences to coliform
methodology (6,8,22,23). Pipe sediment accumulation in low-flow water mains and
dead-end lines is a source of both intermittent turbidity and nutritive materials that
support bacterial persistance and possible regrowth problems that may cause
coliform interferences. The total elimination of dead-end water mains is, of course,
impossible,' but implementation of practices such as periodic flushings, main
cleaning, maintenance of free chlorine residuals and modification or addition of
various treatment processes can substantially improve distribution w&ter quality in
dead-end mains and low flow sections. ' . ', .
Chemical characterization of 5 sediment samples from the Cincinnati distribution
system confirms'our argument that this material serves as a nutrient supply for
bacterial regrowth. Dried sediment samples were pulverized with O.I0 gm of each
TABLE 4 THE EFFECT OF VARYING LEVELS OF RESIDUAL CHLO-
' RINE ON THE TOTAL PLATE COUNT IN POTABLE WATER
DISTRIBUTION SYSTEMS*.
Standard Plate ,' Residual Chlorine (mg/1)
Count** ./ ... '
0.0 0.01 0.1 0.2 0.3 0.4 0.5 0.6
< 1 .. .
1-10
11 - 100
101 - 500
501 - 1000
> 1000
Number of
Samples
8.1%,
20.4
37.3
18.6
5.6
10.0
520
14.6%
29.2
33.7 •
11.2
6.7
4.5
89
• 19.7%
38.2
28.9
7.9
1.3
3.9
76
12.8%
48.9
26.6
9.6
2.1
0
94
16.4%
45.5
23.6
12.7-
1.8
0
55
17.9%
51.3
23.1
5.1
0
2.6
39
'4,5%
59.1
31.8 .
4.5
,0
0
: 22
17.9%
42.9
28.6
10.7
0
0
28
* Data from a survey of community water supply systems in 9 metropolitan areas (54)
** Standard Plate Count (48 hrs. incubation, 35° C)
17
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
I
sample added to 200 ml of distilled-deionized water and allowed to solubilize for 3
hours at ambient temperature. An aliquot of the supernatant was removed, acidified
with HC1 and analyzed with Dohrmann DC-52 carbon analyzer. Additional aliquots
of the supernatant were removed for Kjeldahl-nitrogen and nitrate-nitrogen analysis.
Kjeldahl-N ranged from 250 to 575 mg/1, nitrate-nitrogen from 4400-5200 mg/1 and
total organic carbon from 200 to'550 mg/1. Bacterial regrowth becomes a reality in
slow flow sections or dead end lines where such high levels of bacterial nutrients occur
in pipe sediments especially with water temperatures above 13° C, pH below 10, and
no free chlorine residual or ineffective chloramines present. Laboratory growth
studies using distilled water extracts of pipe sediments showed that nutrient
concentrations were capable of supporting bacterial growth from initial densities of
less than 100 SPC organisms per ml to 300,000 organisms per ml within 10 days at
20°C. ;
' ' 'I ' !
Controlling Interferences Through Methodology
Until all interferences created by inadequate water treatment and distribution
practices are eliminated in problem water supplies, the laboratory must remain
vigilant to the problem of coliform detection in potable waters of varying qualities.
Interferences to coliform detection often go undetected, partly because of a Standard
Methods concept that implies that no coliforms are present in a multiple tube test
when there is no evidence of gas production within 48 hours, or with the membrane
filter procedure when sheen colonies are absent. In the multiple tube procedure,
consideration should be given to confirmation of test samples that produce 5
presumptive tubes with heavy growth but lack visible gas production. For membrane
filter cultures that lack diiferentiated sheen colonies and that have high background
densities of non-differentiated colonies, the entire membrane filter should be placed
in lauryl tryptose broth to test for gas production followed by confirmation in
brilliant green bile broth. This procedure would yield qualitative information on
suppressed coliform occurrences that might otherwise go undetected. Since some
types of turbidities between 1 and 5 NTU physically block membrane filter pores,
impede flow through and prevent development of discrete differential colonies, the
laboratory will be obligated to use the multiple tube procedure until the interference
problem is corrected through improved treatment or distribution system mainte-
nance.
Conclusion
In an effort to detect coliform occurrences at a baseline of one organism per 100 ml,
potable water must be free of interferences from turbidities that exceed the 1 NTU
limit specified iiUheJtetedm_£rjbiary Drinking Water Regulations and bacterial
densitiejjjy^u^blorganismspermhshould not be tolerated. To some extent, these
interferences may be circumvented in the laboratory through alterations in
methodology. However, these measures should only be considered temporary
expedients to the monitoring program. The ultimate solution to these issues resides in
improved treatment practices anjd attention to minimizing deterioration of water
quality during distribution. >
REFERENCES
j , i i •.. .1
i " i j| • • • !
1. Bayliss, J.R. 1938. Bacterial aftergrowths in water distribution systems. Water Works and
Sewage 55:720-722. i
2. Buelow, R.W., R.H. Taylor, E.E.,Geldreich, A. Goodenkauf, L. Wilwerding, F. Holdren,
M. Hutchinson, and H. Nelson. 1976. Disinfection of new water mains. JAWWA 66:283-
288.
18 . "' ' !| ':
-------
INTERFERENCES TO COLIFORM DETECTION/E.E. GELDREICH
3. Buelow, R. W., and G. Walton. 1971. Bacteriological quality vs. residual chlorine. JAWWA
65:28-35. :
4. Committee on Water Supply. 1930. Bacterial aftergrowths in water distribution systems.
Amer. J. Public Health 20:485-491.
5. Fischer, G, 1950. The antagonistic effect of aerobic sporulating bacteria on the coli-
aerogenes group. Zeit. Immun. u. Exp. Then 707:1,6-20.
6. Geldreich, E.E. 1973. Is the total count necessary? Sections VII-l-VII-2. Proceedings
AWWA Water Qualtiy Technology Conference, Cincinnati, Ohio.
7. Geldreich, E.E., R.H. Taylor, and M.J. Allen. 1974. Bacteriological consideration in the
installation and repair of water mains. Sections! VIII-j-VHI-5. Proceedings AWWA Water
Quality Technology Conference, Dallas, Texas. ;
8. 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. JAWWA
6^:596-602.
9. Hutchinson, D., R.H. Weaver, and M. Scherago. 1943. The incidence and:significance pf
microorganisms antagonistic to Escherichia coli in water. J. Bacteriol. 45:2S),
10. Jewell, A.B. 1942. Bacterial aftergrowths in the distribution system. Southwest Water
Works J. 23:13-14. ;
11. Kligler, L.J. 1919. Non-lactose fermenting bacterial from polluted wells and sub-soil. J.
Bacteriol. 4:35-42. I
12. Lee, S.H., and J.T. O'Connor. 1975. Biologically mediated deterioration of water quality in
distribution systems. Proceedings AWWA Water Quality Technology Conference,
Atlanta, Georgia. j
13. McCabe, L.J., J.M. Symons, R.D. Lee, and G.G. Robeck. 1970. Surveylof community
. water supply systems. JAWWA 62:670-687. • I
14. Peterson, N., and M. Favero. 1975. Significance of gram-negative bacteria in wate;
supplies. Proceedings AWWA Water Quality Technology Conference, Atlanta, Georgia.
15. Reitter, R., and R. Seligmann. 1947. Pseudomonas aeruginosa in drinking water. J. Appl.
Bact. 20:145-150.
16. Sanderson, W.W.,and S. Kelly. 1962. Discussion of human enteric viruses in water: source,
survival and removability; pp. 536-541. International Conference:Water Pollution
Research, Vol. 2, London. ! :
17. Sen, R., and B. Jacobs. 1969. Pathogenic intestinal organisms in the unfiltered water supply
of Calcutta and the effect of chioririatio'n. Indian J. MecirRes. 57:1220-1227.
18. Tracy, H.W., V.M. Camerena, and F. Wing. .19,66. Coliform persistence in highly
chlorinated water. JAWWA 58:1151-1159. ,
19. Victoreen, H.T. 1974. Control of water quality in transmission and distribution mains.
JAWWA 66:369-370. ;
20. Victoreen, H.T. 1973. What's the significance of nuisance organisms? Sections IX-l-IX-6.
Proceedings AWWA Quality Technology Conference, Cincinnati, Ohio. :
21. Victoreen, H.T. 1969. Soil bacteria and color problems in distribution systems. JAWWA
61:429-431. ;
22. Walton, G. 1961. Effectiveness of water treatment processes as measured by coliform
reduction. Public Health Service Publication 898. United States Public Ijealth Service,
Washington, D.C.
23. Weaver, R.H., and T. Boiter. 1951. Antibiotic-producing species of Bacillus from well
water. Trans. Kentucky Acad. Sci. 73:183-188.
QUESTION AND ANSWER SESSION
C.D. Cameron, Fairfax County Water Authority, Occoquan, Virgina
What effect would public announcements urging residents to use less water during
water shortages have upon the distribution system water quality? Would not a long
holding time (2 weeks versus 2 days) in system mains and tanks encourage loss of C12
and an increase in total population? i ,
19
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
: ! . '
E.E. Geldreich, Water Supply Research Laboratory/MERL, U.S. Environmental
Protection Agency, Cincinnati, Ohio
There are several factors not denned that will impact on the status of the standard
plate count population in a distribution network during slow flow periods. During
periods of warm water temperature (above 13° C) and increased chlorine demand in
the distribution lines, regrowth of the bacterial population will occur. Some of this
regrowth potential might be suppressed where the finished water pH ranges from 9.0
to 10.0. If chloramines are carried as a residual, longer contact time in the
distribution network would result in some limited control over bacterial regrowth.
As a temporary counterforce, less abrupt changes in water pressure during this
period would reduce the mixing of sediments into the water flow. However, our data
would suggest the overall effect would be a gradual degradation of the water
regardless of limited benefits achieved by longer contact time and water quiescence.
I"
R.R. Colwell, Department of Microbiology, University of Maryland, College Park,
Maryland
Since 90% of the total count is missed when pour plates of water samples are
incubated at 35° C, why not use lower incubation temperatures?
E.E. Geldreich
We recognize that not all viable bacteria from a mixed population in a water
sample can be cultivated on a single medium, at any one incubation temperature or
for one specified incubation time. The critical concern with the general population of
bacteria in potable water lies primarily in the potential interferences they may create
in the test for coliforms. The other determiningfactor is to attempt to include some
indirect measurement for Pseudomonas, Flavobacterium, and other ^secondary
pathogenic invaders which could pose a health risk^ in the ^aspttaTenvjracmient.
Since the coliform test is incubated at 35° Cand the~potential secondary pathogenic
invaders optimally grow at body temperature, the choice of a non-specific bacterial
count in water at 35° C was logical. Furthermore, any attempt to specify a limit for
this portion of the general bacterial population must follow strict adherence to a
specific protocol for medium, incubation temperature and time. Thus the expression
"Standard Plate Count" becomes appropriate and the bacterial population limited to
growth on Standard Plate Count agar, with incubation of pour plate cultures at 35° C
for 48 hours.
20
-------
DISEASE TRANSMISSION/G.F. CRAUN
Impact of the Coliform Standard on the
Transmission of Disease
Gunther F., Craun
Chief, Epidemiology Branch
Field Studies Division
Health Effects Research Laboratory
. • Cincinnati, Ohio 45268
Dr Hendricks, the Symposium Chairman, has asked that I review and summarize
the data on waterborne disease outbreaks in the United States, concentrating on the
water system deficiencies that allowed the outbreaks to occur. I will also discuss the
bacteriological quality of water provided by the water supplies involved in outbreaks
and whether bacteriological surveillance as proposed in the National Interim
Primary Drinking Water Regulations will prevent the occurrence of waterborne
outbreaks. ' ; . .
The Center for Disease Control (CDC) and the Environmental Protection Agency
(EPA) jointly conduct surveillance of waterborne disease. Outbreaks are reported to
the CDC and the EPA by state and local health departments and by water supply
agencies. This report summarizes data reported during 1971-1975. ;
Only outbreaks associated with water used for drinking or domestic purposes are
included in this analysis. To be considered an outbreak at least two cases of infectious
disease must be reported before a common source can be noted and investigated.
Except in unique circumstances, such as a case of chemical poisoning in which the
chemical was identified in the water, a single case cannot be recognized as having
been caused by drinking water. The waterborne outbreaks reported here are those in
which drinking water has been implicated epidemiologically as the vehicle ol
transmission of the illness. In most of the outbreaks, the water was found to be
bacteriologically or chemically contaminated. '
For analysis, the water systems were classified as municipal, semi-public, or
individual. Municipal water systems are denned as public or investor-owned water
supplies that serve communities. Individual water systems are those used exclusively
by single residences in areas without municipal systems or by persons travelling
outside of populated areas (e.g., backpackers). Semi-public water systems, located in
areas not served by municipal systems, are developed and maintained for a group ot
residences (e.g., subdivisions and trailer parks) or at locations where the general
public has access to drinking water (e.g., industries, camps, parks, resorts,
institutions, and hotels). The definition of municipal water system and semi-public
water system correspond to the Safe Drinking Water Act's definitions of community
water system and non-community water system, The major exception is that small
subdivisions and trailer parks are included as semi-public systems in this analysis,
whereas the Safe Drinking Water Act defines these as community water systems.
During the period 1971-1975, 123 waterborne outbreaks were documented in the
U.S., resulting in almost 28,000 cases of illness (Table 1). The three largest outbreaks
occurred in municipal water systems: Sewickley, Pennsylvania (5,000 cases ol
gastroenteritis) in 1975; Rome, New York (4,800 cases of giardiasis) m 1974; and
Pico Rivera, California (3,500 cases of gastroenteritis) in 1971. Two deaths were
associated with outbreaks of waterborne disease during this period.
The mean annual number of reported waterborne outbreaks during 1971-1975
was 25. This is twice as many outbreaks as the mean annual number reported during
the period 1951-1970 and equals the mean annual number reported during the period
1920-1936 (Figure 1). The reason for this apparent increase in number of outbreaks is
difficult to ascertain but is felt to be primarily the result of increased reporting and
surveillance activities. :
21
-------
KJ
Source: G.F. CRAUN
U.S. Environmental Protection Agency
Cincinnati, Ohio
1920-251926-301931-36 1938-401941-451946-50 1951-551956-60 1961-65 1966-70
YEAR
1971-75
n
§
W
i
n
£
SB
I
Figure 1. AVERAGE ANNUAL NUMBER WATERBORNE OUTBREAKS, 1920-1975.
-------
DISEASE TRANSMISSION/G.F. CRAUN
TABLE 1. WATERBORNE DISEASE OUTBREAKS.
1971-1975 i
1971 1972 1973 1974 1975 TOTAL
Outbreaks 19 29 26 . 25 24 123
Cases of Illness 5182 1638 1774 j 8356 10,879 27,829
The 123 waterborne outbreaks were classified b,y type of water system (Table 2).
More outbreaks occurred in semi-public water systems (57%) than municipal systems
(30%) and individual systems (13%); however, outbreaks in municipal systems
affected an average of 504 persons compared to 129 persons per outbreak in semi-
public and 9 persons per outbreak in individual systems. Although most of the
outbreaks (57%) occurred in semi-public water systems, most of the illness (67%)
resulted from outbreaks in municipal systems. '•
TABLE 2 WATERBORNE OUTBREAKS BY TYPE OF SYSTEM.
1971-1975 !
i
OUTBREAKS CASES OF ILLNESS
Municipal Systems 37 18,633
Semi-Public Systems 70 9,058
Individual Systems 16 138
Total ~ i 123 27,829
An etiologic agent was determined in only 49% pf the outbreaks during the period
1971-1975 (Table 3). The remainder were characterized as acute gastrointestinal
illness of unknown etiology. Outbreaks in which ill persons had symptoms including
abdominal cramps, nausea, vomiting, and diarrhe;a 24 to 48 hours after consumption
of water and in which no specific etiologic agent could be determined were grouped in
this entity. This includes a clinical entity known as "sewage poisoning" which is
presumably caused by either coliform organisms or certain viruses that have yet to be
fully characterized. ; '•
TABLE 3. ETIOLOGY OFJVA.TERBORNE OUTBREAKS.
-1971-1975 •!-
OUTBREAKS CASES OF ILLNESS
. i -
Acute Gastrointestinal Illness j 63 17,752
Hepatitis-A : 14 '368
Shigellosis I I4 2'803
Giardiasis ! 13 5,136
Chemical Poisoning 1 12 511
Typhoid j ' J 222
Salmonellosis ! 2 J/
Enterotoxigenic E. coli I 1
Total
123 27,829
23
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
The most commonly identified pathogen during this period was Shigella; there
were 14 outbreaks and 2,803 cases of shigellosis: Most of these occurred in non-
municipal systems.
Giardia lamblia is a flagellated protozoan responsible for giardiasis. Clinical
manifestations of Giardia infection can range from asymptomatic cyst passage to
severe malabsorption. The mean duraction of illness is often 2 to 3 months. Only
recently has G. lamblia been recognized as a cause of acute illness and as the etiologic
agent in common-source outbreaks (3,2). The 'disease has received considerable
attention since 1970, as a result of studies implicating tap water as the probable mode
of transmission of giardiasis in travelers returning from Leningrad. This increased
awareness of physicians and public health workers has no doubt been partially
responsible for the increased reporting of this disease in .the U.S. Fifteen waterborne
outbreaks of giardiasis have been documented in the U.S. since the outbreak at
Aspen, Colorado in 1965-1966. Thirteen occurred during the period 1971-1975,
affecting 5,136 individuals. . .
Until the 1974 outbreak at Rome, New York, where 4,800 persons were affected,
the waterborne outbreaks of giardiasis have generally involved small municipal
systems or semi-public systems in recreational areas. Almost all of the outbreaks
have occurred as the result of drinking untreated surface water or surface water
whose only treatment was disinfection. Giardia cysts are not destroyed by
chlorination at dosages and contact times normally employed in water treatment, but
it is felt that they can be removed by well operated conventional treatment plants
employing coagulation/flocculation, settling, and filtration.
Although adequate disinfection data are not currently available, it is felt that
Giardia cysts areas resistant to chlorination as cysts of E. histolytica. If this is correct,
high concentrations of chlorine and long contact times are required in situations
where settling and filtration cannot be employed; this may necessitate dechlorination
prior to distribution (I). It has been reported that G. lamblia cysts can survive in
water and remain infective for sixteen days (5). Rendtorff (6) conducted a series of
studies with prisoner volunteers and found that £ of 22 men who received 10-25 G.
lamblia cysts became infected and all 13 men who received from 100-1,000,000cysts
became infected. Because an obvious source of gross human contamination could
not be identified in most of the waters systems having giardiasis outbreaks, some wild
and domestic animals are suspected as being important in the transmission of the
disease to man. It is possible in water systems using surface water-whose only
treatment is chlorination at conventional dosages and contact times that sufficient
disinfection will be employed to produce a coliform-free water but not agiardia-free
water.
The were 14 outbreaks of waterborne viral hepatitis affecting 368 people during
1971-1975. There has been .considerable controversy regarding the existence of
viruses in treated water supplies and the possible health consequences. Hepatitis-A
has been epidemiologically implicated in 66 waterborne outbreaks since 1946 and the
data can be examined to determine how these viral outbreaks occurred. Of the 22
outbreaks occurring in municipal systems, three resulted from either inadequate or
interrupted disinfection and five were related to the use of contaminated, untreated
surface or ground water. Half (eleven) of the outbreaks in municipal systems
occurred as the result of contamination of the distribution system, primarily through
cross-connections and backsiphonage. There is no evidence from waterborne
outbreak data that hepatitis-A virus has been transmitted through water systems
with properly operated conventional treatment plants except where distribution
system deficiencies have been found as the source of contamination.
There were 12 chemical poisonings involving fluoride, chromate, selenium,
phenol, furadan, arsenic, ethyl acrylate, a mixture of lubricating oil and kerosene,
cutting oil, fuel oil, and a herbicide.
The four typhoid outbreaks affected 222 people and involved semi-public and
individual water systems. The largest outbreak, responsible for 210 cases of typhoid,
.':!.'
24
-------
DISEASE TRANSMISSION/G.F. CRAUN
TABLE 4. WATERBORNE DISEASE OUTBREAKS BY TYPE OF
DEFICIENCY
1971-1975
OUTBREAKS CASES OF ILLNESS
.Untreated Surface Water* 19 5,729
Untreated Ground Water 38 3,958
Treatment Deficiencies ' 39 , 10,139
Distribution System Deficiencies 15 • 7,468
Miscellaneous 12 535
Total 123 27,829
* Includes seven outbreaks of giardiasis in which surface water was chlorinated but
not filtered ;
occurred in February and March 1973, at the South Dade Migrant .Farm Labor
Camp, Dade County, Florida. This was the largest reported outbreak of typhoid
fever in the U.S. since 1939 (4). No deaths occurred, and the outbreak subsided with
few secondary cases and no transmission 1:o persons not connected with the camp.
Two small outbreaks of salmonellosis occurred during this period: 34 cases of 5.
typhimurium in a semi-public water system and 3 cases of salmonellosis in an
individual water system.
At Crater Lake National Park in 1975,.an estimated 1,000 cases of diarrhea
occurred following contamination of a spring, which served as the drinking water
source for the park, by overflow from an obstructed sewerage line. Enterotoxigemc
E. coli serotype 06:H16 were isolated from ill persons and from water samples.
Untreated or inadequately treated water was responsible for the majority of
waterborne outbreaks during 1971-1975 (Table 4). Treatment deficiencies, primarily
inadequate or interruption of chlorination, and the use of untreated; contaminated
ground water accounted for 63% of all outbreaks and 51% of all illness.
The outbreaks and cases of illness were also classified by cause and type oi water
system (Tables 5-6). As in the previous 25-year period, the major cause of outbreaks
in municipal water systems during 1971-1975 was contamination of the distribution
system- 41% of the outbreaks in municipal water systems occurred because of
deficiencies in the distribution of water. Generally, contamination of the distribution
system has occurred through cross-connections and backsiphonage, and the
resulting outbreaks have been quite contained, affecting relatively few people.
However during 1971-1975, outbreaks caused by contamination of the distribution
system were responsible for 40% of the total cases of illness occurring in municipal
water systems. Two large outbreaks accounted for most of the illness in this
particular category: an estimated 5,000 cases of acute gastroenteritis in Sewickley,
Pennsylvania, which was felt to be related to contamination of an uncovered storage
reservoir for treated water and 1,400 cases of a similar illness in Sellersburg, Indiana,
which was traced to sewage contamination of a water main during construction.
In the previous 25-year period, use of untreated ground water was responsible for
most illness in municipal water systems. During 1971-1975, four outbreaks occurred
in municipal systems because of the use of untreated ground water; however, only 74
(0.4%) cases of illness resulted, because the outbreaks occurred in small water
systems. Nine '(24%) outbreaks and 5,322 (29%) cases of illness occurring in
municipal water systems were related to treatment deficiencies; six (16%) outbreaks
25
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
TABLE 5. MUNICIPAL WATER SYSTEMS DEFICIENCIES
ASSOCIATED WITH WATERBORN^ OUTBREAKS.
1971-1975
OUTBREAKS CASES OF ILLNESS
Untreated Surface Water* 6 5,404
Untreated Ground Water 4 74
Treatment Deficiencies 9 5,322
Distribution Deficiencies 15 7,468
Miscellaneous 3 365
Total • 37 18,633
* Includes four outbreaks of giardiasis in which surface water was chlorinated but
not filtered
and 5,404 (29%) cases of illness to use of untreated surface water. Four of the five
giardiasis outbreaks occurring in municipal water systems were related to the use of
surface water where disinfection was the only treatment. These were included in the
"untreated surface water" category rather than the "treatment deficiency" category
because specific treatment was not provided to remove or inactivate this pathogen.
Conventional disinfection as the only treatment of surface water would not be
expected to destroy Giardia cysts.
Use of untreated ground water was responsible for most of the outbreaks and cases
of illness that occurred in semi-public water systems in the past. Although this is still
an important problem, the 1971-1975 data show that deficiencies in treatment were
also responsible for many outbreaks and illness occurring in these water systems. Use
of untreated ground water and treatment deficiencies accounted for 81% of the
TABLE 6. SEMI-PUBLIC WATER SYSTEMS DEFICIENCIES
ASSOCIATED WITH WATERBORNE OUTBREAKS.
1971-1975 :'_...:
OUTBREAKS CASES OF ILLNESS
Untreated Surface Water* 7 237
Untreated Ground Water 27 3,859
Treatment Deficiencies 30 4,817
Distribution Deficiencies 0 0
Miscellaneous 6 145
Total 70 9,058
* Includes three outbreaks of giardiasis in which surface water was chlorinated but
not filtered
26
-------
DISEASE TRANSMISSION/G.F. CRAUN
TABLE 7. SEASONAL DISTRIBUTION OF WATERBORNE OUTBREAKS.
1971-1975
January
February
March
April
May
June
July
August
September
October
November
December
Unknown
*T^_4.~l
MUNICIPAL
1
2
3
2
1
5
4
5
4
3
4
2
1
«*•*
SEMI-PUBLIC
TOTAL AFFECTING VISITORS
2
1
2
4
9
14
12
12
4
3
3
4
^T/X
1
0
1
4
6
12
10
11
4
1
1
3
£/1
outbreaks and 96% of the cases of illness in semi-public systems. This is most likely
related to the size and nature of these systems, since most semi-public systems would
use ground water. Ground water is generally thought to be free of contamination and
little may be- done to,prevent intermittent contamination of these sources. Where
disinfection is provided, for either contaminated sources or as a preventative measure
for intermittent contamination, little or nothing may be done to insure that
continuous and effective disinfection is maintained. Many of these systems may be
operated for only a part of the year, and this may explain the lack of proper operation
and maintenance.
A large number of waterborne outbreaks each year involves the travelling public
(Table 7). In 1971-1975, 54 of the 70 outbreaks (77%) in semi-public water systems
affected travelers, campers, visitors to recreational areas or restaurant patrons.
Thirty-nine of the 54 outbreaks (72%) involving a transient population occurred
during the months of May-August, the period when outdoor activities such as
picnicking, camping and vacationing are most common. This implies that there is
either increased contamination of these water supplies during this period or, if it is
assumed that the supplies are always contaminated, use by greater numbers of
susceptible individuals during this period..Outbreaks in semi-public water systems
peaked during the summer months in 1971-1975, which is consistent with the
previous 25 years of record. There appears to be little seasonal variation for
outbreaks in municipal water systems, either during this period or the previous 25-
year period.
Waterborne disease surveillance such as this is important for several reasons. First
and most important, it helps stimulate the recognition and investigation of ongoing
outbreaks and terminate them by identification of the water supply deficiencies that
allowed the outbreaks to occur. Second, it identifies deficiencies in water treatment
practices and leads to the formation of control measures. Regulatory agencies can
use the data for establishing water supply surveillance priorities based on outbreak-
related problems. For example, these data would indicate that additional emphasis
should be placed on surveillance of smaller water supplies, particularly the semi-
27
-------
MICROBIOL STAND Al IDS EVALUATION/C.W. HENDRICKS
public water supplies
that serve the travelling public. For these systems, increased
emphasis should be placed on protection of the source water from contamination,
disinfection of these water supplies, and proper operation of the systems to insure
that continuous, effecl ive disinfection is maintained. Outbreaks in municipal systems
would indicate that additional emphasis is needed in the area of cross-connection
control and prevention of distribution system contamination. Third, waterborne
disease surveillance can assist in the recognition of important or new etiologic agents
so that appropriate witer treatment can be provided. For example, the data indicate
that conventional disinfection as the only treatment for surface water sources is
ineffective in preventi ;ig waterborne transmission of giardiasis and that to protect
against transmission of this disease all surface water should be treated by
sedimentation and filtration in addition to chlorination.
/I!, , - , ;: ^______
TABLE 8. SUMM
SYSTE
WERE
OUTER
CAUSE OF OUTER
I. USE OF UN"
1. 1/100 ml,.2
*2. (-), 3/100 r
**3. (-),(-), 200
4. (-), 21 100 r
5. (-),(-),(-),
II. TREATMEN
!.(-),(-), 8/1
t2. 15/ 100 ml
tt3-<2,2MPN
ttK 12 samples
5. (-), 7200/1
f ft|6. 4 samples (
III. DISTRIBUT
1. (-)
2. (-)'s and lo
3. 44 samples
>39/100,<
IV. MISCELLAI
1. 121 100, 22
2. (-)
3.1/100,1/1'
Raw water data (coli
* 246/100
** 100/ 100
t 291 100 n
tt >16 MP
TTT >50/100
tfTt >50/100
28
\RY OF BACTERIOLOGICAL DATA FROM WATER
AS WHERE LOW NUMBERS OF TOTAL COLIFORMS
FOUND IN TAP WATER SAMPLES DURING THE
EAK INVESTIGATION.
! ' • i " :
EAK:
7REATED WATER
.1/100 ml, TNTC
il, 3/100 ml, 8/100 ml,' 19/ TOO ml, 32/100 ml
HOOml
il, 3/ 100 ml, 61 100 ml, 10/ 100 ml, 13/ 100 ml
162/100 ml
T DEFICIENCY
)0 ml, 20/1 00 ml
all(-)
)0 ml
-), 3/100, 10/100 ml
ION DEFICIENCY
w levels within DWS limits
(-), 5/ 100, 8/ 100, 91 100, 10/ 100, 10/100, 26/ 100, >39/ 100,
8/ 100, 521 100
IEOUS OR UNDETERMINED
'100
)0, 3/100, 4/100
,,,''' |j , | ,'], , , i n
brms) for above; not necessarily collected at the same time
nl
nl
.1
M . ' .
ml ,.
ml
i ,„ ::;, , , •„, ,.?; j.,.r • |, ,: ,,„ i ,„!.!,,„ ,,.;„ , ,(1,
-------
DISEASE TRANSMISSION/G.F. CRAUN
Historical water quality data from water systems having a
can be used to determine the effectiveness of bacteriological su
during 1971-1975 with sufficient bacteriological data were examined
the level of contamination at. the time of the outbreak
bacteriological surveillance could indicate if an outbreak woi|ld
In the majority of instances water samples collected
investigation showed high levels of either total or fecal colifi
were outbreaks in which no or low numbers of coliforms we
classfied according to the cause of the outbreak (Table 8). It
these sample results are representative of the extent of conta
occurred in the presence of low coliform contamination, or i
actually larger and the samples were just not collected at the
There were five outbreaks in which the drinking water
implicated, but samples collected during the outbreak investiga
for coliforms. One was caused by a cross-connection, one by
distribution system, and another by contaminated ice. It is fe
were most likely collected after the contamination had passed th
are probably not representative of the water or ice which caused
were, however, representative of samples that might be cojllected
bacteriological surveillance, and as such, the results would certa
the remaining two outbreaks with distribution samples neg
contaminated water was distributed because of an interruption
known only because the treatment deficiency was documented
water was shown to be contaminated during the investigation
samples were'collected at the time nf .the treatment failure,
vaterborne outbreak
veillance. Outbreaks
todetermine(l)
and (2) if routine*1
occur.
during the outbreak
'c rms;. however, there
e found. These were
s difficult to judge if
mination and illness
if tlie contamination was
right place or time.
epidemiologically
tion were all negative
ontamination of the
It that these samples
•ough the system and
the illness. They
for rountine
nly be misleading. In
ative for coliforms,
of treatment. This is
and; the untreated
Unless distribution
was
OtHli L/JIWO YYWll< «^\_/il»^V> l,W»_l Clt-- VIIV- i.J.»ijy ii*""—ililflV ' 1 »-•"*.***»*** fc J.M.J.* v*«.yi^ •" ^-
surveillance would not have warned of a pending outbreak
ro atine bacteriological
outBreaks where negative cotffbrm samples were collected at
samples wereHFOsMvF'To^^
serve
There were several
th"e~sarne time other
to ; illustrate the
importance of collectirfgTimely samples and collecting sample
the systenTiranrouHneibacteriologlcal sampling program is
timely and^representative^sarripirng are a function oi the n
samples required, and certainly a minimum of one or two samples
quarter as permitted inJhe_NjaytjojiajJjalgjir|[LPrirnary Drinkin
(7) cannot proggrly_assess_lhe quality of water being consumed.
"it required"two or more samples to find the contamination
samples collected were negative for coliforms. Unless samp
continuously at a large number of points throughout the system, it is unlikely that
s at proper points in
be effective^ Both
mimum number of
to
per month or per
Water Regulations—.
n several outbreaks,
a|nd in one, 44 of 54
escanbeccjlseted
routine bacteriological sU£V6iltaiiC.e will bj? userul in preventing the occurrence of
"outbreaks. Since it is technically and financially mfeasible to
bacteriological monitoring, alternatives must be considered if
prevented. Monitoring for chlorine residual. is._a reasonabl
especially important for small systems where relatively few
are required.
Outbreaks having routine bacteriological surveillance dat
determine if the occurrence of the outbreak could be
These were categorized by the cause of the outbreak to illu
(Tables 9-12).
Previous bacteriological results were not available for those
distribution deficiencies, but several observations can be
bacteriological surveillance for prevention of these outbreaks.
examples of chemical cross-connections that have not resulted
because the water had a distinct color or odor not because of any
Bacteriologically contaminated water would not necessarily exh
istics, and it is highly unlikely that samples could be collected f
detect cross-connections and prevent outbreaks due to this caus<
occurred during 1971-1975 where bacteriological
provide continuous
outbreaks are to be
alternative and is
bacteriological samples
were; examined to
predicted for that water supply.
trate several points
mtbreaks caused by
n)iade about routine
There are numerous
n illness, but this is
outine surveillance.
ibit these character-
:equently enough to
:. Several outbreaks
entered the
contamnation
29
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRlCKS
'.. ' i " ' i ", '
TABLE 9 TWO OUTBREAKS CAUSED BY USE OF
UNTREATED, CONTAMINATED WATER.
Outbreak in Up'state New York Restaurant
1971
: ' I ' : .'
Acute Gastrointestinal Illness
Overall attack rate 500-1000/581-1163 (86%)
Tap water samples collected in restaurant before and after outbreak were negative
for coliforms except for one sample collected on 1 M9; 71 which showed 200 total
coliforms per 100 ml. Spring on several occasions prior to the outbreak yielded 100
coliforms per 100 ml.
Big Sky Montana Outbreak
1975
Acute Gastrointestinal Illness (Yersinia enterocolitica was isolated from the water.
but stools were not examined for this organism)
Overall attack rate 139/ 314 (44%)
It was reported that between 3/72 and 11 /18/74 samples were collected every 1-2
months and that all had been negative for coliforms. Water samples obtained on
1/14-16/75 yielded from 1 coliform to > 16 coliforms/200 ml. During the
investigation of the outbreak samples from the two wells showed:
I ' •:
1. 4 samples-Negative for coliforms
2. 8 samples - Positive for coliforms
4 - 1/250 ml
2-2/300 ml
1 - 2/300 ml
1 - 7/250 ml
, . il
distribution system and routine bacteriological sampling showed positive results.
This did not prevent an outbreak from occurring because of the length of time
between sample collection and analysis and subsequent recognition that the positive
result represented a real problem. It is often assumed that a positive result is due to
accidental contamination in the system. The Interim Standards (7) require check
samples to confirm positive bacteriologic results and require State notification within
48 hours. It is unclear whether the 48 hours applies to the results of the check sample
or the initial positive-sample. Public notification can range from 7 days by radio and
TV to the time of the next water bill. If the coliform standard is exceeded, by the time
check samples are collected for confirmation, and the maximum time is utilized for
notification, it will be too late to take action to prevent an outbreak from occurring. If
an outbreak does occur, the only consolation will be that sample results are available
to assist the epidemiologist in his investigation. However, with chlorine residual
monitoring, some action could be taken immediately by the operator after discovery
of a low or zero residual. It is unlikely the operator would assume that he had just
made a sampling error when he measures chlorine residual.
Surface water should never be used without treatment, but there are smaller
communities and semi-public systems where this is still done. Bacteriological
surveillance of these distribution systems can only provide a false sense of security.
There can never be assurance that contamination will not occur at any given time
even though the bacteriologic results have been negative. Systems using surface water
. • i .. .• I ;
30
-------
DISEASE TRANSMISSION/G.F. CRAUN
TABLE 10. STEELE ALABAMA OUTBREAK.
1973
Overall Attack Rate 50/100 (50%)
Infectious Hepatitis
Bacteriological Qualtiy of Distribution Systems
Date Samples Results
2/72 1 2/JOO
2/72-3/73 , 30 —
Raw water samples 2/72-2/73 15 samples
negative for coliforms; 7200 coliforms/100 ml found in late 3/73 during
outbreak investigation
and providing conventional chlorination as the only treatment are not uncommon in
some parts of the country. Bacteriological surveillance in these systems is also
meaningless, for conventional chlorination can provide a water free of coliforms and
yet not be effective in destroying Giardia cysts if they are present. ',
In many instances, ground water free of contamination is used without
disinfection. Proper construction and frequent sanitary surveys are required to
prevent contamination of these sources. Two examples of outbreaks caused by use of
undisinfected ground water illustrate this point (Table 9). The outbreak in the New
York State restaurant occurred on November 13-14, 1971, but tap water samples
revealed no contamination untirNovember 19. Although source water samples prior
to the outbreak revealed coliform contamination, the only treatment provided was
softening. A sanitary survey prior to the outbreak would have noted the
contaminated source and lack of proper treatment. The Big Sky, Montana outbreak
(Table 9) illustrates intermittent contamination of ground water felt to be of good
quality and relatively low numbers of coliforms can be associated with illness. The
water supply was epidemiologically implicated and chlorination of the source
stopped the outbreak. The previous bacteriological history had not predicted the
outbreak which occurred in December, 1974-January 1975, for all samples prior to
that time were negative for coliforms. Had a complete sanitary survey been
performed perhaps the sewage lines passing near the wells would have been noted as a
potential hazard as they were; during the outbreak investigation.
The remaining outbreaks to be discussed were all caused by treatment deficiencies,
and show that routine bacteriological surveillance is of little use in preventing
outbreaks due to this cause (Tables 10-12). Frequent sanitary surveys and increased
emphasis on maintaining continuous and adequate disinfection are important in
preventing outbreaks such as these. Routine bacteriological surveillance data from
the Steele, Alabama infectious hepatitis outbreak of February 21-March, 1973
(Table 10) showed relatively little contamination either in the distribution system or
the wells. The one positive distribution sample of 2 coliforms per 100 ml would most
likely not have caused any concern, as many of us would dismiss this as a sample
collector error. Contamination of the limestone aquifer and inadequate disinfection
caused the outbreak. Routine bacteriological surveillance was of little help in
preventing the outbreak. Monitoring of chlorine residuals, however, might have
alerted operators to adjust the level of chlorination during the period of
contamination.
31
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
TABLE 11. OUTBREAK IN A BOY'S CAMP.
1975 '
t
Acute Gastrointestinal Illness—3 separate outbreaks
Overall attack rates of 12/73 (16%); 85/160 (53%); 67/100 (67%)
Bacteriological Quality of Distribution System
F • • . ' i ••
' Date . | Samples Results
5/27/75 [. 2 " ~
6/10/75 ! 3 —
6/24/75 ' 3 —
.7/15/75 3 -
7/29/75 2
8/26/75 I , ./Too
i
Raw Water quality 5/75:8/75: 2 samples negative for coliforms
The outbreak at the Boys Camp (Table 11) is an excellent example of how
meaningless the results of a small number of bacteriological samples are in terms of
Sutine surveillance. The bacteriological record for 1975 would certainly not indicate
a problem with this supply, yet 3 outbreaks were *^™^*"™*^%£°*£
time in which the water supply ;was epidemiologically implicated Samples were
collected every month at approximately 2 week intervals, and the outbreaks occurred
on June 6, July 18, August 24, all between periods of sampling. The 3 wells used as a
water source were chlorinated, but it was reported that the operator did not monitor
for chlorine residuals. He relied instead on bacteriological results. Additional
information on this system was obtained in 1976 when another outbreak occurred. A
thorough engineering evaluation showed one of the wells to be contaminated with 6
to TNTC coliforms per 100 ml and inadequate control of chlonnation. me
contamination that-a routine bacferiological surveillance program failed to uncover
was finally found. The outbreaks most likely would, have been prevented with a
thorough sanitary survey noting geological conditions of the well, etc. and
monitoring of chlorine residuals to assist 5n proper maintenance of disinfection.
the last example, an outbreak: at Richmond Heights, Florida, (Table 12) is one
with which I am personally familiar. Again routine bacteriological sampling oi the
distribution system showed no coliform contamination prior to the outbreak. On
January 14-15, 1974, chlorination was interrupted and untreated well water was
distributed Routine surveillance noted 3 coliforms per 100 ml in a sample from the
system on February 12. The samples collected on February 18 appear to be check
samples but this is not certain. One sample on this date showed 10 coliforms per ,100
ml The 19 negative samples on February 21 appear to be check samples also. During
the period''of time that check samples were being collected, the community was
experiencing an outbreak which affected 1,200 people. The outbreak investigation in
March revealed that one of two'wells was continuously contaminated by effluent
from a septic tank near the well.' A sanitary survey and sampling of the raw water
could have detected these deficiencies prior to the outbreak and a greater emphasis
could have been placed on continuous chlorination and monitoring of chlorine
residuals.
32
-------
DISEASE TRANSMISSION/G.F. CRAUN
Table 12. Richmond Heights Florida outbreak.
1974
Acute Gastrointestinal Illness Compatible With Shigellosis
(10 Culture Proven Cases)
Overall Attack Rate 1200/.6500 (18%) ^
Date
I/ 9/74
1/10/74
2/12/74
2/12/74
2/18/74
2/18/74
2/21/74
3/11/74
3/26/74
3/28/74
3/29/74
•3/31/74
Bacteriological Quality of Distribution System
i
Samples Results
9
2
1
1
1
2
19
9
2
6
3
1
3/ 100 ml coliform
10/100 ml coliform
SPG - 3/26-31/-74 -'Average ISO/ml
Raw Water - 3/27^4/74~~Total coliform
Fecal coliform
* House with Water Softner
>50/100 ml coliform*
3/100 ml fecal*
>50/100 ml
to>50/100ml
Conclusions '
i
Waterborne outbreaks continue to occur and vye should make use of these data to
plan and develop surveillance programs. Water supply deficiencies that allowed the
outbreaks to occur can be identified and the formulation of control measures should
be based on outbreak related problems. Information on etiologic agents can be used
to determine treatment requirements. Historical water quality data from systems
having an outbreak can be used to determine the effectiveness of bacteriological
surveillance. . i
The data indicate that increased emphasis bd placed on surveillance of smaller
water systems, especially those serving the travelling public. The formulation of
control programs should include increased emphasis on (a) providing adequate
treatment of surface water sources to prevent outbreaks of giardiasis, (b) assessing
the potential for contamination of ground water/sources and providing appropriate
treatment, (c) maintaining continuous disinfectipn, and (d) preventing contamina-
tion from entering the distribution system. ]
Thorough clinical, epidemiologic and laboratory investigations must be
conducted to identify possible new etiologic agents responsible for outbreaks of acute
33
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
gastrointestinal illnesses of unknown etiology. Recognition of outbreaks and timely
investigation are necessary for collection and analysis of appropriate water samples
and clinical specimens in this regard.
Routine bactgrj^^L*"'-vei11ance of water distnbution^stems for cohiorms,
purposes
aks. Results ot
roTEttle valrift in preventin^tlieoccurrenceofwaterborne
ou
rbetaEinTTFe length
such sampling cannotje?ecervecLintimefor.corrective action.tQ_be_
-oTTmTe~T£—d^b^^
i^irs-irrruicTn^o-ioii^^
iou IUUK iu cuamt. my. v^...*^." - — -- ------ -• - - ; •—*-.„ -
s. The idea^fcheck sarnpleslboadds to the time delay, since the iirst
assumed t.o be a sampling error rather than rea
contamination. It i
systems. The ^
momtSnri that provides-anlnaication ot
^g_-^r~amiT5Tr^^^
bacteriological results are misleading and meaningless.
I propose, instead, an alternative monitoring program that would allow our
limited microbiological resources to concentrate on research activities, such as
identification of etiologic 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 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 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. Bacteriological 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 the primary disinfection oi
choice ceases to be chlorine,' residuals or markers of alternative disinfectants would
be employed, and the development of these disinfectants should include technology
to measure residuals. 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 chlormator.
REFERENCES
•' '',,' ''T " ! ii ' , :
1 Chang, S.L. and G.B. Fair. 1941. Viability and destruction-of cysts of E. hystolytica.
' JAWWA 33:1705-1715. . .
2. Craun, G.F., L.J. McCabe and J.M. Hughes. 1976. Waterborne disease outbreaks in the
U S.—'1971-1974. JAWWA 55:420-424.
3. Craun, G.F., and L.J. McCabe. 1973. Review of the causes of waterborne-disease outbreaks.
JAWWA 65:74-84.
4. Craun, G.F. 1974. Microbiology—Waterborne outbreaks. J. Water Pollut. Control t-ed.
4(5:1384-1395.
34
-------
DISEASE TRANSMISSION/G.F. CRAUN
5. Rendtorff, R.C. and D.J. Holt. 1954. The experimental transmission of human intestinal
protozoan parasites. IV. Attempts to transmit Endamoeba coli and Giardia lamblia cysts by
water. Am. J. Hyg. 60:327-33%. '
6. Rendtorff, R.C. 1954. The experimental transmission of human intestinal protozoan
parasites. II. Giardia lamblia cysts given in capsules. Am. J. Hyg. 59:209-220.
7. United States Environmental Protection Agency. 1975. National interim primary drinking
water regulations. Fed. Register 40:59566 -59588.
35
-------
- =y ? f-- i===
-------
ALTERNATIVE INDICATORS/G.A. MCFETERS
Alternative Indicators of Water Contamination and Some Physiological
Characteristics of HeteroltropMc Bacteria in Water
Gordon A. McFeiers, John E. Schillinger, and David G. Stuart
Department of Microbiology
Montana State University
Bozeman, Montana 59715
Introduction
The purpose of this presentation is to list some of the most promising
microorganisms and procedures that have been suggested 4o indicate water
contamination of public health significance. In addition, studies dealing with
the comparative die-off kinetics of selected indicator bacteria and pathogens
will be discussed, followed by a brief consideration of reported bacterial
growth in certain aquatic environments. These subjects will not be treated
exhaustively with- a complete listing of literature citations at the end of this
paper.
In discussing the question of alternative indicator systems, it is important
to list certain criteria that are appropriate in judging their efficacy. Some may
seem idealistic, but they are presented within the same altruistic frame of
reference that is basic to the indicator concept; that is safe drinking water.
The criteria follow: .
A. An indicator should be applicable to .all types of water that may be
investigated.
B. If a microorganism, it should be present in greater numbers than the
pathogen in all cases when the Jlatter. is: found.- Alternatively, positive
indication should be present if -there is-a significant health hazard,
C- Any indicator microorganism should not increase significantly in the
absence of a health hazard.
D. Indicator microorganisms should be more resistant to the physiological
stress within aquatic environments and to the action of disinfectants
that are commonly used therein. As a result, the indicators should
exhibit greater survival than pathogens under the "real world" condition
of the water that is to be tested.
E. The indicator reaction or test data should be unique and characteristic
of that microorganism or determination.
F. The indicator methodology should! be of minimal complexity, rapid and
inexpensive. " "'.....'".'""'....""_";„::_".
G. Indicator mircoorganisms should! be harmless to man under usual
conditions.
H. The indicator or test should be proportional to the health hazard that
is present-
Indicator Microorganisms That Have Been Suggested
In this discussion we have arranged the more commonly used indicator
microorganisms first, with others to follows. ;
A. Coliforms, fecal and E. coli.
Since these organisms have already been discussed at this symposium,
we have chosen not to emphasize them in this presentation except to
mention that some of these microorganisms satisfy most of our criteria
37
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
, I
under a wide range of circumstances. They have also served us rather
well for many years (10). However, on occasion they have not been
detected when pathogens have been present, as in the case of the River-
side, California, outbreak of 1,965, and some authors (4) have attached
their efficacy, in part, on the'basis of their occasional non-fecal origin
in nature.
B. Fecal streptococci, * '
This group bacteria is frequently used as indicators of sanitary sig-
nificance- However, it must ;be acknowledged that some species are
capable of reproduction exclusive of warm-blooded animals. This was
documented in two papers by;Mundt in 1963 (15, 16). However, others
(7, 18) suggest that fec.al streptococci provide valuable supplemental
data in determining fecal water contamination if conducted with, the
fecal coliform test. Such a rriultibacterial detection scheme could em-
body the fecal coliform/fecal stretptococcus ratio (8).
C. Standard plate count.
Although it is virtually impossible to enumerate the total population
of heterotrophic bacteria within a water sample by a single test, the
standard plate count may provide useful water quality data, par-
ticularly if it accompanies othbr complimentary bacterial measurements.
V The rationale for this lies jin the observation that a great many
~""~-\ potential or secondary pathogenic invaders that are common to water
C/^ may be detected in this way. As a result, the overall water quality may
be more accurately assessed in addition to the detection of pathogenic
bacteria that are known to reproduce in water. Geldreich (9) has sup-
ported this contention by adding that a standard plate count limit
may also reduce interference in counting low numbers of coliform
bacteria. ...........
D. Clostridium perfringens
The use of this bacterium as an indicator organism has been advocated
by some European workers; most notably, Gunnar Bonde (2). Its spore-
forming charactertistic has made it desirable as an indicator in the
presence of disinfectants or when water sample transportation time is
protracted. In addition, its 'presence has been correlated with high
numbers of classical indicators (2). However, since many authors regard
this organism ubiquitous in nature, and it is a strict anaerobe, its wide-
spread usefulness as a sanitary indicator can be questioned.
E. Klebsietta pneumoniae. \
This bacterium has been employed as an indicator of sanitary sig-
nificance in Britain and water isolates were recently demonstrated as
indistinguishable from clinickl isolates of this opportunistic pathogen
(20). It is not considered by! most to be ubiquitous in nature since it
is usually associated with E. \coli, although it is found in carbohydrate
rich wastes and on vegetables and seeds. In addition, it was found in
the chlorinated drinking water of Chicago as a fecal coliform (19).
It has been pointed out that this organism is capable of proliferating
in waters under some environmental conditions (14). This widely held
view casts some doubt that Klebsiella species may independently serve
as definitive indicators although it may be frequently enumerated as a
total or fecal coliform.
F. Aeromonas species.
Several species of Aeromonas are pathogenic to a variety of aquatic
fauna and man. These microorganisms are widely distributed in nature,
38
-------
ALTERNATIVE INDICATORS/G.A. MCFETERS
they are considered ubiquitous by some authors (6) and are capable
of prolonged survival in many aquatic environments. Because of these
qualities, we regard Aeromonas species of limited value as indicator
organisms in potable water.
G. Pseudomonas species.
These opportunistic pathogens represent particular problems in stored
water and some medical procedures!using water. They are ubiquitous
in nature and may actively reproduce in potable and distilled water to
populations as high as 10 Vml (3). | Therefore, these microorganisms
are a significant health hazard, especially to the hospitalized patient.
They may also persist in such water for extended periods of time.
Because of this capability, thesie bacteria are frequently found in the
absence of coliforms; one study reported that one-half of the water
samples containing Pseudomonas aeruginosa lacked coliform bacteria
(17). | :
H. Pathogenic microorganisms. j
Some pathogenic microorganisms like Salmonella species occur in
contaminated water in a statistically reproducible manner if the water
is grossly contaminated. However, under conditions where the con-
tamination is present in a lower level, such as potable water, the
occurrence of pathogens is less predictable. No single'pathogen suffi-
ciently satisfies the criteria of an effective indicator organism- There-
fore, until one such pathogen is identified and appropriate detection
methods are developed, we feel that the pathogenic microorganisms
have many shortcomings as indicators. At some time in the future,
developments may allow a pathogen like one of the enteric viruses to
serve as such an organism. i
. ... 1. .Others, .,. _...,,..,.,.. .,_,.i.,.,„.,....,. ...,_...._. . .
Five additional'microorganisms will be mentioned. However, the present
level of information on each regarding distribution and population
dynamics in water is inadequate to make a judgment as to their suit-
ability as indicators. They arejCaryophanon latum, the Bifidobac-
terium, bacteroides, coliphage and certain acid-fast bacterial ~ ~
i
Other Indicator Procedures: Rapid Tests ;
The promise of rapid alternatives or supplementary indicator systems is a
highly desirous yet elusive goal of water microbiology. This realization awaits
the resolution of many theoretical ami practical problems through research
and testing of procedures such as the ATP biometer, liberation of 14CO2
from radio-labeled lactose and the Limulus 'lysate test.
The Limulus lysate test for bacterial endbtoxin warrants further mention.
There is a growing body of evidence that this test could be the most promising
of the rapid indicator procedures. Studies that were done by Tom Evans in
our laboratories (5) illustrate the relation o|f this test to some conventional
bacterial water quality criteria. The results seen in Figure 1 are typical of
our findings and suggest that the Limulus lysate assay for endotoxin parallels
results of some conventional indicator bacterial tests in natural waters ranging
in quality from pristine streams to the discharge of a sewage treatment plant.
Answers to questions such as the longevity of endotoxin under conditions of
water renovation and disinfection along witli further refinement of the assay
will bring this procedure within the scope of practical usefulness in the near
future. i
39
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
10^
s
cc
Ul
o
§
10'
--» CONCENTRATION Of ENDOTOXIN
NUMBER.OF HETEROTROPHIC
BACTERIA/ml
NUMBER OF TOTAL COLIFOHM
BACTERIA/100ml
NUMBER OF GRAM . NEGATIVE
BACTERIA/ml
NUMBER OF ENTERIC
'BACTERIA/mi
JL
-L
_L
10°
EFI HB EG4 OF2 EG5 EG5a
SITE
Figure 1 BACTERIAL AND ENDOTOXIN PROFILE OF THE EAST GALLATIN RIVER
3 DRAINAGE FOR SAMPLING DATES 7/11/75 AND 7/15 7Ei EACHI POINT
REPRESENTS THE GEOMETRIC MEAN OF FIVE REPLICATE SAMPLES
ENUMERATED BY THE SPREAD PLATE TECHNIQUE.
Factors Influencing the Recovery and Enumeration oi Indicator Bacteria
An understanding of the population dynamics and physiological injury of
heterotrophic bacteria in water is basic to the realistic application of the
bacterial indicator concept. This assertion is based on the observation that
the aqueous environment, from which the indicator bacteria are recovered is,
at most, an extremely dilute solution of bacterial nutrients. As a resuh, when
bacteria are released into the water from their nutrient-rich natural habitat,
thev eo through a spectrum of/debilitation during which enumeration becomes
progressively more difficult until they finally die. On the other hand, there is
evidence that some heterotrophic indicator bacteria metabolize and even grow
in some aquatic environments that are oligotrophic by all other standards.
cc
la
cc
ui
ca
13
40
-------
ALTERNATIVE INDICATORS/G.A. MCFETERS
An overview of these considerations will be given within the context of studies
that have been carried out in our laboratories-
In studying the physiological behavior and the population! dynamics of
heterotrophic microorganisms in water, methodology can be very important.
For example, coliform bacteria.are capable of cryptic growth,; using the or-
ganic chemicals liberated from dead cells that have undergone autolysis, if
the population is on the order of 106/iml in an enclosed container. For that
reason we developed the membrane diffusion chamber (Figure 2) to allow
the retention of a population of microorganisms within an enclosure that is
immersed in a body of [ water (Figure 3) (12). Water is free to diffuse through
the enclosure, carrying remnants of dead cells away and allowing a. more
realistic experimental approximation of what is likely to occur in the natural
aquatic environment. In addition, we found that the population of cells must
not be excessive (Le. less than lOVml) in the performance of these
experiments. . ...
Studies were carried out using these chambers to investigate the compara-
tive survival of classical indicator bacteria and some • waterborne pathogens
in water (13). It was demonstrated that coliforms and enterococci persisted
E
o
2
2 4
10
cim
Figure 2. SCHEMATIC DRAWING OF THE MEMBRANE DIFFUSION CHAMBER.
41
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
Figure 3. PHOTOGRAPH OF MEMBRANE DIFFUSION
CHAMBERS IN OPERATION SUBMERGED IN
WATER.
i
42
-------
ALTEELNATWE INDICATORS/G.A. MCFETERS
as well as or better than the pathogens that were studied (Table \ 1). Also, the
fecal coliform to fecal streptococcus ratio of human fecal wastes decreased
while that from a bovine source increased with time in water to a point
where the ratios were not useful in distinguishing the origin of the pollution
within 24 to 48 hours. In another study that was carried out in our labora-
tories, largely by Gary Bissonnette, the physiological debilitation of indicator
bacteria in water was examined (1). This work demonstrated that indicator
bacteria progressively lose the ability to grow and form colonies^ on selective
media as they age in natural water (Figure 4) and that an appropriate
nutrient environment and moderate temperatures allow the repair of such
injury within three hours (Figure 5). These studies indicated that this occur-
rence is widespread among indicator bacteria in nature and that injured cells
may comprise as much as 90 percent of the population.
An ongoing investigation describing the interaction between indicator
bacteria and algae in the high alpine zone also relates to the broad question
of indicator bacterial behavior in water to the collection and interpreta-
tion of. meaningful bacterial water quality data. Relatively high numbers of
total coliform bacteria in a mountain stream have been shown to be associated
with an algal mat community. In laboratory studies, indicator bacteria includ-
ing total coliforms, fecal coliforms and members of the genus Klebsietta
TABLE 1. COMPARATIVE DIE-OFF RATES (HALF-TIME)* OF
FECAL INDICATOR BACTERIA AND ENTERIC PATHOGENS.
.__'_'- -••-"•" •••.•'•— "'. .-'" - '"_"_"___'_ . Bacteria No. of
"^Bacteria ••--•-- Half-time I strains
(h) analyzed
Indicator bacteria
Coliform bacteria (avg) 17.0 29
Enterococci (avg) 22.0 20
Coliform from raw sewage 17.5
Streptococci from raw sewage 19.5
, Streptococcus equinus 10.0 : 1
S, bovis - 4.3 1
Pathogenic bacteria
Shigella dysenteriae '."!."."~.'.'.V.".'".'l ]'.' ..... 22A* 1
S, sonnei 24.56 1
S. flexneri 26.86 . 1
Salmonella enteritidis ser. paratyphi A ........ 16.06 1
S. enteritidis ser. paratyphi D 19.26 1
S, enteritidis ser. typhimuriuni 16.06 ', 1
S, typhi 6.0 2
Vibrio choleras 7.2 3
S. enteritidis ser. paratyphi B 2.4 1
"The half-time was determined grapbicsilly from Fig. 2 and 4 as the time
required for a 50% reduction in the initial population.
6 The half-time was determined graphically from Fig. 4 as the time required for
a 50% reduction in the population at 24 h.
43
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
DAYS
Figure 4. COMPARATIVE RECOVERY FOR E.COLI C320MP25 IN MEMBRANE
FILTER CHAMBERS LOCATED AT SITES BR2 AND EG5 DURING A
4- DAY EXPOSURE PERIOD. SAMPLES WERE SURFACE : OVERLAY
PLATED USING TSY AND KF AGARl
increased by two to three orders of magnitude at low temperatures using the
products of algae from the mat community. The likelihood that these bacteria
were reproducing this way in nature was further reinforced by the finding that
they metabolized radioactive algal organic extracellular products in controlled
laboratory experiments. These findings indicate that under some circum-
stances heterotrophic bacteria are capable of growth in a high quality water.
This was previously suggested by Hendricks and Morrison (11) who pro-
posed that stream sediments adsorb organic substances that are subsequently
used by bacteria. The implications of these studies go far beyond the alpine
streams in which they were done. For instance, surfaces within water storage,
treatment and distribution facilities may adsorb and concentrate potential
44
,; | „
-------
ALTERNATIVE INDICATORS/G.A. MCFETERS
2 3~4
HOURS (in TSY broth)
Figure 5. REPAIR OF INJURY IN TSY BROTH FOR E. COLI C320MP25 CELLS
HAVING BEEN EXPOSED TO THE STREAM ENVIRONMENT OF SITE EG6
FOR 2 DAYS. CONTROL OR O - TliME CELLS AND 2 - DAY EXPOSED
CELLS WERE ENUMERATED OVER A 6 - HOUR GROWTH PERIOD IN TSY
BROTH USING TSY AND DLA AGAR SURFACE - OVERLAY PLATES.
bacterial nutrients and provide favorable conditions for bacterial growth both
before and after disinfection. Activated carbon filters are an example of an
increasingly popular water treatment process that is particularly susceptible
to this occurrence. With that in mind, the proliferation of bacterial indicators
or potential pathogens in potable waters must be considered in examining
alternatives for determining water quality. The growth of indicator bacteria
might give a false signal of water contamination, but multiplication of or-
ganisms such as Flavobacterium, Pseudomonas, and Klebsiella species, creates
a dangerous condition for some consumers that must be recognized if the
test procedures are to serve our needs. With the presently used bacterial in-
dicator surveillance protocol, this situation probably does exist in many finished
waters that are defined as potable.
45
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
I
I
Philosophical Basis of Indicator Bacteria and Recommendations
A discussion of "the philosophical basis of the indicator bacterial concept
is essential in discussing alternative indicator protocols for determining water
quality. The philosophical basis that we consider fundamental to the indicator
concept -is that an indicator system that is more encompassing is better than
one that is more restrictive, in terms of the source of the contamination that
will be detected. Additionally, false negatives are to be avoided more than
false positives if one has to select the lesser of the two evils. As a justification
for this position, we cite the distinction between sewage contamination and
feces contamination and make .the generalization that an effective indicator
system will detect not only feces but also sewage contamination. The impor-
tance of this notion is that water contamination from sewage rnay vary
widely in amounts of fecal pollution and, more importantly, that the broad
spectrum of health threatening chemicals and infectious agents that may be
transmitted through water are not exclusively associated with fecal material.
Translated into choices that have been presented, E. coli is most restrictive
since it is most frequently associated with feces followed by fecal coliforms
that are usually linked with fecal pollution, and then total coliforms that may
or may not be associated with fecal material. However, total coliforms are
a good indicator of sewage or the intrusion of surface or contaminated ground
water if found in finished drinking water. As another advantage, the total
coliforms are more easily and economically detected with less chance of
false negatives than fecal coliforms and E. coli. Also, the maximum con-
taminant level (MCL) for total coliform bacteria might be increased somewhat
because of the greater sensitivity in detecting a wide spectrum of contaminant
sources plus laboratory contamination with: this group of organisms. This
would help to satisfy some of the objections to the National Interim Drinking
Water Standards that have been voiced in many quarters recently.
Using such an indicator system or even E. coli, as in Britain, there is still
legitimate cause for concern regarding large numbers of potential pathogens
and other microorganisms in drinking water that are capable of reproducing
there. This concern was recently expressed by the British (21) in an article
that described the proliferation and detection of these organisms within water
distribution systems. We propose therefore that a test similar to the present
standard plate count be performed as a routine part of the microbiological
surveillance of all potable waters and that appropriate MCL's be set. Such a
test should be incubated for a minimum of two days at 22 to 30°C to ac-
commodate slower growing bacteria that are adapted to the reduced ambient
temperature of their aquatic environment. However valuable the information
gamed, the time lag required for the performance of this test presents a real
logistic problem since each re-test following an excessive result will take at
least two days. The implications of this are obvious in terms of the volume of
water used in every two-day period. For that reason the utilization of the
Limulus lysate test would be useful to check for excessive bacterial popula-
tions when the standard plate count or coliform counts are high and the
operator needs rapid assurance that his remedial action has been effective.
This idea may be somewhat futuristic since the Limulus lysate test has not
yet been perfected or widely accepted. However, we feel that this wiH become
a reality within the next two to five years.
Based on the arguments presented in the preceding sections of this dis-
cussion, we propose the following recommendations for a bacterial indicator
system in finished water.
46
-------
ALTERNATIVE INDICA'
B.
C.
A. Total coliforms should be retained as the primary
organisms with a slight relaxation of that MCL (
ml).
A test similar to the standard plate count should
regular basis as an index of othex microorganisms
in water. The MCL level might be near the 207nil
in Germany or as high as 500/ml.
In the event either the total coliform or standarld
exceed the MCL, a Limulus lysate test should be
becomes available.
These recommendations represent a realistic yet
to use the enumeration of heterotrophic bacteria
to measure the safety of finished drinking waters.
there are imperfections in this plan, as in others
posed, largely due to the statistical nature of the
inherent in searching for heterotrophic bacteria within
alien environment of treated water. In spite of these
good reason to believe that this protocol will improve
able health safety record of finished drinking
States.
e determined on a
that are important
limit that is used
plate count data
performed when it
comprehensive plan
and their products
We 'recognize that
lhat have been pro-
task and difficulties
the hostile and
difficulties, there is
the already laud-
Water in the United
REFERENCES
Es:h
1. Bissonnette, O.K., J.J. Jezeski, G.A. McFeters, and D.G. S
of environmental stress on enumeration of indicator b;
""waters. Appl. Microbiol. 29:186-194.
2. Bonde, G.J. 1962. Bacterial indicators of water
quantitative estimation. Teknisk Forlag,, Copenhage.
3. Carson, L.A., M.S. Favero, W.W. Bondl, and N.J. Peterse;
cal, biochemical, and growth characteristics of p. cepacia
Appl. Microbiol. 25:476-483.
4. Dutka, B.J. 1973. Coliforms are an inadequate index
Environ. Health 36:39-46.
5. Evans, T.M. 1975. The Limulus lysate assay a rapid
bacterial water quality. MS Thesis. Montana State
Montana.
6. Fliermans, C.B., R.W. Gorden, T.C. Hazen, and G.W.
distributions and survival in a thermally altered lake. A;
biol. 53:114-122.
7. Galvand, M.M. 1974. Fecal contamination—the water:
sponsibility. Sewage Works. 12:66-69.
8. Geldreich; E.E. and B.A; Kenner. 1969. Concepts of l
stream pollution. J. Water Pollut. Control Fed. 41:
9. Geldreich, E.E. 1973. Is the total count necessary?
Proceedings AWWA Water Quality Technology
Ohio.
10. Geldreich, E.iE. 1967. Fecal coliform concepts in stream
Sewage Works. 114-.R98-R110.
11. Hendricks, C.W. and S.M. Morrison. 1967. Multiplication
selected enteric bacteria in clear mountain stream water. Water
12. McFeters, G.A. and D.G. Stuart. 1972. Survival of coliform
waters; field and laboratory studies with membrane fill
Microbiol. 24:805-811.
'ORS/G.A.MCFETERS
group of indicator
e. less than 3/100
uart. 1975. Influence
.cteria from natural
pollution. In: A study of
. 1973. Morphologi-
firom distilled water.
of water quality. J.
j.nd sensitive test of
University, Bozeman,
. 1977. Aeromonas
ipl. Environ. Micro-
part I, analysts' re-
ecal streptococci in
:R336-R.352.
S< ction
Conference
vii-i—yn-2.
Cincinnati,
pollution. Water &
and growth of
Res. 1:567-576.
bacteria in natural
er chambers. Appl.
47
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
Coi nparatr
13. McFeters, G.A.,
Stuart. 1974.
gens in well wate
14. Menon, A.S. and
microbiology of
Environ. Technol.
15. Hundt, J.O. 1962.
Appl. Microbiol.
16. Mundt, J.O. 1962
Appl. Microbiol.
17. Nemedi, L. and
aeruginosa in wa
18. Phirke, P.M. and
of stream pollution,
19. Ptak, D.J., W.
dence of Klebsiel a
20. Seidler, R.J.,
environment:
pneuriioniae
29:819-825.
21. Water Research
water during di
Centre, London.
cu
fron
L. J. McCabe, Jr., £
Protection Agency,
What are some cf
of the Limulus lysat
range of ambient w;
.K. Bissonnette, J.J. Jezeski, Carole A. Thomson, and D.G.
:ive'survival of indicator bacteria and enteric .patho-
r."Appl. Microbiol. 27:823-829.
. W.K. Bedford. 1973. Study of the seasonal effects on the
:iorthern pulp and paper mill aeration lagoon. Can. Dept.
Develop. Rpt. 4-AR-73-1. Ottawa, Canada.
Occurrence of enterococci in animals in a wild environment.
11:136-140.
Occurrence of enterococci on plants m a wild environment.
77:141-144.
B. Layni. 1971. Incidence and hygienic importance of P.
er. Acta. Microbiol. Acad. Sci. Hung. 75:319-325.
S.R. Verma. 1972. Significance of enterococci as indicators
in. Indian J. Environ. Health 74:328-334.
jinsburg, and B.F. Willey. 1973. Identification and inci-
i in chlorinated water samples. JAWWA 65:604-608.
. Knittel, and C. Brown. 1975. Potential pathogens in the
tural reactions and nucleic acid studies on Klebsiella
clinical and environmental sources. Appl. Microbiol.
Centre. 1976. Deterioration of bacteriological quality of
tribution. In: Notes on water research. Water Research
QUESTION AND ANSWER SESSION
health Effects Research Laboratory, U.S. Environmental
iincinnati, Ohio
the more significant problems associated with the use
assay to determine natural water quality over the broad
rters that might be tested?
G. A. McFeters, Department of Microbiology, Montana State University,
Bozeman, Montana
As you have suggested, there are still some potential -problems in the ap-
plication of the Likiulus lysate test for determining water quality. For in-
stance, we feel that more needs to be known regarding .the persistence^ _of
endotoxin in waters of v3P"°LfI!fL chemical types as well as in
water treatment and
to dinerent
Beyond this, we
the assay procedure
renovation processes. JPossible references of the test
aiemistries or "additives Should also be investigated.
also acknowledge the need for
due
are commercially a
allow the correlaticn
itself (at present there are four different test kits that
ailable)." Also, more studies need to be performed to
of this new test_with_ other^slablisjig^^
\yaieE-qiiality. Such
ditipns.
testing should be done
world"water_can-
48
-------
STATISTICAL CONSIDERATIONS/L. MEUNZ
Some Statistical ConsideratioBS in Water Quality Control
Larry Muenz
National Cancer Institute, Biometry Branch
National Institutes of Health, Landow Building, C-509
Bethesda, Mainland 20014.
As an outsider I am particularly grateful for the opportunity to address
you today regarding water quality control. I am a statistician, not a bac-
teriologist or water quality expert, and have been asked to look at some
statistical aspects of water sampling. My analyses are based on the use of
the membrane filter method, although they could certainly be applied to
other techniques. Indeed, what I have done is an example of the general class
of quality control sampling analyses. Each sample must be classified as being
either good or bad depending upon whether or not one can detect some rela-
tively rare phenomena. My analyses today are also based upon the idea that
the use of coliform indicators is desirable, that they really indicate something
we want to know. I do not take into account two facets of the problem that
no doubt need ultimately to "be considered. They are (a) in measuring a
certain number of E.coli> using them as indicator bacteria, we have indicated
the presence of an uncertain number of pathogens of various sorts and
(b) the doses of pathogens needed to induce illness in say 1% of the people
drinking water with that level of contamination changes radically from one
sort of bacterium to another. Regarding (a), and this is an issue that the
experimentalists need to consider, it would be important to have a probabilis-
tic relationship between measured number of E. coli and probable numbers
of various sorts of pathogens believed to be present.
-"Specifically, I was asked to investigate three topics. Theser are (1) how
many samples a month should be taken for various size towns, (2) how
much water should be taken hi each sample and (3) are the present sampling
rules well defined? My knowledge of the rules is based upon the EPA interim
regulations of December 25, 1976, i.e. specifically, the table appearing in
paragraph 141.21 and the instruction of 141.14. \
The proposed regulations specify a. number of samples that increases
steadily with population. It is not evident that this is necessary or desirable.
For example, a nationwide voting poll doesn't need substantially more people
hi its sample than one restricted to a single city. A critical aspect is homo-
geneity: are we taking a sample from a homogeneous source? If so, many
big samples are no better than a smaller number of smaller samples. Figure 1
sHows'the "number of samples per 1,000 population according to the proposed
EPA regulations. There are several observations. First, it evidently declines
with increasing town size. This seems desirable. There is a small insert in
the beginning of the graph; this shows the number of samples per 1,000 popu-
lation for small towns up to a population of approximately 90,000. The extent
of this insert on the larger graph is indicated by a tick mark on the hori-
zontal axis of the larger graph. The insert shows an extremely ragged line,
definitely not a consequence of an analytic model. It is not necessarily bad
to use seat-of-pants procedures to determine the number of samples per head
but in this case it appears not to be rational for small communities.
The next question refers to the amount of water per sample. A standard
sample is now 100 milliliters. There is nothing unique about this value other
than that it corresponds to a convenient size dish or test tube. Standard
statistical theory can tell how large a sample is needed to makej sure that we
49
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
V)
z
8
en
tu
a.
Q
I
O
o
Ul
_J
a.
<
C/J
cc
Ul
I
1.2
1.1
\
V
090
2500
POPULATION (THOUSANDS)
I
Figure 1. NUMBER OF SAMPLES TAKEN PER 1, 000 PERSONS.
5000
will find something if something is there. Again, the critical issue is the
same as previously raised. Do the tubes properly reflect the heterogeneity of
the source? .
The regulations say that we should examine a certain number or. tubes
containing 100 milliliters each, that number depending upon community size.
For each batch of tubes, say 20 of them, the sampling rule is composed of
two parts. First, if the mean number of bacteria per tube is too high, declare
the sample to be contaminated. The threshold value is now set at one bac-
terium per tube. Second, if more than 5% of the tubes contain more than 4
bacteria—even though the overall mean may be less than one—also declare
the sample contaminated.
To study 'this two part sampling procedure we need an analytic model to
describe the number of caliform bacteria in a sample. With such a model
statistical theory can find the best way of testing the bacterial level of the
sample and then the proposed procedure can be compared with the optimal
one. It may be that the optimal procedure is excessively complex and the
proposed procedure is an effective, simple approximation.
The sampling problem is described by a number of parameters. These are:
1. The level of contamination that we consider just at the border line of
acceptable and not acceptable, that is the threshold. 2. A measure of vari-
50
-------
STATISTICAL CONSIDERATIONS/L. MEUNZ
ability of contamination from one water sample to another. 3. The probabil-
ity that with a given decision procedure and a certain level of contamination
that we will "pull the alarm" at exactly the border line level. 4. The proba-
bility that we will fail to detect some high level of contamination, one sub-
stantially greater than the border line value. 5. The historical information
on the frequency of waterborne disease in communities of various sizes.
It is important to distinguish between the case in which the water is uni-
formly mixed and homogeneous and one in which the water has an important
degree of heterogenity. With uniform mixing a relatively small number of
samples will suffice to indicate the average number of bacteria per volume of
water. In the heterogeneous case we can consider an extreme, indeed un-
realistic, example with an extraordinarily high degree of variability. This
might be reflected hi a situation in which all the contamination was con-
tracted into the form of a bolus. I call this the "bowling ball" model in which
unless you hit the particular focal point of contamination with your sam-
pling procedure you will detect no contamination at all. To have a reasonable
probability of detection we would need an extraordinarily large number of
samples. In the more realistic situation in which contaminants are somewhat
clumped it is evident that we will still meed a larger number of .samples than
in the situation of homogeneous mixing. :
When the water is uniformly mixed the relevant distribution for the num-
ber of particles per standard volume of water is the Poisson distribution
which has a single parameter and for which mean and variance are the same.
If we assume that this mean varies from sample to sample and can itself be
considered a random quantity we arrive at the negative binomial distribution.
Professor Pipes of Drexel University and some of his graduate students have
considered the practical problem of distinguishing between these two possible
distributions, Poisson and negative binomial. Based on a large number of
water samples they have almost invariably come to the conclusion that the
negative binomial fit the data better than does the Poisson. While this is
reassuring it isn't at all surprising since the negative binomial distribution has
two parameters and the Poisson only one. Thus we have more flexibility in
fitting with the negative binomial. Since the negative binomial arises for the
case of the Poisson where the mean is allowed to vary, in the extreme case
where the mean doesn't vary we get back to the Poisson, that is, one is a
special case of the other. The question then becomes: is the negative binomial
distribution enough of an improvement relative to the Poisson to warrant the
addition of another parameter? Using the data collected by Dr. Pipes' stu-
dents I conclude the answer is 'yes'. Making special provisions for a model
with a high degree-of heterogeneity, namely the negative binomial, is justified
in real data.
As I have indicated the sampling rule is in two parts, one referring to the
mean and one to the number of tubes containing an excessive number of
bacteria. As was indicated in a handoult received prior to this meeting there
are problems with this rule for very small numbers of samples. For example,
one bad tube can guarantee that the mean will be too high regardless of zero
counts in the others.
Although this rule may not be ideal I have analyzed it using the negative
binomial model referred to above. The model is parameterized by a mean
level of contamination, a coefficient of variation for the prior distribution of
the mean and a sample size, -that is, number of tubes. I have calculated the
probability that a given water sample will be accepted for various combina-
tions of the parameters. Figure 2 shows the results of a computer simulation.
51
-------
zs
PROBABILITY OF ACCEPTANCE OF SAMPLE
I
i
SSSSS
mnj_|3>m
3mX<2
m z m > 2
£232g
s|S8B
glsci
isi^q
aia
mil
= ro3oS
00
-------
STATISTICAL CONSIDpRATIONS/L. MEUNZ
Because I have used a simulation rather than exact calculations the answers
have a great deal of variability themselves and sometimes behave counter-
intuitively.
Some general observations are that: 1. There is still a; high probability of
accepting a relatively bad sample, one with a mean of three coliform per 100
mis., if the number of tubes is small enough. 2. There is a general tendency
for the acceptance probability to increase as the variability of the mean in-
creases. This is entirely reasonable if we think back to my bowling ball model.
3. Things do behave in a rational fashion in that the probability of ac-
ceptance goes down with increased mean and with increased numbers of
tubes.
If my model for bacterial distribution is a reasonable representation of
reality then this rule is not the optimal one. It is suboptimal in the sense that
for a given probability of considering uncontaminated samples to be con-
taminated we could achieve a higher probability of cqrrectly detecting a
contaminated sample with some other rule. Unfortunately the optimal rule
is a good deal more complicated. It says "reject when the mean is too high,
but the threshold that defines too high must be based on the variance of the
number of bacteria found per tube."
The last topic I want to consider is historical data regarding the level of
water-borne disease by community size,, The goal of quality control is to
.0011
.00055 •
m
<
m
O
cc
Q.
COMMUNITY SIZE (Log 10)
Figure 3. PROBABILITY ANY ONE PERSON WAS INVOLVED IN AN OUTBREAK
OF WATERBORNE DISEASE DURING 1946 - 197lO.
53
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
keep to a low level the probability that some person will contract water-borne
disease. In a community with a poor historical record for the incidence of
such disease we would have to be more stringent with the sampling rules in
order to keep the probability down to the same level for rail citizens in all
comknunities. . , , i .,.,„• , ,
The historical record is a confusing one. My last figure shows the proba-
bility that a single individual will be involved in an outbreak of water-borne
disease at some time in the time interval 1946-1970 as a function of com-
munity size (on a Iog10 scale). Small towns (3-5 thousand) and rather large
towns (50-100 thousand) have higher rates than towns of other sizes. A
possible explanation for this is that the very smallest towns don't need elab-
orate quality control mechanisms so they have a low disease rate. Larger
towns (the first peak in the curve) do need such surveillance but don't intro-
duce it until town size is near ten-thousand. Towns of 50-100 thousand popu-
lation need quality control of a higher order of sophistication but this isn't
to be found (the second peak) until town size approaches one-half million.
Of course, this reasoning is entirely a product of having seen the data first.
If we accept the odd behavior of this curve as representing reality we see
that there could be objections to a sampling plan which weighted the inten-
sity of sampling using historical information. From a political viewpoint, it
might be difficult to justify less sampling for towns of 10-thousand population
than for those which are either smaller or larger.
After all this, please note that the maximum value of the line indicates a
probability of less t'han one part in a thousand over a 25 year span.
I urge you to remember that sampling and quality control are problems to
which the statistician can make a major contribution and his advice should
be solicited regarding any proposed regulations.
i
iESSl
QUESTION AND ANSWER SESSION
S. M. Morrison, Department of Microbiology, Colorado State University,
Fort Collins, Colorado
How does your model take into account heterogenity with respect to small
and large towns that have their own treatment and distribution systems?
L. Muenz, National Cancer Institute, National Institutes of Health, Bethesda,
Maryland
It doesn't do so _in any explicit way. The coefficient-of -variation parameter
can be adjusted to agree with historical information from a town of a given
size but this sort of adjustment is not built into the model; it's at the discre-
tion of the user. A more complicated model would already have the com-
munity size-variability relation explicitly included although I wonder how
useful such a mixture of mathematics and survey information would be.
P. D. Haney, Black and Veatch, Kansas City, Missouri
', , ' " ',-.'• :." ..... " ...': ..... ;,:, ..... . ; , , i ' . , " ..... :;
In the 1940's the U.S. Public Health Service published a graph giving the
relation between number of water samples/month and population. J. K. Hos-
kins, who was Assistant Surgeon Genera.! at the time said that the mathemat-
ical derivation wais so complicated that they only published the curve arid
not the mathematics, Perhaps you could "track"' this down.
54
jjiii^^ :,.,. i.:,,,!!,:,!,.,'.,:,it
-------
STATISTICAL CONSIDERATIONS/L. MEUNZ
L. Muenz
My plot shows the number of samples/month/1000 persons as a function
of community size using -information from the E.P.A. interim regulations.
Assuming that this is the same number of samples/month as referred to in
the 1940's there is no way to ascribe a mathematical origin to this curve. The
demonstration of this is given by looking at the "ink blot" of points corre-
sponding to very low populations. This region is blown up in the insert and
a tick mark (') on the horizontal axis of the larger graph shows how far
the insert extends. The behavior of the insert is completely non-analytic
even if most of the larger graph seems rather smooth. No conceivable mathe-
matical function could give rise to the jagged line appearing in the insert.
Thus the whole curve is likely to be empirical; this, of course, doesn't mean
that it's bad, just that it would be difficult to defend.
55
-------
:l
Mi!*!/ I
I, i
-------
CHLORINE RESIDUAL SUBSTITUTION/L.J. MCCABE
CMorine Residual Substitution — Rationale
Leland 3. McCabe
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
Consideration of the use of chlorine residual as an indication of the safety
of a drinking water supply was much easier before the discovery of the
chloroform problem. In the past a chlorine residual was an unquestioned
good. Now the value of the residual must be determined so that trade-off
considerations can be given to the possible increase in chloroform and the
other trihalomethanes. Both toxicologies! and epidemiological studies indi-
cate an adverse effect of chloroform ingestion or the consumption of water
that has been chlorinated.
The value of carrying a disinfection residual to overcome contamination
throughout the distribution system of a community water supply has not
been determined. Johns Hopkins University now has a research project to
determine the effectiveness of the residual to contain introduced contamination
to a distribution system (3). The value of the residual in a special situation
has been reviewed. It is believed that residual chlorine is a substantial safety
factor on shipboard distribution systems but credible evidence is not avail-
able to insist that such a residual be maintained (2).
Besides the safety factor that a disinfection residual may provide to a
distribution system there 'is another benefit that must be considered. If the
water supply operations is set up to maintain a residual, the absence of the
residual can quickly signal the inadequacy of treatment or the introduction
of foreign material into the distribution system. Thus, the measurement of
a disinfection residual can provide a continuous-real-time monitoring of the
safety of the distributed drinking water.
To control the spread of waterborne disease, information on water quality
must be available in time to influence decisions concerning the use of the
water. Microbiological methods are not adequate as they only provide his-
torical data on the quality of water severail days -ago. Microbiological methods
can be very helpful in the investigation of the causes of waterborne disease
outbreaks but they do not lend themselves to monitoring.
The need to rethink what we have been calling water quality monitoring
or surveillance was demonstrated in the Community Water Supply Survey
(4). The survey design called for sampling all water supplies in one metro-
politan area of each of the Public Health Services regions. A late change
was made in Region I to include the State of Vermont instead of New Haven,
Connecticut. Some virus studies were being conducted on the New Haven
supply and it was thought best not to include this supply as part of the
national survey. Nationally, it was found that at 85 percent of the water
systems, bacteriological samples were not collected at the rate prescribed in
the drinking water standards. Table 1 is a copy of the table from the report
of the survey. Samples were collected at less than half of the prescribed
frequency at 69% of the supplies surveyed. The prescribed frequency is
admittedly an arbitrary number but when efforts were made to examine the
required frequency the established frequency was universally supported by
the state sanitary engineers in 1968. The results of the bacteriological sam-
pling are shown in Table 2 and it can be seen that the problem of poor
57
-------
in
09
n
TABLE 1. BACTERIOLOGICAL SURVEILLANCE EVALUATION BY COMMUNITY SIZE.
Population Served in Thousands
(All data are per cent of System Totals)
Overall System
Totals
=.5
.5-5
5-10 : 10-25 25-50 50-100 >100 Number Per Cent
Met bacteriological
surveillance criteria
Did not meet bacterio-
logical surveillance
criteria
Did not collect
samples at the
rate in the bac-
teriological sur-
veillance criteria
Systems totals—number
18 12 10
12 36 100 10
96 82 88 90 97 88 64 869 90
94 74 83 85 92 88 64 827 55
446 315 75 59 36 16 22 969 —
Note: 60 per cent of the survey population were served drinking water by
systems over which a bacteriological surveillance program met the criteria.
w
o
f
S3
I
p
I.
-------
CHLORINE 1RESIDUAL SUBSTITUTION/L.J. MCCABE
TABLE 2. WATER SUPPLY SYSTEMS IN VARIOUS POPULATION
SIZES WHERE DELIVERED WATER EXCEEDED THE COLIFORM
DENSITY LIMIT.
Population Range
^500
501-5,000
5,000-10,000
10,001-25,000
25,001-50,000
50,001-100,000
> 100,000
Totals
Number of
Systems
446
315
75
59
36
16
22
969
Number of
Systems
Exceeding
79
29
5
5
1
1
0
120
Per Cent of
Systems
Exceeding
18
10
7
9
3
6
0
12
Per Cent of
Survey Population
Exceeding
15
8
: 7
7
3
6
0
2
quality water is greater with the smaller supplies. Something needs to be
developed that can address that problem.
The usefulness of the chlorine residual to maintain the water within the
bacteriological limits of the standards was also demonstrated in the CWSS.
Although the practice of chlorination caused a dramatic decline in the fre-
quency of the samples containing colifoims, these organisms are not eliminated
merely by claiming that chlorination is practiced. Unless a chlorine residual
was maintained in the distribution system, a significant portion of the samples
from the distribution system contained coliforms. The coliforms are elimi-
nated nearly completely in systems providing at least a trace of chlorine
residual (1).
Buelow and Walton (1) also analyzed Cincinnati's water surveillance data,
during a period when a free chlorine residual was being maintained in the
plant effluent. This is illustrated in Figure 1 when the free residual was car-
ried in 1969-1970. All monthly average coliform counts were below 0.1 per
100 ml. and in most cases the samples with coliforms were collected from
locations without chlorine residuals.
If the reality of these data is faced, it is obvious that something new must
be done in the way of improving bacteriological surveillance. The idea of
substituting the chlorine residual for coliform testing emerged from con-
sidering the results of the Community Water Supply Survey. The concept was
outlined by Robeck (5) but he cautioned that five specific problems should
be addressed before the substitution is allowed for a water supply.
1. The number and location of samples for which chlorine residuals are
to be substituted.
2. The form and concentration of chlorine residuals to be maintained.
3. The frequency of chlorine residual determination.
4. The analytical method to be used.
5. The past bacteriological history of sampling location.
The problems were considered and paragraph(h) of Section 141.21 of the
National Interim Primary Drinking Water Regulation (6) allows for _the
substitution of chlorine residual for 75% of the microbiological sampling.
At least 4 chlorine residuals are substituted for each microbiological sample
dropped but the residual must be measured daily. The benefit of the use of
this provision to the safety of the smaller water supplies is obvious. Daily
monitoring of a water supply would not be achieved by the coliform test,
59
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
I
o
u
a.
e
8
u.
Id
1.6
K
A
.3
.2
.1
0
JA
N.
iru
—
f
_
_j_r
1
12 MONTH AVG. MF COUNT/100 ml
1966 0.11
1969-1970 0.02
_n _T~~^ Ln-J-T-r-1
MAR. MAY JUL. SEP. NOV. JAN. MAR. MAY JUU SEP. NOV. *MAY JUL. SEP. NOV. JAN. MAR.
1966 1967 1969 1970
Figure 1. MONTHLY AVERAGE COLIFORM COUNTS PER 100m1
unless the supply served 25,000 people. For a population served of 25,000 to
28,000, 30 microbiological samples are collected per month.
If the rules are followed the chlorine residual procedure will assure much
quicker action in correcting a problem of unsafe water. "When a particular
sampling point has been shown to have a free chlorine residual less than 0.2
mg/1, the water at that location shall be retested as soon as practicable and
in any event within one hour." Thus, the operator has one hour to increase
or restore the chlorine application and make sure the residual reaches the
sampling point. With coliform testing, considering sample transit times, time
to perform the test, and time to make the notification, the reaction to bad
water might be half a week's delay. A good example of the response follow-
ing the loss of chlorine residual was the investigation of a cross connection
in Tampa, Florida (Communication from Hillsborough County Health De-
partment to the State Division of Health, November 27, 1973). A pressure
sewer had been mistakenly connected to the water main and the location of
the problem was accomplished with the chlorine residual testing.
Chlorine residual testing could have been required in addition to the re-
quired microbiological monitoring but to promote the use of the more effec-
tive monitoring it was decided to eliminate the more costly microbiological
testing. Savings effected can be used for other water quality testing. Responsi-
bility for first line quality control must rest with the water works operator.
The present system of collecting limited samples at small water supplies and
sending them to the state laboratory for testing introduces a mystical aspect
tp quality control. The operator is not involved personally and reports come
back a week or more later that may indicate deficiencies in the past. When
the operator determines a chlorine residual and does not find a pink color
' ' ' , , Ml , . , , ,, I | I ||
60
-------
CHLORINE RESIDUAL SUBSTITUTIQN/L.J. MCCABE
(DPD test), he is the only one to know there is a problem and he must make
a decision to protect the public's health. The responsibility for an acute
problem is where .it should be and at the only place where necessary correc-
tive action can be taken. The publication of the causes of waterborne out-
breaks must be intensified to reinforce the responsibility of the water-works
operator in the prevention of disease. This will require the reporting and
investigation of outbreaks and this activity can be supported with EPA's
water supply program grants. The state laboratory microbiolbgist should no
longer be overburdened with routine bacteriology samples and should be able
to participate in outbreak investigations to determine etiological agents and
their source. Table 3 shows a tabulation of waterborne disease outbreaks by
agent. The large percent attributed to the nonspecific category gastro-
enteritis points out the need for more extensive microbiological workups
during outbreak investigation.
Section 141.21 paragraph (h) requires the collection of coliform samples
when the chlorine residual cannot be restored within one hour. This sample
has a great likelihood to contain collforms but it is to be averaged in to
indicate the monthly means for compliance with section 141.14. This pro-
vides a much stricter level of quality for a supply using the chlorine residual
procedure. For example a water supply serving 5,000 population is required
to collect 6 microbiological samples and may substitute the chlorine test for
4 of them. Thus 2 coliform samples would be collected per month but sur-
veillance would be provided daily by the chlorine residual test. If the residual
were low on one day of the month and the follow up coliform count was 4
per 100 ml the system would not meet the drinking water standards even
though the supply likely had good quality water all but one day of the
month. To promote the use of the chlorine residual substitution, which pro-
vides really a better monitoring system, the days with adequate chlorine
residual should be considered zero coliform days .in computing an average.
TABLE 3. WATERBORNE DISEASE OUTBREAKS
DISTRIBUTION BY ETIOLOGY.
1946-1974
Gastroenteritis
Infectious Hepatitis
Shigellosis
Chemical Poisoning
Giardiasis
Typhoid
Salmonellosis
Amebiasis
Poliomyelitis
Enteropathogenic E. coli
Tularemia
Leptospirosis
Number of Outbreaks
Community
• - 'Water Systems
52.6
16.3
9.6
5.9
5.2 - .
4.4
4.4
0.7
0.7
0.0
0.0
0.0
135
Other
\Vater Svfitems
TT **iw|. kJj^utiJ.ji3
47.5
13.7
10.2
4.0
2.5
15.8
2.8
1.2
0.0
1,2
0.6
0.3
322
61
-------
f! P *
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
Recommendation
I would think that complete chlorination substitution would be usable for
the smaller supplies. If this were allowed for supplies up to 9400 P°P^ion
served it would also take care of the averaging problem with a small number
of microbiological samples. A bacteriological sample is required when the
chlorine cannot be restored within an hour. The operator of the system wou d
of course continue to work to get the residual restored and it will faave^o
be assumed that .this will be done as diligently as possible When he getsthe
residual back up to the prescribed level the public is again protected. What
WoSd be the best microbiological test to perform on this sample that was
Taken before the problem was corrected? A sample for Cl^dmm perfr^es
would seem to provide data on what was needed to be known and the sample
could be just mailed to the laboratory without a concern **««"«***;*
these organisms are more resistant to chlorine and should be ubiquitous
enough to indicate when polluted water enters the distribution system.
For large water systems the full benefits of the chlorine -residua should.be
promoted Continuous monitoring is possible with feedback controls o adjust
The treatment and give alarms when departures occur. In these systems the
microbiology applied to samples collected when the residual chlorine cannot
™e restoredI within one hour should be rather complete. The testing should
assist in determining the source of contamination and ought to be more
elaborate than just a coliform test.
REFERENCES
1 Buelow, RiW., and G. Walton. 1971. Bacteriological quality vs. residual
2 CoesCorol. 1977. Final report of the drinking water
Section ad HcS Advisory Committee. Department of Health, Education
and Welfare, Atlanta, Georgia. ,
3 Kruse CW 1976. Grant Application R804307— Biological evaluation of
' the benefits of maintaining a chlorine residual in public water systems.
US Environmental Protection Agency, Washington, D.C.
4. McCaS, L.J., J.M. Symons, R.D. Lee, and G.G. Robeck. 1970. Survey of
community water supply systems. JAWWA 62:670-687.
5 Robeck G G. 1974. Substitution of residual chlorine measurement tor
SSbution bacteriological sampling. Proceedings AWWA Water Quality
Technology Conference, Dallas, Texas. ....
6. United States Environmental Protection Agency. 1975 .National interim
primary drinking water regulations. Fed. Register 40:59566-59574.
QUESTION AND ANSWER SESSION
R. R. Colwell, Department of Microbiology, University of Maryland, Col-
lege Park, Maryland
How can chlorine residuals supply information concerning actual bacteria
populations in drinking water? If chlorine is that effective in the treatment of
pathogens in sewage, why do we need to chlorinate the drinking water if we
do it to sewage?
"U,« I", !•(! •
::• .hi,, Sim,:, , 'Iff"
62
''Jr'.'iWJI H!*ifc'.; tW;ii.t!(>'
n' H1'™
-------
CHLORINE RESIDUAL SUBSTITUTION/L.J. MCCABE
L. J. McCabe, Health Effects Research Laboratory, U.S. Environmental
Protection Agency, Cincinnati, Ohio
The chlorine residual test is not supposed to give a quantitative evaluation
of the bacterial population. When the chlorine residual exceeds the specified
amount it is an immediate qualitative evaluation that pathogens are not
present.
Pathogens are also derived from non-point sources that are not treated. It
will be a great day when all sewage is disinfected adequately and with reli-
ability, until this Utopia, it would seem desirable to disinfect drinking water.
A residual disinfectant provides some measure of protection in distribution
systems.
W. Ginsberg, Water Purification Laboratories, Chicago Bureau of Water,
Chicago, Illinois
Do you believe water treatment plant operators are likely to take four
times the number of chlorine residual samples when they often do not take a
sufficient number of coliform samples now?
L. J. McCabe
We want to provide to the operator the most effective test for him to use.
Their use of any test is a matter of personal modification and; enforcement. I
think these would be enhanced with a. chlorine residual measurement. In the
case where neither test were used, the chlorine residual requirement would
waste less money because it would not be necessary to have an unused micro-
biology capability standing by. With the use of the chlorine residual, the
microbiological capability can be used for more in .depth analyses of the most
meaningful samples and be ready for waterborne disease outbreak in-
vestigations.
63
-------
1. <-!' ' i ii
-------
MICROBIOLOGY OF POTABLE WATER/R.R. COLWELL
Public Health Considerations of the Microbiology of
"Potable" Water • • • '•
R.R. Colwell, B. Austin and L. Wan x
Department of Microbiology
University of Maryland
College Park, Maryland 207.42
1 Department of Civil Engineering
Howard University
Washington, D.C. 20001
Water is essential for every biological function and the convenience of ob-
taining a ready supply of drinking water from municipal supplies is accepted
without question in civilized societies. In fact, the occasional odor of chlorine
when a tap is turned on has been interpreted by the layman as a guarantee
that the water has been purified and is, therefore, safe to consume. Unfor-
tunately, there are problems, other than microorganisms, as, for example,
taints, odors, floccules and disturbing colorations^ which cause temporary
distress and, perhaps, a telephone call of complaint to the supplying agency;
otherwise, in North America water is available and used without further
thought.
Natural watejrs_which flow in rivers and streams, and settle in wells and
lakes, Contain a Jffetincjnini^^
un^lssolYed^jjiOT^^ micror
flo7ararF"potential pathogens and, therefore, hazardous lo the health of the
ctSnTBTImi^whereas ioAer__jmcrop^uiis^s_areJjeneficial in mineralizing
dfganic matter, concentrating heavy -metals in tficMlng~~filte"rs, and 'acting"
~~'
function of "water supply agencies to provide "clean" water, in quantities
necessary to cope with the demand of the community. This water is collected
from rivers and wells and sometimes from estuaries or the sea. The water
that is collected is treated with flocculants. e.g. alum, to precipitate suspended
material, filtered through sand beds to remove precipitates and floes of living
organisms, and chlorinated Jo_gliminate potentially harmful organisms. Addi-
tional compounds often are added to "the. water (such as~3iuprides; and the
finished water is distributed to storage reservoirs or through1 pipes to con-
sumers. The water is not •microbiolo.gicallv sterile since spores and many
Gram-negativerods survive chlorination, and it is not pure, as fluorides_and
other compoundlTmav be added. However, thT quality ot tn¥ finished water
'should be safe for_drinkin~a_and-usa^. in.hospiMs.^phannacies and industry.
Periodically, samples of water are collected from ffiecfistribution system and
chemical analyses done, but, unfortunately, rather primitive techniques are
used to assay microbiological quality. The microbiological analyses are recog-
nized to be time-consuming and may provide questionable information, par-
ticularly since the few bacteriological tests that are most commonly done are
performed by inexperienced or inadequately trained personnel. Very rarely
is a registered microbiologist employed to carry out the bacteriological tests.
Increasingly, problems with drinkiing water are being brought up for
public discussion. Setting aside the problems of toxic chemicals and speaking-
only to microbiological criteria for waiter quality, there are several problems
that need to be addressed, as, for example, the difficulties in .establishment
of standard laboratory procedures, standardization of , laboratories, and
development of criteria for qualification of professional environmental and
65
-------
ff
:* M. " . !
Ji H1,, :• :
if-!!
f ft
4:'
"Si1:"
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
sanitary microbiologists. The major problems are, clearly, (1) standardiza-
tion of personnel and training; (2) standardization of laboratories carrying
out the analyses; (3) standardization of reporting results and banking of the
data for coordination and intercalibration around the country; and (4) stand-
ardization of media and reagents^ as well as procedures for sample gathering,
plating, culturing, reading tests, interpretation of test results, and recording
of data. Clearly, the need for professionalization of water testing and water
quality determination is most urgent. Papers in this series deal with this prob-
lem and it can be seen that there are solutions, but that effort, time and
money wiilbe required. . .
The microbiological standard for drinking water has not changed in its
essentials since 1914 (10), and is determined by reference to the presence
of coliform bacteria. The definition of coliform is vague and includes many
asporogenous, Gram-negative rods (21). Their isolation from water is de-
pendent upon the antiquated Most-Probable-Numbers Technique (1) or by
the newer mebrane-filtration method (1). The presence of "coliforms" in
numbers greater than 1/100 ml (27) is taken to mean that the water is con-
taminated. The solution is often re-chlorination of reservoirs and pipe water,
and the problem, thereafter, quickly forgotten. It takes at least 48 h for
coliforms to be recognized by the standard procedures employed (23), .by
which time the contaminated water will have been distributed and consumed
by many individuals in the community. Thus, at best, the bacteriological
methods presently employed provide only an indication that trouble occurred
in the system and the event may have been 2 or 3 days ago, hardly to be
considered satisfactory water quality control. Thankfully, water-borne infec-
tions are rare in the U.S.A., even though the potential exists for dissemina-
tion of disease-producing.microorganisms. Thus, modern microbiology tech-
niques should be incorporated in routine examination of water. Gas-liquid
chromatography appears to offer immediate detection of bacteria and bac-
terial products by direct analysis. Methods such as pyrolysis and laser scan-
ning also offer potential value in direct assessment of water quality. The
application of molecular genetic techniques to water bacteriology appears to
be almost totally lacking. Reasoner (23) correctly stated that "the develop-
ment: of quantitative procedures for pathogenic bacteria has lagged behind the
development of quantitative procedures for non-pathogenic bacteria." Un-
fortunately, it can also be stated that water bacteriology has not advanced to
a degree commensurate with that of other aspects of microbiology, most
notably of microbial genetics and physiology. If the aim of water quality
determination is to locate the presence of pathogens or indicator bacteria,
then it is essential not only to be aware of other types of organisms occurring
m water, but also to understand their physiology, ecology, and role in public
health.
Bacterial Microflora in Treated Water
Although it is relatively easy to remove some microorganisms from water
by filtration and chlorination, it is difficult to reduce the_contgn^ of organic
material to a level below which it w2LJMJl_supparJ_mic]X>^^
^ „_-.- -, means of disinfectants are generally
ineffective and invalidate the description of pure ^water(13JL-The viable,
aerobic, heterotrophic bacteria which can be routinely isolated from treated
water include Acfiromobacter, Arthrobacter, Bacillus, Clostridium, coryne-
forms, Flavobacterium, Gallionella, Klebsiella, Leptothrix, Micrococcus, My-
rcobact.erium, Proteus. Pseudomonas and Serratia (4, 26, 29, 30, 32). These
66
-------
MICROBIOLOGY OF POTABLE V
rATER/R.R. COLWELL
survive inadequate treatment or enter via contaminatior
within the distribu-
tion system. Although none of these organisms are primary pathogens, some
may be responsible for taints and odors, spoilage of commercial products
for which water is used, and also may act as secondary invaders in human
disease processes.
Pseudomonas and Flavobacterium are problematic genera, some species of
which are opportunistic pathogens (14). The latter genus, a: heterogeneous
taxon of convenience (3, 20), comprises a variety of sp;cies, some of which
are tolerant to high levels of chlorine, with certain strains resisting as much
as 100 mg/L. Pseudomonas and Flavobacterium spp. have been noted to
cause heavy growth in the sprinkler heads of scrub siiiks and water baths
in hospital operating theatres and preparatory areas (15|). Yellow-pigmented
rods, described as flavobacteria, have been isolated frota patients with sys-
temic infections following heart-lung surgery. One species, F. meningo-
„ - I* *_ _A1_4-».«.!*•* MAnlnt-n w£ +rt rt *"»*l! !•». « J-»+-I rtf /* O O ^ rt +*A 1 t 4-CT »-»»*£lCie» »*»^-»O P M/%1 t I ft
septicum, is relatively resistant to antibiotics (22) and
not be tolerated in treated water. The dangers resulting f;x>m bacterial metab-
olites, such as pyrogens, should not be overlooked, anc in one study (24)
it was demonstrated that water used for pharmaceutical
viable bacteria producing pyrogens. The problems caused by bacteria in
water extends to scientific laboratories where, deionized '
have been shown to contain pigmented Gram-negative
bacterial growth occurs in the resin columns, with the
comprising as many as 10s cells/ml (2:5). Surely, meth
its presence should
purposes contained
and soft water units
rods. In particular,
bacteria population
ids mtist be devised
"to ensure the sterility of all water that is to be used as drinking water.
Purification ol Water
The effectiveness of chlorination justifiably can, be criticized because of
the resistance of many viruses and bacteria to chlorine treatment. Treatment
with ozone is considered to be an alternative method of water treatment
since ozone is bacteriocidal and removes taints and oc.ors without leaving
unpleasant residues (11). Methods not presently employed in water treat-
ment, but of potential'Value, include heat sterilization a; id membrane filtra-
tion, both of which can be adapted for purification of large volumes of
water. Reverse osmosis and ultrafiltration are perhaps
and effective in reducing bacterial numbers (6). In all
claving and ultrafiltration, viable organisms remain in ft
ment and this fact raises several questions (13):
(a) are the microorganisms remaining in drinking
tional treatment merely harmless indicators of an
(b) are they beneficial by assisting in the breakdown
. pounds transported by the water?
(c) are* they a hazard to the system, the user or as
end product of industrial processes intended for i
by humans?
the most feasible
cases,; except auto-
e water after treat-
yater after conven-
unsterile condition?
of organic corn-
contaminants in an
ingestion or injection
Microbiological Problems with Water Treatment and Distribution
McCabe and co-workers (19) found that, of 969 public:water supply
systems serving 18.2 million individuals, 613 utilized grpumd-vyater systems,
67% of which distributed water to 2.8 million consumers without any disin-
fection. Of 621 wells, 9% produced water that contained cbliforms, 2%
contained fecal coliforms, and 11% of the wells dispensed water with >
: 67
-------
III I'.j I ,1 II
in ::R I- - iii
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
500 heterotrophs/ml as determined by plate counts. How many pathogens,
particularly organisjms causing gastro-enteritis, were distributed in those
waters can only be speculated. . .
It is particularly disquieting to realize that even water in hospitals is
far from pure. In one interesting study (18) it was recorded that as much
as 30-40% of water contained in so-called sterile tanks and pipes serving
operating theatres derecbnt~ammate
-------
MICROBIOLOGY OF POTABLE WATER/R.R. COLWELL
before a sanitary problem is detected, when it may already be too late for
action to avoid health hazard.
Tap Water
Bacteriological examinations of tap water frequently reveal the presence
of large numbers of bacteria. Goliforms and fecal streptococci occur infre-
quently in chlorinated water, and from their presence is inferred 'recent'
contamination.
In a bacteriological study of tap water undertaken in our laboratory,
(Table 1 ) , it was shown that samples collected in the Washington, B.C. area
were negative for coliforms, as determined by incubation for up to 7 days at
35 °C in lactose broth. However, after this extended incubation time the
plate counts on glucose tryptone yeast-extract agar (GTYEA) were alarming-
ly high, although lower than for bottledjwate£-gP^|3III5fler incubation
for only 48 h these plate counts were very low, yet the prolonged incubation
period at 25 °C was necessary for development of the majority of the bac-
teriai~c"olomes. ihis trend was repeated, in parr, on iiosin-MetnyIene~-Blue
agar (JtJMBjTwhich allows the selective growth of coliforms. Several samples
(Table 1) contained /3-haemolytic colonies on blood agar incubated aerobic-
ally, and anaerobically under hydrogen;. Thus, it is recommended that plate
, the incubation periods be increased to' 7-14
of bacteriological media be employed in routine
examination of water..
Drinking Fountains
The convenience"of drinking ^fountains for providing refreshing, cool
water is accepted as part of everyday life, yet the bacteriological standard
of the water at public drinking fountains is low. The drinking fountain head,
unless regularly cleaned, is often dirty. In 1965, 78% of drinking fountains
tested (12) were contaminated, yet in 1975, 76% were in a similar poor
state. It is interesting to note that a wide range of microorganisms were re-
covered from the fountains. Although it is assumed that drinking water al-
ways contains a residual chlorine level, the water from the drinking foun-
tains tested by Herman (12) was found to be devoid of the disinfectant. This
may be explained by the separate water delivery systems for drinking foun-
tains, in which the water is cooled and constantly circulated, with the result
that the chlorine level usually falls below the effective level of 0.6 mg/L,
thus permitting a diverse range of microbial growth at the air-water interface
on the fountain head. Clearly, the bacteriological quality of water at com-
munal drinking fountains needs to be constantly monitored and, where
necessary, the water delivery systems modified.
Bottled Water
Bottled water has gained popularity in the U.S.A. with annual sales ex-
ceeding $150 million (31). In 1972, there were an estimated 700 water
bottling plants (8), with California the largest customer. The quality
standards of bottled waters varies between states, with some complying with
the Public Health Service Water Standards (22), whereas others have lesser
or no standards. However, from the bottle labels it is often implied that the
quality is superior to the municipal water supply, and that this quality will
69
-------
TABLE 1. BACTERIOLOGICAL EXAMINATION ;OF TAP WATER OBTAINED FROM THE DISTRIBUTION SYSTEM
WITHIN THE WASHINGTON, D. C. METROPOLITAN AREA"
. ; No. of colonies/ml
Presumptive • '
colifonns/
100 i
Source after
U.S. Capitol . .
U.S. Supreme Court
Washington-Suburban Sanitary Station
Alexandria City Hall
Fairfax County Offices :
Montgomery County Offices .-
Prince Georges County Court House
nl
7 days
0
0
0
0
0
0
0
GfYEA"
2 days
5
20
0
0
5
10
0
7 days
7.8X102
1.4X104
1.1 XlO2
7 XlO3
4 XlO3
50
0
EMB°
2 days
0
40
0
0
0
0
0
7
4
9
1
days
60
XlO3
5
XlO2
.2X10S
0
0
Hektoen
Agar
7 days
0
0
0
0
0
0
0
Sheep Blood Agar
(Aerobic) (Anaerobic)
2 days
5
30
0
0
0
20
0
7 days
10
30
5
0
0
30
0
3 days
40
70
0
0
0
60
0
a Water was allowed to run for.2-3 min before collecting samples in sterile botfles. Inoculated media, except for the presumptive coliform
test, were incubated at 25°C for the time periods indicated.
"Glucose tryptone yeast-extract agar (Difco).
c Eosin methylene-blue agar (Difco). .
o
I
8
a
1
-------
TABLE 2. BACTERIOLOGICAL EXAMINATION OF BOTTLED WATER. INCUBATION OF MEDIA, EXCEPT THE
PRESUMPTIVE COLIFORM TESTS, WAS AT 25°C.
No. of colonies/ml
Presumptive
coliforms/ GTYEA °
100 ml
Source (Brand name) after
Aqua Pure0
Deer Park
Great Bear
Mountain Laurel
opting vvaicrui9
Mountain Valley
Mineral Water
Poland Water obtained from :
Grocery Store
Manufacturer
U.S. Supreme Court
7
0
0
0
\j
0
0
0
0
days 2 days
10'
4.4X10°
5 XlO2
u
0
4 XlO6
6.6X105
1.8X10°
7 days
102
con"
1.5X108
JU
0
con
con
con
EMB"
2 daysj |
0
2 XlO5
6.3 XlO5
6 A1U
0
5 XlO4
3.5X104
1.8X104
7 days
0
con
6.6X105
z xiir
0
7 XlO*
3.5X10*
1.8X10*
Hektoen
"Pnfpru*
Agar
2 days •
0
5.1 XlO3
0
u
0
0
0
0
7 days
0
6.4X10°
0
u
0
0
0
6 XlO2
Sheep Blood Agar
(Aerobic) (Anaerobic)
2 days
50
2xlOa
3 XlO3 .
i.axiir
0
1.7X10*
1.5X10*
3.7X10*
7 days
50
con
5.4X104
J..JXHT
0
1.8X104
1.5X10*
6.4X10*
3 days
0
1.1X10"
3 XlO3
6 XlO-
0
0
0
1.8X10°
§
o
R
o
S
w
w
*fc>*
a Glucose tryptone yeast-extract agar (Difco)
" Eosin methylerie-blue (Difco)
0 Pasteurized
d con = confluent growth
"From hot springs, >90°C
* Refrigerated
8
-------
'I!,,11" 'I H,1 lift!
'ffl i
it
MICROBIOL STANDARDS EVALUATION/CW. HENDRICKS
be maintained throughout the shelf life of the product. Unfortunately, this
is usually far fronji true, and the bacteriological quality of freshly bottled
water varies from [brand to brand (9). Furthermore, the bacterial count
tends to increase With storage. The usefulness of the bottled water standards
adopted by the Fo'od and Drug Administration, which became effective on
May 22, 1974 (28 j, is questionable. They do not include a plate count limit
for bacteria, but tjbie coliform limit of less than .2.2 organisms/102 ml is
similar to the dririking water standard. It is a help th.at bottled water is
categorized as a fiod, but it is a pity that bottlers often recommend it for
preparing baby food, pharmaceutical prescriptions and hot and cold bev-
erages! Thus, certain members of the community, namely the young, aged
and infirm, may bz more susceptible to infection by opportunist organisms
which could occur in the water.
In a study undertaken in this laboratory, the bacteriological quality of
bottled water obtained in Washington, D.C. was examined (Table 2). Al-
though the coliform numbers, as determined by the production of acid and
gas (in lactose brbth) within 48 h, were negligible; in some cases there
were large numbers of aerobic heterotrophs as determined from plate counts
j_ ~_ c-v-r'Vrj A .^^^iirv, TV.Q einnififant nninf tn hp marlfi IS that. CXCeOt
medium. The significant point to be made is that, except
coliform tests, the test media were incubated at 25°C.
37°C, little or no growth was obtained. Thus, it is mis-
made on GTYEA
hi the case of the
When incubated at -*/ •»-', ***.».*^ v* *+~ e,~—..~ ..—
leading to enumerate bacteria in water by incubating the inoculated test
media at 37°C.
It can be seen (table 2) that pasteurized water, i.e., Aqua Pure, and re-
frigerated mineral [water, was of good bacteriological quality; however, the
numbers of bacteria present hi other brands was disturbingly high. These
data were obtained several years ago (Washington Star, Washington, D.C.,
July 11, 1971). Monitoring of the microbiological quality of bottled water
should be done on! a regular basis to ensure the public health safety of this
product. Of particular concern was the large number of /Hiaemolytic colonies
which developed oln blood agar. These colonies may reflect the presence of
potentially harmful organisms in the water, and the numbers greatly exceed
those found in tap! water (Table 1). Furthermore, it is readily obvious that
the bacterial counlts for stored, bottled water, viz., Poland water obtained
from a grocery st|ore and from the U.S. Supreme Court at the tune the
study was carried put, were higher than in water obtained directly from the
bottler. Clearly, long-term storage of bottled water can have an adverse effect
in providing for gjrowth of bacteria. It would seem appropriate for bottlers
to either pasteurize the water, or require refrigeration during storage.
Recommendations
Too little is known about the microbial ecology of water at source, treat-
ment, distribution iuid storage. It is essential that the normal microbial flora
of drinking water lie regularly monitored, and the taxoriomic composition of
the microorganisms present in source water" be determined on a regular
basis. Thus, seasonal variations in the flora can be measured so that altera-
tions in the flora lan be recognized to be abnormal increases in the -total
number of microorganisms or selective increases in a few taxa due to season-
ality.
It is desirable not to have any viable organisms in drinking water, and
methods to improve the treatment of water, including removal of viruses,
, i ",,, * I ;; ; M | „„,,,• :
i ' '
-------
MICROBIOLOGY OF POTABLE WATER/R.R. COLWELL
and maintain its purity until consumption, should be employed in water
treatment systems.
Methods presently employed for the examination of water are inappropri-
ate and give a poor idea of water quality. It is important to examine the
effectiveness of a wide range of media, such as blood agar and low nutrient
media, pre-treatment at low temperatures, and of using a wide range of in-
cubation temperatures, especially low temperature incubation of 20-25 °C
for enumeration of total viable bacterial populations. Thus, psychrophiles,
mesophiles and thermophiles should be considered because of -the variety
of conditions and uses for drinking water. Non-selective enrichment tech-
niques have been successful in the recovery of Salmonella spp. (17) and
Vibrio cholerae (5) from the Chesapeake Bay, and these methods may be
effective in determining the sanitary quality of drinking water. It is im-
portant to remember that no single technique is suitable for the isolation of
all pathogens; therefore, a combination of methods is recommended, with
the minimum being inclusion of total viable counts at 25 °C.
The concept of a residual chlorine level in drinking water needs to be
carefully scrutinized because of chlorine resistance amongst viruses and many
bacteria. Although chlorine-resistant coliforms may not occur, the presence
of Flavobacterium and Pseudomonas is undesirable and represents a potential
health hazard. Examination of drinking water for the presence of Pseudo-
monas should be incorporated in the routine analyses of the water since a
suitable method exists for the enumeration of P. aeruginosa (16). Also, other
Pseudomonas spp. are reasonably easily identified and enumerated.
The taxonomy of Flavobacterium must be radically improved, and isola-
tion procedures for pathogenic species of this genus should be developed..
The problems appertaining to antibiotic resistance and plasmid transfer
between taxa should be studied and the significance of these phenomena to
public health^safety determined.
In summary, the present methods for bacteriological analysis of drinking
water, although unsatisfactory in many respects, should not be discarded
in favor of the chlorine residual or other such chemical test. Rather, the
bacteriological testing should be improved, expanded and modernized. It is
never wise to throw the baby out with the bath water, and substitution of
the chlorine residual for sound microbiological analyses would do just
that!
REFERENCES
1. American Public Health Association. 1971. Standard methods for the ex-
amination of water and wastewater, 13th ed. American Public Health As-
sociation Inc., New York.
Bayliss, J.Ri 1930. Bacterial aftergrowth in water distribution systems. Water-
works and Sewerage 77:335-338. ' .'
3. Bean, P.G., and 3JR.. Everton. 1969. Observations on the taxonomy of
chromogenic bacteria isolated from cannery environments. J. Appl. Bacteriol.
52:51-59.
" •. Clark, P.M., R.M. Scott, and E. Bone. 1967. Heterotrophic iron-precipitating
bacteria. J. Am. Water Works Assn. 59:1036-1042.
5. Colwell, R.R., and J. Kaper. 1977. Vibrio species as bacterial indicators of
potential health hazards associated with, water. In Bacterial indicators/health
hazards associated with water. Am. Soc. Testing Mat. pp. 115-125.
73
&•
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRlCKS
6. De Roos, R.L., D.R. Vesley, A.G. Du Chene, and L.J.. Hart. 1976. High-
purity water supplies for biomedical research laboratories. Health Lab. Sci.
7. 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. Am. Water Works Assn. 64:596-602.
8 Geldreich, E.E., H.D. Nash, D.J. Reasoner, and R.H. Taylor. 1975a The
bacterial quality of bottled water. News Environ. Res. Cincinnati, U.S.,
9 GeldreichPE.E., H.D. Nash, D.J. Reasoner, and R.H. Taylor. 1975b. The
necessity of controlling bacterial populations in potable waters—bottled water
and emergency water supplies. J. Am. Water Works Assn. 67:117-124.
10 Haney, PJD. 1977. Evaluation of microbiological standards for drinking water.
•'••/„ Engineering Control Practices Technical Symposium, Office of Water Sup-.
ply. U.S. E.P.A. pp. 13-14.
11. Hann, VA. 1956. Disinfection of drinking water with ozone. J. Am. Water
' Works Assn. 48: 1316-1320. .
12. Herman, L.G. 1976. Sources of the slow-growing pigmented water bacteria.
/•;' Health Lab. Sci. 15:5-10. .
/(J3) Herman, L.G. The slow growing pigmented water bacteria: Problems and
VKsources. Adv. Microbiol. In press.
/(14 /Herman LG., and H.J. Fournelle. 1964. Flavobacteria: A water-borne po-
"V7tential pathogen. In Illrd International Congress of Chemotherapy; Inter-
national Society of Chemotherapy, pp. 1352-1555. G.T. Verlag.
15. Herman, L.G. and C.K. Himmelsbach. 1965. Detection and control of hospital
sources of flavobacteria. Hospitals 39:72-76.
16.*Highsmith, A.K., and R.L. Abshire. 1975. Evaluation of a Most Probable
Numbers technique for the enumeration of Pseudomonas aerugfnosa. Appl.
Microbiol. 50:596-601. ,„*,.•*
17. Kaper, J.B., G.S. Sayler, M.M. Baldini, and R.R. Colwell. 1977. Ambient-
temperature primary nonselective enrichment for isolation of Salmonella
spp from an estuarine environment. Appl. Environ^ Microbiol. 55:829-835.
! g Kelsey, J.C., and M.M. Beeby. 1964. Sterile water "Ibr operating theatres.
"The Lancet, July 11, No. 7350:82-84.
19. McCabe, L.J., J.M. Symons, R.D. Lee and G.G. Robeck 1970 Survey of
coramunity water supply systems. J. Am. Water Works Assn.. 62:670-687.
20 McMeekin, TA., DJB. Stewart, and J.G. Murray. 1972. The Adansonian
taxonomy and the DNA base composition of some Gram negative, yellow
pigmented rods. J. Appl. Bacteriol. 55:129-137.
21. Orndorff, S.A., B. Austin, LA. McNicol, and R.R. Colwell. 1978. Isolation
of lactose-fermenting Pseudomonas spp. from the aquatic environment.
Can. J. Microbiol. In press.
22. Pederson, M.M., MA. Marso, and M.J. Pickett. 1970. Nonfermentative
bacilli associated with man. Am. J. Clin. Pathol. 54:178-192.
23. Reasoner, D.J. 1974. Microbiology-detection of bacterial pathogens and
' their occurrence. J. Water Pollution Control 46:1395-1408.
24. Saunders, L., and E. Shotton. 1956. Water for pharmaceutical purposes.
J. Pharm. Pharmacol. £: 83 2-847.
25. Stamm, J.M., WJE. Engelhard, and J.E. Parson. 1969. Microbiological
study of water softener resins. Appl. Microbiol. 13:376-386.
H|k Tennessee Dept. Public Health. 1966. Epidemiologic studies of atypical and
^ acid fast bacilli in Tennessee. P.H.S. Grant No. CC00078; 133 pp.
27 United States Public Health Service. 1962. Drinking Water Standards. Dept.
Health, Ed. and Welfare, P.H.S. Pub. No. 956, Washington, D.C.
28. United States Food and Drug Administration. 1973. Bottled water quality
'standards: addition of fluoride and current good manufacturing practice
regulations. Dept. Health, Ed. and Welfare, Federal Register 38, 226:
r., 32558-32565.
\29\ Van der Kooij, D. 1977. The occurrence of Pseudomonas spp. in surface
V
|||;,; -''I , (;v _ ( ,,,-11 ;|'|^ , 45.187:197. ' " ; "; ;;
; , ' ' ii! f<,ii ' !' ' ...];•• ; • , • , ; , :' ' ,; "i ,: :
1 " ' :" 74
-------
MICROBIOLOGY OF POTABLE WATER/R.R. COLWELL
5^9. Victoreen, H,T. 1969. Soil bacteria arad color problem in distribution systems.
J. Am. Water Works Assn. 62:429-431.
31. Waters, E.P. 1974. What about bottled water? F.D.A. Consumer., U.S. Dept.
Health, Education and Welfare, Washington, D.C.
Willis, A.T. 1957. Anaerobic bacilli in a treated water supply. J. Appl.
BacterioL 20:61-64.
75
-------
llliillL |i|J
I ...... , ' :"!.
" Wf
It ,PK ' ' : "1
i!'! It:'.
f |
•;< .f "
I'lit, : t ft
1 II'NIII !„
,Lii-"« * i
-------
Water Analysis by the Membrane Filter Procedure
SESSION 2
Coliform Detection and
Control
Chairman: Mr. Edwin E. Geldreich, Water Supply Research Laboratory/
MERL, U.S. Environmental Protection Agencj
, Cincinnati, Ohio
77
-------
friiLL .......... :kliii;id£ ..... lili ........ li^ ........ k ..... ii
si, ..... i!i ..... kiih ..... Mi:* .......
J H' <;
:r'lj
-------
SAMPLING COLLECTING AND HANDLING/H.D. NASH
Sampling Collection and Handling
Harry D, Nash
Water Supply Research Laboratory/MERL
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
The ideal transported water sample is one which reaches the laboratory
with its bacterial flora unchanged, both qualitatively and quantitatively, as
it was at the time of collection. The type and degree of change is best
determined by analyzing the samples at the time and location of collection
for comparison of results obtained by the same procedural analysis after
transit. Unfortunately this action is impractical for a routine drinking water
monitoring program. Changes which may occur are dependent upon many
variables and combinations of the individual variables. Factors which in-
fluence the bacterial flora during transit are quality of the sample bottle,
bacterial population, density of each type comprising the population, concen-
tration of nutrients both organic and inorganic, fluctuations in temperature
during transit, turbidity, clumping and antagonistic action; all of these ex-
hibit their effect on the bacterial flora during the transit/storage period. These
factors can and do vary for every water sample, regardless of water classi-
fication as to source, surface, ground or potable water. Sampling locations
must be carefully selected to insure representative sampling, collection pro-
cedures must be conscientiously followed to prevent accidential contamina-
tion, and every attempt must be made to transport samples promptly to the
laboratory for processing. Failure to comply with these facts regarding sample
collection and handling will produce data which are confusing, misleading
and detrimental to any monitoring program.
Collection of Potable Water Samples
Proper collection procedures and techniques are fundamental to obtaining
reliable results for any water sample. Errors relating to sampling are usually
greater than those associated with laboratory analysis.
The laboratory is usually responsible for supplying sterile sample bottles.
Each bottle must be properly prepared with respect to the type water being
sampled with thiosulfate added to bottles for sampling chlorinated water
supplies,.or a chelating agent (EDTA) for waters containing metals, copper
and/or zinc. Quality control tests must be conducted on each batch of
bottles sterilized to assure sterility of all bottles. Sterile bottles should not be
stored more than 30 days to reduce possible contamination due to accidental
damage during storage. After sample collectors receive sterile bottles, it is
their responsibility to prevent accidental contamination during the collection
procedure. Bottles must not be carried in vehicles unprotected and certainly
not for several days to even weeks. Samples should be collected from a distri-
bution system tap which does not have any attachments. Taps which have
anti-splash devices and/or home purification attachments should be avoided
or removed before sample collection. It should be known that the tap is
supplying water from a service pipe directly connected with the main and is
not served from any type of storage tank (1,2,5). The tap should be opened
and water allowed to run to waste for 2 to 3 minutes, or for a time sufficient
to clear the service line. The recommended time for flushing varies from 2 to
5 minutes. Flushing time is relative; the intent is to flush the service line to
79
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
i
I
assure that the sample collected is representative of water in the main. Treat-
ment of water taps before collecting potable water samples is unnecessary if
water is allowed to run to waste and reasonable consideration given to the
choice of sampling taps (5). Thomas et al (13) reported that flammg taps
resulted in no significant differences for both total and fecal conforms nor for
the standard plate count.
The sample bottle must remain closed until it is to be filled. The stopper
must be protected from contamination when removed and during the actual
sample collection. The bottle should be filled without rinsing, leaving an
ample air space so that the sample can be thoroughly mixed before analysis,
and the stopper replaced immediately (1,5).
CompHance with recognized sample collection protocol can prevent ac-
cidental contamination and reduce the occurrence of spurious samples. Sample
collection is an important part of the testing procedure; the test is no better,
nor more reliable than the sample which arrives at the laboratory for
analysis. Laboratories must be certified, evaluated, approved—choose your
word—the intent is the same; however, there are few programs which provide
adequate and repeated training for sample collectors. Such training is
recognized as worthwhile but economically impractical; whereas failure to
have such a program may really be the economical impractabihty. Serious
consideration must be given to the implementation of a program designed to
train and periodically recertify sample collectors, similar to the milk pro-
ran. m1, ; , i '.,«
grin'application, transit time and storage are so interrelated, they should be
considered together. Standard Methods (1) provides recommendations for
transit/storage times for stream, source and potable water samples. Stream and
source water samples should be held below 10°C during a maximum transit
time of 6 hours. Such samples should be refrigerated when received m the
laboratory and processed within two hours. When local conditions dictate
delays in 'sample delivery longer than 6 hours, consideration should be
given to field examination or by use of the tentative delayed-incubationtotal
cbliform procedure. In a study on sample storage time-temperature effects,
Geldreich et al (6) observed that MPN coliform counts conducted on samples
held at 5°C for periods up to 48 hours were in acceptable agreement
with initial MPN results than counts observed on samples stored at 13-32°C
or ^t 35°C. Also, they observed better stability for samples held at 5°C and
examined by the MF technique. The more precise results obtained by the
till'', '<• I . Ili • ! i;i !t"'j| • direct MlF count as compared to the statistical MPN estimate revealed a more
subtle change in bacterial densities during storage. These recommendations
and information basically apply to stream and source water samples. Since
these requirements may not be realistic for individual potable water samples
transported to the laboratory by mail service,.the time interval between.col-
lection and analysis should in no case exceed 48 hours, preferably within 30
hours as specified in Standard Methods. Even the recommendation for refriger-
ation for potable water samples can be interpreted as optional and not
a requirement.
Presently, most State programs do not enforce the 30-hour transit/storage
time limit. Because of a multitude of reasons, e.g. inefficient mail service
and inability to utilize other means of transportation due to geographical
locations of remote water supplies, State programs allow examination of
samples which arrive within 48 hours, or more realistically, within two
calendar days. Few, if any State programs require refrigeration of potable
water samples during the transit period.
80
-------
SAMPLING COLLECTING AND HANDLING/ELD, NASH
Conscientious investigations were conducted by many workers in an at-
tempt to determine a transit/storage time of water samples which would not
adversely affect the bacterial flora either qualitatively or quantitatively. Cox
and Clairborne (4) reported that refrigerated water samples showed relatively
little change after 48 hours, while paired samples stored at room tempera-
ture showed a noticeable change. However, the lowest coliform density ob-
tained at zero hours storage was 390/100 ml which would not qualify their
samples as potable water. Shipe and Fields (12) indicated that a chelating
agent (EDTA) added to stream samples known to contain copper and zinc,
0.16 mg/1 and 6.5 mg/1, respectively, reduced the rate of decrease in E.
coJi density. They concluded that this chelating agent would.be of value
in maintaining the coliform index near the level existing at the time of
sample collection for a storage period of 24 hours. Data indicated that
addition of EDTA at the time of sampling resulted in approximately 69%
of the cells remaining viable after 2 hours and 40% after 24 hours. How-
ever, there is no indication of survival of coliform organisms in water when
the initial density was between 1 and 10/100 ml. Calwell and Parr (3) and
Hoather (8) reported a gradual decrease in coliforms with increased storage
time whether samples were iced or not. Lucking (10) concluded that the time
element between collection and examination is important and testing must
begin within 48 hours of collection. Lansome et al (9) concluded that sam-
ples having a coliform density between 1-20/100 ml need not be refrigerated
up to a 48-hour storage period with no adverse effects on coliform detection.
Previous investigations relating to transit/storage time of'samples have
omitted information regarding the influence of increased nonrcoliform den-
sities on coliform survival and detection, especially with respect to potable
water supplies. However, Lucking (10) and Lansome et al (9) certainly
recognized this as a factor directly relating to the over-all problems of transit
and storage. They also recognized that this and other factors such as changes
in temperature, availability of nutrients, oxygen tension and chemical concen-
tration are extremely varied in widely separated geographical areas, therefore
making the transit/storage problem unique for each water supply. Excessive
non-coliform densities may interfere with coliform detection when examining
potable waters expected to have low coliform densities (7). Analyses of
bacteriological data from a survey of distribution water transmitted by
969 public water plants did suggest that high non-coliform densities can in-
terfere with detection of low levels of coliform. Data in Table 1 shows that
the frequency of detecting total and fecal coliforms increases as the standard
plate count increased to levels up to 500/mI; but decreases in frequency
when the non-coliform densities exceed 1000/ml. These findings indicate that
high densities of non-coliform bacterial populations may adversely affect
coliform detection. The factor of high density non-coliform organisms versus
low density coliform organisms is just being investigated and may be cor-
related to transit/storage time for potable water as well as stream and
source water supplies. Data obtained on the distribution system water quality
in Northern Kentucky reflects the potential interference of coliform detection
exhibited by high densities of non-coliforms after a 24-30 hour transit/storage
period. The Northern Kentucky treatment plant laboratory data, based on
samples examined less than six hours after collection, indicated satisfactory
water quality throughout the distribution system. Data obtained on similar
samples collected during the same sampling period and from the same distribu-
tion system but examined after a 24-30 hours transit time, were reported by the
State Laboratory as being potentially unsatisfactory. Results of these samples
showed excessive or confluent bacterial growth. In some instances sheen was
81
-------
'; 1 i1!""1 II!1 !,». XI i,
MCROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
PLATE COUNT VS. COLIFORM DETECTION
WATER NETWORKS FOR 969 PUBLIC
WATER SUPPLIES.
-.! .
General Bacterial
Total Coliform
Fecal Coliform
Density Range
per 1 ml
-10
11-30
31-100
101-300
301 - 500
501 - 1,000
> 1,000
Number of
Samples
1013
371
396
272
120
110
164
Occurrences
47
28
72
48
30
21
31
Percent
4.6
7.5
18.2
17.6
25.0
19.1
18.9
Occurrences
22
12
28
20
11
9
5
•« /\T
Percent
2.2
3.2
7.1
7.4
9.2
8.2
3.0
TOTAL
2446
277
* Standard Plate Count (48 hrs. incubation, 35°C)
observed associated with the confluent growth, but distinct cohform colonies
did not develop which prevented quantitation of the coliform density Un-
published data obtained during a study of this Northern Kentucky distribu-
tion system indicate a statistical significant increase of the non-coMorm
density after a 24-hour storage period. The density increase appears to be
temperature dependent and is greater during the summer months when the
distribution sysLn water temperature reaches 26°-27°C. This increase in
non-coliform density occurs even though the free chlorine residual is main-
tained at a concentration usually above 1 nag/1 throughout the distribution
system. , ,
Limited data is available relating to transit/ storage time specifically for
potable waters/However, the data available does indicate that samples ex-
amined immediately after collection will provide the most valid results
Since this is impractical for potable water, data indicate that such samples
should be examined as soon as possible after collection. The transit/ storage
period ranges from 6 to as many as 72 hours. However, 24 hours for potable
waters appears to be acceptable by the majority of investigators while a
few indicate that 48 hours is also acceptable. All of the studies ^ indicate
that the transit/storage problem is dependent upon a multitude of factors
and the influence of these factors, individually or combined, will vary for
every water examined, be it stream, source or potable.
The bacteriological flora of a water sample will either increase or de-
crease upon storage. Which change and the degree of change are difficult to
predict and will vary with each sample, water supply and geographical loca-
tion. If these changes, which are related to transit/storage result ra un-
acceptable data, .there are alternate steps which should be considered. Samples
could be sent to laboratories nearer the water supply. Additional laboratories
or regional laboratories could be established. Alternate means of transporting
samples other than the mail service, such as bus lines, private courier or
United Parcel Service, should be seriously investigated. The delayed mem-
82
-------
SAMPLING COLLECTING AND HANDLING/H.D. NASH
brane filter total coliform test could also be considered, however, on a case
by case basis and not adopted as a cure-all procedure.
Perhaps the transit/storage dilemma for potable water cannot be resolved
until research is conducted dealing only with potable water samples. Such
a research project must address many types of water supplies, in many geo-
graphical areas to obtain results which can resolve the problem. Until
definitive data are compiled, it is necessary to make recommendations
based upon the data presently available, taking into consideration the con-
flicting interpretations of that data as applied to potable water samples. A
strong recommendation is made that potable water samples, be processed
within thirty hours after collection; .this is a maximum transit/storage time
period. If the laboratory is required by State regulations to examine samples
after 30 hours and up to a 48-hour transit/storage period, the laboratory
should indicate that the data may not be valid because of excessive delay be-
fore sample processing. Samples arriving for processing after a 48-hour period
should be refused without exception and a new sample requested.
REFERENCES
1. American Public Health Association. 1975. Standard method^ for the ex-
amination of water and wastewater. 14th ed. American Public Health Asso-
ciation Inc., Washington, D.C. '- -
2. Bordner, R.H., P.V. Scarpino, and J.A. Winter. (Eds.). 1977. (In prepara-
tion). Microbiological methods for monitoring the environment, part I—
water and waste. Environmental Monitoring and Support Laboratory, United
States Environmental Protection Agency, Cincinnati, Ohio.
3. Caldwell, EX., and L.W. Parr. 1933. Present status of handling water
samples—comparison of bacteriological analyses under varying tempera-
tures and holding conditions with si>ecial reference to the direct method.
Amer. J. Public Health 25:467-472.
4. Cox, K.E., and F.B. Claiborne. 1949. Effect of age and storage tempera-
ture on bacteriological water samples. JAWWA 47:948-952. ;
5. Geldreich, E.E. 1975. Handbook for evaluating water bacteriological labora-
tories. 2nd ed. United States Environmental Protection Agency; Washington,
D.C.
6. Geldreich, E.E., P.W, Kabler, H.L. Jeter, and H.F. Clark. 1953. A delayed in-
cubation membrane filter test for coliform bacteria in water, Amer. J.
Public Health 45:1462-1474.
7. 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. JAWWA 64:596-602.
8. Hoather, R.C. 1952. The bacteriological examination of water. J. Inst.
Water Eng. 61:426-442. :
9. Lonsame, B.K., N.M. Parhad, and N.U. Rao. 1967. Effect of storage tempera-
ture and time on the coliform in water samples. Water Res. /:309-316.
10. Lucking, H.E. 1967. Death rate of coliform bacteria in stored Montana wa-
ter samples. J. Environ. Health 29:576-580.
11. Public Health Laboratory Service Water Sub-Committee. 1953. The effect
of sodium thiosulfate on the coliform and bacterium coli counts of non-
chlorinated water samples. J. Hyg. 51:572-577.
12. Snipe, E.L. and A. Fields. 1956. Chelation as a method for maintaining
the coliform index in water samples. Public Health Reports. 71:974-978.
13. Thomas, S.B., CA. Scarlett, and W.A. Cuthbert. 1954. The flaming of tap
before sampling of the bacteriological examination of farm water supplies.
J. Appl. Bacteriol. 77:175-181. :
83
-------
ifcii'iiiffr,1'!11, 'SS!^ '""
ii ' ,"l<,i!!: 111:!
-------
EPA METHODS MANUAL/R.H. BORDNER
The EPA Methods Manual and the Microbiological Standards
Robert H. Bordner and John A. Winter
Environmental Monitoring and Support Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45211
Uniform application of the best available methods is essential for generat-
ing the valid data needed to enforce water quality standards. The EPA
methods manual, Microbiological Methods for Monitoring the Environment,
Part I Water and Wastes (2) provides these methods in support of the special
needs of the Agency.
The purpose of this report is to preview the microbiology manual in
preparation, show how it supports the Agency's responsibility to monitor the
quality of our Nation's waters, and demonstrate particularly how it performs
this function for the National Interim Primary Drinking Water Regulations
(NIPDWR).
Responsibility for Standardization
The Federal Water Pollution Control Act Amendments of 1972, the Marine
Protection, Research and Sanctuaries Act of 1972 (commonly referred to as
the Ocean Disposal Act) and the Safe Drinking Water Act of 1974 direct
the Environmental Protection Agency to protect and improve the quality of
our water resources and to control pollution. To achieve these goals the Agen-
cy must set, enforce and monitor standards for water supplies, ambient wa-
ters and wastewater discharges. In EPA, the Environmental Monitoring and
Support Laboratory (EMSL)—Cincinnati is responsible for selecting, vali-
dating and standardizing analytical methods for water and for establishing
quality control procedures.
With the publication of the microbiological methods manual, EMSL—Cin-
cinnati has completed its mission to provide standardized methods for chem-
istry, biology and microbiology. In addition, EMSL has prepared a quality
control handbook for water and wastewater laboratories.
These manuals are not replacements for nor in competition with Standard
Methods for the Examination of Water and Wastewater, (1) but rather
provide the basic methods for EPA to carry out its enforcement responsi-
bilities. The EPA microbiology manual cites as valuable sources of further
detail and background information Standard Methods and EPA's Handbook
-for Evaluating Water Bacteriological Laboratories (3).
Development of the Manual
The standardization process for microbiological procedures began at a
seminar held in San Francisco January, 1973 where the Agency's needs for
standard microbiological methods were agreed on and a steering committee
was formed to develop the EPA manual. It represented the Office of Research
and Development, the Regional Offices, the Office of Water Programs and
the National Environmental Investigations Center (Office of Enforcement).
Committee members were assigned reisponsibility for preparing sections on
selected microbiological parameters. First drafts were presented and reviewed
at the second meeting of the committee in January, 1974 at Cincinnati where
the basic design, format and content were also established for the manual.
: 85
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
The editors, Robert Bordner and John Winter, EMSI^Cincinnati, developed,
formatted and edited these materials and added technical detail and infor-
mation reflecting Agency policy. This draft was reviewed by EPA programs
and Regional offices. The many constructive review comments, changes and
corrections are now being incorporated into the final version of the manual
which is scheduled for distribution in September, 1977.
Legal Authority
The manual will be the basic* reference lor monitoring water and wastes
in compliance with applicable water arid effluent standards established by the
Agency. Publication in the Federal Register will establish the legal authority
for the methods used in the NPDWR program as was done for the pro-
cedures used in the National Pollution Discharge Elimination System
(NPDES) for the analysis of effluents.
The Code of Federal Regulations (CFR) Title 40, Part 136 (6) lists the
procedures required for Section 304 (g) of the Federal Water Pollution Con-
trol Act (NPDES). Table 1, an excerpt from that Federal Register, shows
the approved test procedures for bacterial parameters. References are made to
Standard Methods for the Examination of Water and Wastewater (1) and
the methods of the U.S. Geological Survey (5). When released, the EPA
microbiology manual will be referenced in the column provided in the
table. The amendments also provide blanket approval of alternate test
procedures by applications to EMSI^-Cincinnati and for individual approval
by application through the EPA Regions.
It is anticipated that the alternate test procedures for the Naional
Primary Drinking Water Regulaions will be referenced similarly by amend-
ment to the Federal Register.
,
I",
;,:,,, ;
. '• '',','jff''
'' "'
l.;:i
Objectives of the Manual
To carry out the legal requirements set forth in the three water quality
acts the EPA manual must provide:
1. "fhe best available 'standardized methods for water quality determina-
tions, enforcement actions, compliance monitoring, resource planning and
research activities. ....... The procedures will be the reference methods for the
' evaluation of modified ..... or'new meffiods.
^'^e'accompIls!imen^of the objectives of ..... trie manual will assure the uni-
form application of the most reliable procedures in all laboratories. Uniform
procedures facilitate comparison, interchange and common use of data
in cooperative field surveys, interlaboratory studies and information storage
and retrieval systems.
2, Methods that have the general consensus of microbiologists within EPA.
]^[any methods are in use by researchers or by water analysts that have not
been proven by application in the various waters and geographical areas of
this country. These, methods are not necessarily acceptable for uise in the
'
3. Methods that are tailor-made to meet the specific needs of the Agency.
Procedures acceptable for some agencies may not meet the needs of EPA.
For example, if a decision is made for more stringent holding time require-
ments for potable water samples than has previously been practiced, such
specification will be put in the manual.
I I I
86
-------
TABLE 1. LIST OF APPROVED TEST PROCEDURES FOR WASTEWATER *
Parameter and units Method EPA
4.
5.
6.
7.
8.
BACTERIA
Coliform (fecal)5, number MPN; membrane filter
per 100 ml.
Coliform (fecal)5, in pres- .. _ do 6- «
ence of chlorine, num- ;
berperlOOml. ;
Coliform (total)5, number do6.
per 100 ml.
Coliform (total)5, in pres- MPN; "« membrane filter
ence of chlorine, num- with enrichment
ber per 100 ml. - • I • •
Fecal streptococci5, num- MPN; « membrane filter;
ber per 100 ml. plate count.
r
References
(page nos.)
USGS
methods 14th ed. PT. 31 methods7
standard 1975
methods ASTM
922
937 45
922
928,937
916
928 35
916
933
943
. 944 5Q._ .
947
I
00
5 The method used must be specified.
6 The 5 tube MPN is used.
7 Slack, K.V. and others, "Methods for Collection and Analysis of Aquatic
Biological and Microbiological Samples: IIS. Geological Survey Techniques of
Water-Resources Inv. book 5, ch. A4 (1973)". : icwiumues m
8 Since the membrane filter technique usually yields low and variable recovery
from chlorinated wastewaters, the MPN method will be required to resolve any
controversies.
1 FEDERAL REGISTER, VOL. 40, NO. 232-40 CFR 136, DECEMBER 1, 1976
ffi
w
-------
MICROB*OL STANDARDS EVALUAridN/c.w.
i f:,;,
.111 ...... j ..... v.
4 Procedures with limited options: For enforcement purposes the analyst is
required to use one preferred' procedure whenever possible. At the First
Microbiology Seminar on Standardization of Methods in San Francisco (4)
in 1973 Jeter reported four variations of membrane filter tests and sixteen
ways to conduct multiple tube tests for total coliforms in Standard Methods.
The EPA manual must provide a single, direct method wherever possible.,
For example the use of lactose broth for the MPN analysis of total coli-
forms will not be included in (the EPA manual. , ~ •* • fnr
5 Validation of methods previous to publication in the manual. Criteria tor
validation are specified for nevf or additional methods to be included in the
manual. No methods will be added to the manual without sufficient support-
6D|'Quick response to changing needs of the Agency is possible by the single
agency action of EPA. For example the Water Supply Laboratory Certifica-
tion Program requires criteria; and procedures to evaluate all water supply
laboratories. This very timely and nationally-important information was in-
serted inthe microbiology mamial without delay. If the Maximum Contami-
nant Levels or methods of establishing levels for total coliforms in drinking
water should change, the manual could easily reflect these changes.
Special Features of the Manual
•"'.•: ' , , '.-:' • -'. ,::: •• ": ;.••-, ,-j „!
The EPA manual features several special characteristics:
1 A looseleaf style is planned'for the manual so that method improvements
or additions can be made easily. For example, a recent development in
the manufacture of "membrane filters produced a filter of larger pore size for
microbiological analyses. A statement was added to the manual allowing the
use of this filter, which is not permitted by Standard Methods. If the National
Academy of Science recommends additional parameters for drinking water,
for example the standard plate count, the methods would be added to the
manual. I f '
2. New editions are planned as demanded to keep the contents ot the
manual current These revisions can be published in as little as 18 months.
As improved test procedures are developed and validated and other parame-
ters are proved useful they will, be considered for the manual.
3. The manual guides the 'analyst in selection and use of the best method for
his needsi In response to the recent controversy concerning use of the mem-
brane filter procedures for chlorinated effluents, the EPA manual limits the
methodology. ...
4. Important supporting information for good laboratory practice is in-
cluded in the manual. Examples are a section on quality control of micro-
biological procedures, an area jthat has been largely ignored. Another section
is devoted to laboratory and field safety. A third applies basic statistics to
microbiological analyses, and a fourth discusses legal considerations for the
micfobiologist. .
5. The manual will be widely distributed to Federal, State, local and private
laboratories performing water"analyses'.
6. it is written in a simple, direct, stepwise style that is easily understandable
and prepared hi a "cook book]' form so that it can be used at the bench by
the analyst. The looseleaf form will also be conducive to use at the bench.
7. The format assigns a decimal number to each paragraph for easy
reference and location, similar to the American Society for Testing and
Materials Books of Standards.
88
ll|M
I!;,;;;!! ,,„!, ,JilK .; iiiilllil Illiillll
-------
EPA METHODS MANUAL/R.H. BORDNER
Summary
In summary, the EPA microbiological methods manual fills the specific
needs and legal requirements of the Agency! to standardize methods in sup-
port of its enforcement, compliance monitoring, ambient monitoring and re-
search programs. The methods chosen are the best available, allow limited
options and represent the consensus of EPA microbiologists. The manual
format allows for rapid response to changing Agency needs,; offers oppor-
tunity for frequent revisions, guides the analy'st in the selection and use of the
best methods for his/her purposes and features a stepwise "cook book" form
for application at the bench level. j
REFERENCES
1. American Public Health Association. 1975. Sjtandard methods for the examina-
tion of water and wastewater. 14th ed. The American Public Health Associa-
tion.
2. Bordner, R.H., P.V. Scarpino and J,A. Winter (Eds.). 1977 (In prepara-
tion). Microbiological methods for monitoring the environment, part I—
water and wastes. Environmental Monitoring and Support Laboratory, U.S.
Environmental Protection Agency, Cincinnati* Ohio.
3. Geldreich, E.E. 1975. Handbook for evaluating water bacteriological labora-
tories, EPA No. 670/9-75-006. 2nd ed. Municipal Environmental Research
Laboratory, Office of Research and Development, United States Environ-
mental Protection Agency, Cincinnati, Ohio..
4. Jeter, H.L. 1973. Total coliforms, p. 20-24.f Proceedings of the First Micro-
biology Seminar on Standardization of Methods, EPA No. R4-73-022. Office
of Research and Development, United States Environmental Protection
Agency, Washington, D.C. ;
5. Slack, K.V. et al. 1973. Methods for collection and analysis of aquatic
biological and microbiological samples, Book 5, chapter A-4. In: Tech-
niques of water-resources investigations of the United States Geological
Survey. United States Printing Office, Washington, D.C. i
6. United States Environmental Protection Agency. 1976. Guidelines estab-
lishing test procedures for the analysis of /pollutants—amendments. Federal
Register 40:52780-52786. {
QUESTION AND ANSWER SESSION'
i
C.W; Hendricks, Office of Drinking Water, U.S. Environmental Protection
Agency, Washington, D.C. :
When will the manual be available? '
R.H. Bordner, Environmental Support Laboratory, U.S. Environmental Pro-
tection Agency, Cincinnati, Ohio j,
I " ' "
The target date for printing is September, 1977. The manual should be avail-
able by the end of the calendar year.
i'
S.M. Morrison, Department of Microbiology, Colorado State University,
Fort Collins, Colorado 1 :
In response to the Drinking Water Law many[ laboratories are buying equip-
ment and instrumentation and preparing to jperform tests they have never
done before with technicians who are completely unfamiliar with the tests
-------
1 ilff, f'f ]>T
ill! ........... i ......... ....... • ...... am ...... y;:im .......... ~
•w ....... » ....... ; ................... w ...... wm-im ..... HIT ...... rw,1*:™ ..... r: .......... • ...... ..... rsis ..... «! ....... wmx ...... IWSHKI? ........... iis?. ........... '
i! !!;
'! i I HBW • Hill
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKSi
and equipment. This is particularly true of the smaller laboratories. Have
any provisions been made for this problem?
" ..... ' .. . . . 1 ................. il ...........
R.H. Bordner
EPA recognizes the need for training as an important part of the program to
upgrade the laboratories. Training by the States is certainly encouraged; to
pur knowledge funding has not been provided by EPA.
"' ........... ' .......... ............. • !
S.M. Morrison
I would like to see the use of regional of district laboratories which do have
this expertise encouraged.
W. Litsky, Department of Environmental Sciences, University of Massachu-
setts, Amherst, Massachusetts
You described briefly the relationship of the EPA methods to Standard
Methods. How do the methods in the EPA manual relate to those being
developed by other organizations such as ASTM (the American Society for
Testing and Materials)?
...... • ! .................................. ]•
R.H. Bordner
EPA members are participating in the activities of ASTM as well as
Standard Methods and .other methods — standardizing groups. The methods
must be reviewed by the EPA Microbiological Methods Steering Commit-
tee. Criteria for inclusion of new methods in the manual, in the future
state that the method should satisfy the needs of the Agency, show some
advantages over present methods, be validated by the developer and other
laboratories, be relatively easy to use and practical, etc.
, , M ..... ,, * ...... ',;,;,,," , ',,!,, J| , ..... , , „
Q,W. Hendricks! j
You mentioned internal review j of the manual.' Are "there provisions for the
manual to be reviewed by experts in aquatic microbiology outside the
Agency? !
*, ""- , , .. ................. ,• ........ ; ....... ,
R.H. ' ' .................... ~' ........................ " ............... ' ................ ......... ....................... ' ................ '' ...... ' ....... ................. ...... ........... ..................................... ..........................
We dp plan pn an external review.
90
-------
STANDARD METHODS/A.E. GREENBERG
Evaluation of Standard Procedures: Standard Methods
Arnold £. Greenfoerg
Bioenvironmental Laboratories Section
California Department of Health
Berkeley, Caltfornia 94704
In 1905, after almost ten years of deliberations and discussions, the
American Public Health Association published the first edition of what is
now called Standard Methods for the Examination of Water and Wastewater.
It was a book devoted largely to the enumeration and identification of bac-
teria in water but it did include methods for chemical analysis as well. The
intervening 72 years have seen tremenous expansion of the manual, wide-
spread legal or quasi-legal acceptance of the methods it contains, and a
total of fourteen editions. The 15th edition is in preparation an;d expected to
be available by 1978.
Although APHA has continued as the publisher, it was joined by the
American Water Works Association in 1925 and in 1947 by the Water Pol-
lution Control Federation in preparing this widely used and widely accepted
manual of methods. The three sponsoring societies have different internal
procedures for dealing with the administrative and technical details of pro-
ducing Standard Methods but all of these are coordinated by a Joint Edi-
torial Board that consists of the chairman of each society's standard methods
committee. Within this framework, APHA has major responsibility for the
sections on microbiology.
The current edition includes 131 pages (out of a total of 1193) devoted to
the bacteriological examination of water and wastewater. Principal emphasis is
given to counting organisms indicative of fecal pollution such as total coli-
form bacteria, fecal coliforms, and fecal streptococci by either a multiple
tube dilution or a membrane filter technique and to enumeration of the
total bacterial count. Brief sections are devoted to detection of pathogens
such as Salmonella, Shigella, enteropathogenic E. coli, and leptospires. Tenta-
tive procedures are included for viruses in water, Staphylococcus and Pseu-
domonas aeruginosa in swimming poods, fungi, yeasts, actinomycetes, and
nematodes. A section on the identification of nuisance iron and sulfur bac*
teria is included also.
The sections devoted to indicator organisms do not emphasize a strict
taxonomic approach to identification since that has never been considered of
primary importance in public health. Tfcius, for example, methods for Escheri-
chia coli or Enterobacter aerogenes are much less important than procedures
for fecal coliforms which are defined as coliform bacteria capable of pro-
ducing gas from lactose at 44.5° C. This difference stresses the pragmatic ap-
proach to counting organisms originating in the gut of warmblooded animals
rather than specific identifications. :
On both a national and an international level in recent years there has
been a proliferation of organized interest in standardization of methods hi
aquatic microbiology. In the United States, as we have heard, the Environ-
mental Protection Agency is in process of preparing a manual of bacteriologi-
cal methods that will be a companion volume to their already published
Methods for Chemical Analysis of Water and Wastes (EPA-625-/6-74-003,
1974) and Biological Field and Laboratory Methods for Measuring the
Quality of Surface Waters and Effluents (EPA-670/4-73-001, 1973). It has
been argued that EPA as a regulatory agency requires procedures that will
91
-------
IE il1!1:*1 :' .1 • I
!l!i lilritt i II '
If" 111* •'.' • I "• ("I1
I ill 1
-------
STANDARD METHODS/A.E. GREENBERG
Joint Task Group responsible for preparing the section on quality control,
the remedy is obvious and immediately at hand.
In conclusion, the Water Supply Program of EPA has available to it
the best efforts of bacteriologists, including those from EPA, in providing
analytical methods for evaluating microbiology standards. These exist in
Standard Methods for the Examination of Water and Waste-water which
should be adopted as the sole methods source. Exclusion of unacceptable
alternative .methods can be achieved by appropriate page or section citations
in EPA regulations.
93
-------
II! 1't.ir
1 g1 lull," I I,
Bi&J I' f ii
111 11)1!':' I i
a, 'illlilllll I1!1" i !" »'i, !!," !" '' if >i
.mi1 mil'',.'. '".i t:'i at:1'
"!• tih • 5 : i', •';i; • HI
it IIP "iinii ;i,i B"1:
:'l /'lull' /'Illli!
." s i!|i:.ii!
, ill, III" •',, ijlll
''Kir ",!<: is. ?
1*",);1-, iihii'ii 'ii•"'<', •
'i .1 ;-,iiilr W ' i'JH I
i Hill", i, ill. .'• "-!"M
• Hffi1^ :t ;>:'! I
-------
VIRUS DETECTION/D.O. COVER
Virus Detection
Dean O. Cliver
Food Research Institute
University of Wisconsin
Madison, Wisconsin 53706
Field Virus Survey Unit (US Environmental Agency)
I have been assigned the topic of "Virus Detection", as it relates to Evalua-
tion of the Microbiology Standards for Drinking Water. I plan to interpret
this charge rather liberally. ,
I shall begin by noting that the purpose of standards is to .protect public
health, that viruses are a class of pathogens sometimes transmitted by water,
and that I have been unable to find the; word "virus" in the National Interim
Primary Drinking Water Regulations. In the spirit of the Regulations, I
had better define the term before I use it further. The viruses share three most
important properties:
1. They are transmitted from host to host in the form of discrete particles
too small to be seen- with a light microscope.
2. They multiply only inside suitable living host cells.
3. They cause mysterious, incurable diseases (in the sense that when
your physician tells you, you have a "virus", he is saying he doesn't
think he can do anything for you). Fortunately, most individuals, if
they are otherwise healthy, recover from most virus infections without
any substantive medical assistance.
The only virus disease proved to be 'transmitted currently in water in the
United States is hepatitis A, formerly known as infectious hepatitis (11).
95
-------
•i Jsif, ::' I
i '1,,, ! if 1,,
• Hi
ll Bill1 , li"i 111 I,,, I
I'lS ri.si •
«i life
I;,11 Ulln , , i I'"1 , II. M.I i',I. .,<
li:l'?l:tl&
is ill'''", i .. i!' '•*•:
•i:l! li I,1 ill
i
ill
IIP
'Hill i I
Illllll
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
Still some of the gastroenteritis which results from drinking contaminated
water is probably caused by virus (2) and many viruses known to cause other
diseases are detected in sewage-polluted waters (10). I don't want to be-
labor the point, but I think it is worth noting that all of the viruses thought
to be transmissible in drinking water are shed in human feces, and that these
virus-containing feces have been diluted by a factor of approximately 2000
as they occur in domestic wastewater. Therefore, virtually all threat oi
virus transmission through drinking water could be eliminated by perfecting
and requiring the use of some ^of the alternatives to common water-flush
toilet (also known as the "W.C.").
We can assume that this is unlikely to happen in the near future and that
viruses will continue to occur in raw water sources supplying some com-
munity systems, I'ougnt, then, to point out that conventional water treat-
ment techniques were not specifically developed to remove or inactivate
viruses, and that viruses have a perverse tendency to adhere to each other (5)
y wM as to suspended solids in water (12). These factors, and the relative
resistance of viruses to chemical disinfectants (9), would appear to compli-
cate the task of producing virus-free drinking water from contaminated raw
water or from wastewater. Still, known incidents of virus disease associated
.. "water have resulted either from chronic under-treatment of
was ".known to be contaminated, or. from cross-contamination
g^^'s'^f^^^ .............. virus-containing wastewater into other-
••'jjjyjgg safe finished/ water during distribution (4). Rather than viewing vkus
" if ansmission ...... through ...... water as an exotic problem requiring new and unprece-
lente! countermeasuresT I suggest that it be seen as another potential con-
|feguence ........ of making some of the same old mistakes. Such lapses are infre-
quent and are not equally probable: under-treatment has most often been
seen to occur in private or semi-public water systems; incidents involving
community water supplies have most frequently entailed recontamination
of the finished water during distribution (4).
One would like to have some btHer meahs than human disease to detect
the presence of virus in drinking water. The task of recovering virus is
complicated by the extremely small quantity of virus material that is likely
to be present in a water sample and by the tendency of the virus particles
to adsorb to suspended solids (13). However, quite elegant methods have
been devised for ' recovering virus particles from samples as large as 1900
liters (~ """"500 gallons) or greater. The virus particles in the concentrated
||^pe g-g«| g— st-fl to be detected by the infections they may produce in
tissue cultiires (3). Such infections become perceptible only slowly: inocu-
lated tissue cultures must usually be Incubated and observed for days or
weeks. The potential for success varies: no tissue culture is known to sup-
port replication of the hepatitis A virus, but tissue cultures are probably
incp sensitive than the host's body to some of the other intestinal viruses
•''wnicn' ' imglit" occur "in water (Oliver, unpublished data). Although, there
will always be room for improvement in such techniques, we have now
reacned the point where a "Tentative Standard Method" for detecting
viruses in water exists (3). The method has been shown to recover experi-
mentally-inoculated viruses efficiently from large-volume samples of water
(6) . No virus has been detected in approximately 175 large-volume samples
of finished community drinking water by this method, even though the raw
water sources of some of these communities were known to be highly con-
taminated (1; E.W. Akin, EPA Cincinnati, personal communication).
The most obvious explanation for these negative test results is that vkus
was absent from the samples, although other reasons can be envisioned.
i 1 I II I
96
-------
VIRUS DETECTION/D.O. CUVER
Because it is difficult to detect viruses in water, it is reasonable to seek
"indicators" (organisms or substances whose presence in water is likely to
accompany that of virus). In fact, there are many possible indicator systems:
Coliforms, fecal coliforms, and fecal streptococci suggest fecal contamina-
tion of water; however, they differ from viruses in the attrition they are
likely to undergo in the environment or during water treatment. Turbidity,
standard plate count, and free residual chlorine (or lack of it) have, at
best, an indirect association with the likelihood that virus has contaminated
water. Coliphages (bacteriophages specifically infectious for Escherichia coli)
as a group are indicative of fecal contamination of water and may lose their
infectivity at a rate comparable to animal viruses (8); however, there seems
to be some problem in selecting a "best" host strain of E. coll in which to
detect the coliphages (14).
Finally,, the vaccine strains of the human polioviruses have been proposed
as indicators because they are almost always present in urban sewage, are
likely to lose infectivity at approximately the same rates as many other in-
testinal viruses, and may be more rapidly detectable than viruses in general,
using a fluorescent antibody test procedure (7). Unfortunately, polioviruses
are not always present in sewage, and the detection methods are a good deal
more complex and exacting than most Standard Methods microbiological
procedures.
By now, you are probably wondering how it is ever possible to determine
that a community's drinking water does not contain virus. In fact, it is not
possible to demonstrate the complete absence of virus from each day's total
production of finished water. The most direct test of the safety of drinking
water is drinking the water. However, given the twd-liter-per-day consumption
figure upon which the Maximum Contaminant Levels are supposed to be
based, only about 1 % of the total water used at home (and, therefore, even
less of the daily per capita total water consumption) is. drunk.. The rest is
used in a variety of ways that are unlikely to allow manifestation of any
virus it contains.
The water could also be tested for virus by the procedures mentioned
previously. I believe that the surveys now in progress represent the most
proper use of the Tentative Standard Method and its counterparts—these are
not procedures appropriate for routine quality control of finished waters
by individual suppliers; neither can they reasonably serve as the basis for
establishing or enforcing Maximum Contaminant Levels for viruses. What
the survey results seem to show is that virus-free water is attainable by means
of the best current treatment methods, if these are conscientiously applied,
even when raw water sources are far from ideal. Less rigorous treatment is
likely to be necessary with high-quality, protected raw water sources. The
evidence indicates that the means to produce .virus-free finished water are
at hand, at least for raw waters that do not consist principally of sewage
effluent. The most important task at this time is to ensure that these means
are employed, absoluely without fail, by water suppliers. Whether virus
testing involves one "large" sample or several small ones per day, only a
minute portion of a community's daily water output could be monitored in
this way, at a cost in resources that could better be applied to other surveil-
lance tasks. The findings of the EPA community water survey indicate that
the results of such testing are likely to be negative, but this obviously does
not guarantee that the. consumer is receiving virus-free drinking water.
I believe that the most consumer protection that can be obtained for the
consumer's dollar will derive from means already at hand. Suppliers should
be helped in selecting treatments appropriate to the quality of their raw wa-
97
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
ter For any source other than extremely well-protected ground water, treat-
ment should probably include at least the following: 1) coagulation and
sedimentation, 2) filtration, and 3) chemical disinfection (I am not now
prepared to recommend treatments for direct reclamation of urban effluents).
The physical: treatments which are suggested should serve to remove proto-
zoan cysts, a very real cause for concern, as well as most solids that might
protect virus from chemical disinfection. Beyond selection and adaptation of
treatments appropriate to the quality of the raw water, adequate consumer
protection requires that the supplier by supervised to ensure that the treat-
ments are properly performed all of the time. Microbiological testing of the
finished product alone won't do the job. There must be monitoring of both
microorganisms and turbidity, as welt as a continuous record of the proces-
sing as it Is done. Then the quality of the finished water must be sustained all
the way to the user's home. The presence of a free chlorine residual, or the
absence of coliform organisms, at the tap is reasonable evidence that cross-
contamination or back-siphonage has not occurred to degrade the finisnecl
water during distribution. ' _ _. _ ' .'"
I believe that this discussion supports a few conclusions: First, it is
reasonable to test'for".viruses"in finished drinking water on occasion, but it
is not 'reasonable to expect drinking water suppliers to-do it. Second, Maxi-
;m;gm*'contaminant levels (as defined in the Safe Drinking Water Act of
1974), are not appropriate for viruses. Finally, the basis for protecting the
public from water-borne virus disease lies rather in detecting lapses in the
practices of treatment and distribution than in detecting the virus in the
^finished product.
;.. .-. • • ; REFERENCES ' ' ,
1. Akin, E.W., D.A. Brashear, and N.A. Clarke. 1975. A virus-in-water study
Of finished water from six communities. Health Effects Research Laboratory,
Office of Research and Development, United States Environmental Protection
Agency, Cincinnati, Ohio.
2. Albrey, M.B., and A.M. Murphy. 1976. Rotaviruses and acute gastroen-
teritis of infants and children. Med. J. Aust. 7:82-85.
3 American Public Health Association. 1976. Section 913: Detection of
enteric viruses in water and wastewater, pp. 968-975. In: Standard methods
for the examination of water and wastewater. 14th ed. American Public Health
Association Inc., Washington, D.C.
4 Craun, G.F., and L.J. McCabe. 1973. Review of the causes of waterborne-
' disease outbreaks. JAWWA. 65:74-84. .
5. Floyd™ R., and D.G. Sharp. 1977. Aggregation of poliovirus and reovirus by
dilution in water. Appl. Environ. Microbiol. 55:159-167.
6. Hill, WK W. lakubowski, E.W. Akin, and N.A. Clarke. 1976. Detection of
virus in water: sensitivity of the tentative standard method for drinking
water. J. Appl. Environ. Microbiol. 31:254-261.
7. Katzenelsoii, E. 1976. A rapid method for quantitative assay .of poliovirus
frpm water with the aid of the fluorescent antibody technique. Arch. Virol.
50:197-206. • t . •"
•g. Kott, Y., R. Netta, S. Sperber, and N. Betzer. 1974. Bactenophages as viral
pollution indicators. Water Res. 8:165-171.
9. Liu, O.C., H.R. Seraichekas, E.W. Akin, D.A. Brashear, E.L. Katz, and WJ.
Hill, Jr. 1971. Relative resistance .of twenty human enteric viruses to free
chlorine in Potomac water, pp. 171-195. In: Snoeyink, V., and V. Griffin
(Eds.), Virus and water quality: occurrence and control. Proceedings of the
Thirteenth Water Quality Conference. University of Illinois Bulletin, Vol.
69, No. 1, Urbana, Illinois.
98
-------
VIRUS DETECTION/D.O. COVER
10. Metcalf, T.G., C. Wallis, and J.L. Melnick. 1974. Virus enumeration and
public health assessments in polluted surface water contributing to trans-
mission of virus in nature, pp. 57-83,, In: Virus survival in water and waste-
water systems. Center for Research in Water Resources, The University of
Texas, Austin, Texas.
11. Mosley, J.W. 1967. Transmission of Viral Diseases by Drinking Water, pp.
5-23. In: Transmission of viruses by the water route. Interscience Publishers,
New York.
12. Schaub, S.A., and C.A. Sorber. 1976. Viruses on solids ,in water, pp.
128-138. In: Viruses in water. American Public Health Association Inc.,
Washington, D.C.
13. Sobsey, M.D. 1976. Methods for detecting enteric viruses in water and waste-
water, pp. 89-127. In: G. Berg et al. (Eds.), Viruses in water. American
Public Health Association Inc., Washington, D.C.
14. Vaughn, J.M., and T.G. Metcalf. 1975. Coliphages as indicators of enteric
viruses in shellfish and shellfish raising estuarine waters. Water Res. 9:613-616.
99
-------
i. ','!, I1!1!1!*' "mi .ill< i' !t ii. i.l* ', Hi''1" I'i "I !Fi
'"'II ,' " 111 III1 i'1 II" II1 Wl;lll
'' : I'll
-------
DISINFECTANT RESISTANT ORGANISMS/P.T.B. SHAFFER
Disinfectant Resistant Organisms *
Peter T. B. Shaffer
Carbo.rundum Company
Niagara Falls, New York 14302
Much of the dialogue at this symposium has been based on laboratory
disinfection studies in which the die off rates of standard strains of viruses
have been measured as a function of time at various levels of chlorine
residuals. While this specific information is technically sound, it overlooks
the variations in resistance to chlorine inactivation among various strains
of viruses or among individual members of a specific virus strain. Thus, it
becomes questionable to measure a chlorine residual and assume that this
will assure a given level of disinfection of all viruses. Any indirect standard
must be viewed with some misgivings. If a water is to be assured to be free of
a particular organism, that organism must be monitored even if infrequently.
Virus isolates recovered from well-treated finished water (2) have been
questioned, to no small extent on the basis that they were recovered from a
water of very low turbidity in the presence of chlorine residuals in excess
of one milligram per liter (1). We have examined isolates from one such
study in considerable detail and are convinced that they are sufficiently dif-
ferent from any available laboratory strains to assure that they are natural
isolates and not artifacts resulting from contamination (3).
Most recently, two of these isolates were grown to high titers, carefully
purified by sedimentation in an ultracentrifuge, and subjected to chlorine in-
activation. Their inactivation rates were compared with those of two
standard poliovirus 1 strains (LSc and Mahoney) grown up and purified in
the same way. They showed a chlorine inactivation rate orders of magni-
tude lower than that of either of the two standard controls. These data
shown, in part below, will be published in detail shortly.
VIRUS LOG SURVIVAL
STRAIN FRACTION (30 MINUTES)
Mahoney
LSc
Isolate 1
Isolate 2
-4.6
-7.79
-1.74
- -3.83
FREE CHLORINE RESIDUAL
(mgll)
0.39
1.34
0.92
1.31
The point to be noted is that all viruses, in fact all strains, or even all
members within a single standard strain, do not respond in ,the same way
when subjected to disinfection conditions typical of those applied hi a well
operated drinking water plant. To assume that all viruses will exhibit the
same die off rate as that shown by standard laboratory strains is not tech-
nically sound. Therefore, to replace a direct standard, whether it be coliform
or virus, with an indirect standard of turbidity and chlorine residual can not
be defended on the basis of experimental evidence.
* This paper was submitted after the Conference.
101
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
REFERENCES
in. ..... i';
'' '
nil ilH i, II '!•!:,;•
liiiii."iii/iiii1' i ' I '•,'" I".''
lili'MflllNIIUI!1 i i " ill, n/f
fi JH"i> : i , 1
1!'
Hf'!!'
1 '"31
1. Akin, E.W. and W. Jakubowski. 1976. Viruses in finished water—the Occo-
' quan experience. Proceedings AWWA Water Quality technology Conference,
San Diego, California.
Hoehn", R;C., C.W. Randall, F.A. Bell and P.T.B. Shaffer. Trihalornethanes
and viruses in a water supply. J. Environ. Eng. (to be published).
Shaffer^ P;Y.B., .RJE. Meierer and C.D. McGee. 1976. Isolation of natural
viruses from a variety of waters. Proceedings AWWA Water Quality Tech-
nology Conference, San Diego, California.
4, %|ffer; 'P;T.B., "T.G. Metcalf, O.J. Sproul and D.G. Sharp, Chlorine re-.
sistant poliovirus from finished drinking water (to be published).
",• !!
' li'lll N "Ml,,,', ,|
) lifl
lit'ill „
i'(>" , «;<;
i „!;;,: ,. ' j.. .,;.,
!,"': '' *l ' ' "i1
l|l "U;1!'!!' 'Vlr " ill
102
-------
TURBIDITY AND DISINFECTION/J.C. HOFF
The Relationship of Turbidity to Disinfection of Potable Wafer
John C. HoS
Water Supply Research Division/MERL
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
Introduction
Although turbidity has been used as a measure of potable water quality
for many years, increased attention has been focused on this parameter as a
result of its inclusion. at a Maximum Contaminant Level (MCL) of 1
Nephelometric Turbidity Unit (NTU) in the National Interim Primary
Drinking Water Regulations. Turbidity in drinking water can be caused by a
large variety of particles which can originate from many sources. Turbidity
may be present as a result of incomplete removal by treatment! processes of
particles present in the raw water source or may result from problems with
water treatment techniques such as coagulation. Turbidity may also originate
within the distribution system as a result of corrosion, growth of micro-
organisms, line repairs, etc. ;
Because of the current emphasis on turbidity in potable water, the signif-
icance of turbidity, particularly as it relates to disinfection efficiency during
water treatment, will be discussed from the historical point of view as well
as in the light of recently developed information. Problems associated with the
use of turbidity as a contaminant limit and some possible alternatives for
future consideration also will be discussed.
Turbidity as a Potable Water Standard
Turbidity in water is caused by the presence of suspended matter such as
clay, silt, finely divided organic and inorganic matter and; microscopic
microorganisms. Turbidity is defined and expressed as an optical property
of water and is measured by determining the degree of light scattering by
the particulates present in the sample (1). Because of the nature of the
measurement, turbidity determinations cannot be related either to the
quantity (mg/1) or characeristics (size, shape, chemical nature) of the
particles present in the sample. Another problem is that different types of
turbidimeters, after being standardized on the same turbidity standard, will
usually give readings differing by as much as 500% when used to measure
the turbidity of a sample containing another turbidity causing substance
(12).
Despite these limitations, turbidity has been found to be a useful parameter
for measuring the quality of public water supplies. Turbidity first became a
part of the USPHS Drinking Water Standards in 1942 when a tolerance
limit of 10 units was established. In 1946 this was changed to a tolerance
limit of 10 for filtered water only and a recommended limit of 10 for other
waters. In 1962, the recommended limit was reduced to 5 and the tolerance
limit was dropped (2). The most recent change embodied hvthe National
Interim Primary Drinking Water Regulations establishes a turbidity Maximum
Contaminant Level (MCL) of 1 with up to 5 units allowed if the supplier
can demonstrate that the higher limit does not do any of the following:
1. Interfere with disinfection.
103
-------
rvnri (Fffn '• s rvt
v-wsHcmf/n IK1 nil iisii
1' Mi :'••['
|H 'lilll
ll llll''1";: :,:
I"! I if I/:
I'1::"';,!: ,•.;"•, ;;'; SI;
...... ill
lli'ii:ii
toiiEL
("I i, (h I) i
iiT't: •!'
i ill n i i in i
I! IB I Mi: Ill
iiif
i Mi1"
II
lii ("IP1"!, ' 'II
•[ '::,! I • '/" •>*•'"!
''si'.;]* ;;;i, , •-«)
' ill' ill i|U". .1 Ji,! hi" ill"!
^' ^ " MICROBIOLSTANpiiLpSE^ALUATI6^^^
2. Prevent maintenance of effective disinfection throughout the distribution
system.
3. Interfere with microbiological determinations.
Of the three, the most criti«d'and'^Vpne'_which has generated most con-
cern is the possible interference of turbidity with disinfection efficiency. The
second criterion relates to complex problems of turbidity as a source of
'nutrients"and available surface for growth of microorganisms in distribution
lines, as "well as the more directrelationshipbetweenturbidity and disinfectant
demand™ Significancei"of'the "tKird 'cnterion'''h"as'been"shown at high turbidities
but an efiect in the turbidity range of 1 to 5 has not been clearly demonstrated.
Theoretical Aspects of the Relationship Between Turbidity and
Disinfection
For disinfection to be effective, there must be contact between the agent and
the organisms that the disinfectant is to kill. Some of the particles which
make up part of the turbidity may be particle-microorganism complexes in
which a microorganism is adsorbed onto a larger particle. In other cases one
or more smaller particles may "be "adsorbed onto a microorganism or a par-
ticle may completely encase a microorganism. In each of these cases varying
degrees of protection of the microorganism from the disinfectant may be
expected. [
The microorganisms of concern in this respect include pathogenic protozoa,
bacteria and viruses. SeVeral characteristics of these groups of microorganisms
must be considered relative to possible effects of turbidity on their inactiva-
tion. One of these is size. The sizes of representatives of these three groups
differ by a factor of over 100X (poliovirus 0.03 /^m diameter, E. coli 0.3 X
1-5 /im, E. histolytica cysts 5-20 /*m). This would indicate that in the case
of virus protection by turbidity, contact between the disinfectant and the
microorganism eould be prevented by much smaller complexes than would
be needed for protection of enterobacteria or protozoans. Size is also im-
portantin relationship to the significance of surface charge phenomena in
causing aggregation or adsorption of particles. These factors become in-
creasingly important as size decreases and are very important with regard
to particles in the size range of enteroviruses. Another factor of importance
is the relative innate disinfection resistance of the microorganisms them-
selves. In general, the enterobacteria are the most sensitive to disinfectants',
followed by the enteroviruses, while the protozoans exhibit the greatest re-
sistance (19).
Based on these factors, several assumptions regarding turbidity effects can
be made. In order to protect ah enterovirus from disinfectant action, a par-
ticle, whether produced by adsorption of many extremely small particles onto
one virus particle or from dissolution of fecal or other material to form
small particles some of which may completely enclose a virus particle, must
be larger than 0.03 ju,m hi diameter. Particles which will protect bacteria or
cysts musf be correspondingly larger. Thus, the particles present in water
which are of concern because of possible disinfection efficiency effects are
those ranging upwards from 0.03 /xm.
"... ^Tffie^extensive} body of literature on the physio-chemical mechanisms
involved in virus adsorption onto surfaces and the influence of these inter-
actions on the fate of viruses in the environment was recently reviewed by
Britton (5). The adsorption phenomenon is used for virus removal
in various waste water and water treatment processes (coagulation, floccula-
104
-------
TURBIDITY AND DISINFECTION/J.C. HOFF
tion, filtration and also is applied in methods developed for concentration
and detection of low levels of vkus from large volumes of water. Because of
the way in which viruses, particularly enteroviruses, are produced, excreted,
and transported (intestinal tract —^ feces —> water) and the number and
variety of particles present in natural waters it is likely that many entero-
viruses exist in the environment as adsorbed, rather than free, particles (4).
Results of Research on Protective Effects of Turbidity
Although the concept that certain particles present in water may protect
microorganisms from inactivation by disinfection certainly is logical, little
direct evidence of such protection has been demonstrated. Clarke and Chang
(9) stated that turbidity was responsible for disinfection failure which re-
sulted in the Delhi infectious hepatitis outbreak. An AWWA committee (3)
stated that "Effective disinfection can be carried out only on water free from
suspended material." Gulp (11) stated that for good viral disinfection turbidi-
ties of 0.1 JTU were preferable. Cookson. (10) stated that the ability of dis-
infection to produce virus-free water depends on removal of turbidity to no
more than 1 JTU.
A few studies have provided evidence that the presence of particulate
matter interferes with effective inactivation of viruses and bacteria in water.
In an early rather crude study, Neefe et al (15) showed that feces from an
infectious hepatitis patient when suspended hi water caused hepatitis in two
of five human volunteers after exposure to a chlorine residual of 1.1 mg/1
for 30 minutes. In a similar experiment, when the water was first coagulated
and then filtered before disinfection to the same chlorine residual, none of
five volunteers contracted hepatitis. .Change et al (7) demonstrated that enteric
bacteria and enteroviruses ingested by aquatic nematodes were completely
protected against high residual chlorine levels (80-90 mg/1) even when
more than 90% of the nematodes were immobilized. Tracy et al (18) re-
ported the persistence of coliforms in drinking water which was highly
chlorinated. The problem was most severe when turbidities were between
5 and 10 units. They concluded that the most likely explanation was that
the coliforms had been ingested by Crustacea present in the water and were
thus protected from contact with the disinfectant.
In a discussion included as a part of a paper by Clarke et al (8), W. W.
Sanderson and S. Kelly of the New York State Health Department, Division
of Laboratories presented evidence that turbidity interfered with disinfection
in a water supply. They studied an impounded water supply receiving no
treatment other than chlorination. The concentration of free chlorine residual
in samples from household taps after a minimum of 30 minutes contact time
varied from 0.1 to 0.5 mg/1 and the total chlorine residual was between 0.7
and 1 mg/1. These samples consistently yielded confirmed coliform organisms.
Turbidities in these samples varied from 4 to 84 TU's and microscopic exami-
nation showed iron rust and plankton to be present. They concluded ". . .
coliform bacteria were imbedded in particles of turbidity and were probably
never in contact with the active agent. Viruses, being smaller than bacteria,
are much more likely to escape the action of chlorine in a natural water. Thus,
it would be essential to .treat water by coagulation and filtration to nearly
zero turbidity if chlorination is to be effective as a viricidal process."
Results of our research on this problem as well as the results of other
recent research provide more definitive information regarding the relationship
of low levels of turbidity and disinfection. Our research in this area has
been designed to determine whether significant differences in disinfection
105
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
-'!/' -1
1 ft, , Ill'1"11
-1
-2
S
cc
<
>
OC
lif"1"!"!'. " 'liilll! i,i.".!,'
,-4
-5
-6
„„, ,*y
D VIRUS ONLY - 0.28 NTU
HOCI (0.50 - 0.46 mg/l)
A VIRUS * BENTONITE - 7.1 NTU
HOCI (0.50 - 0.44 mg/l)
A
n
A
D
A
D
D
n
468
TIMI=(min.)
10
12
Figure V.' iMACif IVATl"oN OF BENTONITE - ASSOCIATED POLlOVlRUS 1 BY
CHLORINE (pH6.0, 5° C).
efficiency can be specifically related to turbidities of 5 NTU and 1 NTU. All
experiments were conducted at 5°C and pH 6.0. Under these conditions vir-
tually 100% of the free chlorine residual is present as hypo-chlorous acid
(HOCI) which is the most effective chlorine species for disinfection. The
LScZab strain of attenuated ppliovirus type 1, treated to remove chlorine
demand," 'was'used. '
Studies of effects of inorganic turbidity on disinfection were done using
;|Virus "SorSe'd l!;"on"":"oentonite clay and "virus precipitated with aluminum
phosphate (A1PO4). Volumes of virus associated with bentonite or A1PO4
sufficient to produce a turbidity of ~ 5 NTU were added to stirred reactors
containing 0.5 mg/l HOCI. Centrifugation results showed that 65-80% of
the bentonite-yirus mixture was bentonite associated and over 90% of the
AlPO4-virus mixture was"".'AlPO4 associated. An example of results using
106
-------
TUBSIDITY AND DISINFECTION/J.C. HOFF
-1
-2
o
-------
MICROBIOL STANDARDS EVALUATION/ClW. HENDRlCiKS
O)
E 1
O
O
X
-1
3
g
cc
cc
'3
05
-6
s
O
© VIRUS-BENTONITE
TURBIDITY-5-5 NTU
A VIRUS-AIPO4
TURBIDITY-4.2 NTU
E VIRUS ONLY,
TURBIDITY-0.2 NTU
G
TIME (min)
Figures. INACTIVATION OF POLIVIRUS ASSOCIATED WITH AIPO4 AND
BENTONITE BY 1.5 mg/1 HOCI.
conducted at higher HOCI levels (-1.5 to 3.0 mg/1). Results of an experi-
ment comparing inactivation of clay associated, A1PO4 associated, and free
virus at 1.5 mg/1 HOCI are shown in Figure 3. Inactivation rates were very
similar, with bentonite associated virus showing a slightly lower rate of in-
activation than AlPOY associated or free virus. Figure 4 shows a com-
108
I,
-------
TURBIDITY AND DISINFECTION/J.C. HOFF
O
g 1
HEP 2 CELL-ASSOCIATED VIRUS
TURBIDITY-1.4 NTU
ARTIFICIAL HEP 2 CELL-VIRUS
MIXTURE TURBIDITY-1.O NTU
VIRUS ONLY O.15 NTU
El HEP 2 CELLS PRESENT, VIRUS ADDED
SEPARATELY TURBIDITY-1.4 NTU
3 4
TIME (min)
Figure 4. EFFECT OF CELLULAR MATERIAL ON RATE OF POLIVIRUS
INACTIVATION BY CHLORINE.
parison of disinfection rates of free virus, cell-associated virus and an
artificial cell-associated virus mixture prepared by mixing virus with cells
which had been harvested, washed, and stored at 5°C for 1-2 days. The
artificial mixture was stored ovemite -at 5°C before use. The results show a
pronounced protective effect of virus associated with cells. Also shown in
109
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
O) 2
T-T
-G
A VIRUS ONLY
TURBIDITY-0.15 NTU
© HEP 2 CELL
ASSOCIATED VIRUS
TURBIDITY 1.4 NTU
30 40
TIME (min)
Figure 5. INACTIVATION OF CELL- ASSOCIATED POLIVIRUS BY
CHLORINE (60 min.).
Figure 4 is an inactivation curve in which washed cells were added to the
disinfection reactor followed immediately by virus. This simulated conditions
in which virus inactivation was occurring in the presence of chlorine demand
created by the cellular material. The results show that the presence of chlorine
demand had no effect on the inactivation of the virus. Figure 5 shows the
results of a longer term experiment on inactivation of cell associated virus.
110
-------
TURBIDITY AND DISINFECTION/J.C. HOFF
HOCI levels declined from an initial level of 3.0 mg/1 to 1.5 mg/1, most of
the decline occurring in the first 5 minutes of the experiment. Even though
the turdidity was only 1.4 NTU, virus was detected even after 50 minutes
exposure to 1.5 mg/1 HOCI. Free virus exposed to 3.0 mg/1 HOCI was
reduced more than 5 logs in less than two minutes.
In two other recent studies, the effect of inorganic particles on virus
disinfection has been studied. Boardman (6) reported that inorganic particles
including alum, calcium carbonate and kaolin did not interfere with bac-
teriophage T7 inactivation by chlorine but did interfere with the rate of in-
activation of poliovirus. The poliovirus results are shown in Table 1. The
experiments were conducted at 22.5°C at a residual chlorine level of 0.21
mg/1. The chlorine residual determinations indicate that most of the residual
was present as combined rather than free chlorine. Although turbidity was not
measured, the amounts of alum (20 ing/1) and kaolin (50 mg/1) were
similar to the levels of A1PO4 (50 mg/1) and bentonite (35; mg/1) used
in our research. Boardman related the protection shown to extensive virus
aggregation enhanced by the^presence of particulates serving as foci for virus
adsorption. Experiments designed to show this were inconclusive. Boardman
also noted that the protective effects wens greatest in systems in which infor-
mation of particulates occurred in the presence of the virus (CACO3 & alum).
Stagg et al (17) compared the inactivation by HOCI of bacteriophage
MS-2 associated with bentonite clay and in the freely suspended state. Both
freely suspended and clay associated phages were inactivated ; rapidly but
the rate of inactivation of clay associated phage was retarded. At equivalent
HOCI concentrations ranging from 0.02 to 0.6 mg/1, approximately twice
the length of time was required for equivalent inactivation (99% ) of clay-
associated phages as for freely suspended phages. Reaction rates in both cases
were essentially first-order. Time required for 99% inactivation of freely
suspended phage was 10 seconds and for clay associated phage 19 seconds at
0.5 mg/1 HOCI at 22°C. Turbidity levels for the experiments were in the
range of 2 to 4 Jackson Turbidity Units (JTU).
Use of cell associated virus in studies of disinfection of combined sewer
overflows by chlorine and chlorine dioxide was also reported by Moffa and
Smith (13). Cell associated virus, when mixed with simulated combined sewer
overflow water was much more resistant to inactivation by ClOs than was
TABLE 1. EFFECT OF VIRUS ADSORPTION ON POLIOVIRUS
INACTIVATION BY 0.21 mg/1 RESIDUAL CHLORINE^ AT 22.5°C.
Preparation pH Percent Inactivation
in 10 Minutes
Polio Only
Polio Only
Polio + Kaolin (50 mg/1)
Polio -1- Alum (20 mg/1)
Polio + Ca CO3 (600 mg/1)
7.0
10.1
7.0
7.0
10.1
99.7
97,3
95.8
78.4
84.7
'free residual range—0.02-0.08 mg/1
combined residual range—0.12-0.14 mg/1
From: Boardman (6)
111
-------
MlCROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
freely suspended virus. The results are shown in Figures 6 and 7. Turbidity
and disinfection demand characteristics of the water were not given. However,
since the cell associated virus was in simulated combined sewer overflow
water, it is likely that a high chlorine demand was present.
More recently, we have initiated some studies on the effects of turbidity
in the range of 1 to 5 NTU on coliform inactivation by chlorine. These
experiments also were conducted at pH 6.0, 5.0°C in 0.05 M phosphate
buffer. The coliforms used in these experiments were naturally occurring
coliforms associated with solids in unchlorinated primary sewage effluent. The
effluent was collected and centrifuged at 2,000 RPM for 15 minutes to sedi-
ment the solids. The sedimented solids were washed to remove soluble chlorine
demanding substances by repeated centrifugation and resuspension in 0.05 M
phosphate buffer, pH 6.0. The concentrated washed solids, when diluted to
produce turbidities of 1 and 5 NTU contained coliform concentrations of
1 X 10s — IX 10V100 ml. Inactivation of the coliforms associated with
the solids was studied in stirred batch reactors containing 0.5 mg/1 HOC1.
Coliform levels were determined by standard coliform MPN tests. Completed
tests were conducted on all positive tubes. Inactivation of a washed culture
of E. coli (ATCC 11229) by the same level of HOCl was also determined.
, ..
SYMBOL DOSAGE mg/l
o
D.
10
10
10 -
0 30 60 90 120 150 180 210 240 270 300
TIME, seconds
Figure 6. SINGLE-STAGE DISINFECTION OF POLIVIRUS 1 WITH
CIO2 IN A NO-DEMAND SYSTEM. (ADAPTED FROM: MOFFA, 13)
112
-------
TURBIDITY AND DISIlSfFECTION/J.C, HOFF
Typical results are shown in Figure 8. It is evident that the paturally occurring
coliforms associated with effluent solids were much more resistant to chlorine
than was the washed E. coli culture. At both 1 and 5 NTU's survival ratios
were similar, with coliforms still detectable after 60 minutes exposure to
0.5 mg/1 HOC1. ;
The results of the recent virus studies described above | indicate that virus
associated with inorganic particles such as clays or A1JPO4 is inactivated
by HOC1 at rates similar or perhaps slightly lower than freely suspended
virus. The rates in both cases are essentially first order witri no evidence of an
extremely resistant fraction. In contrast, disinfection rates of cell associated
viruses were not first order, with inactivation rates beconiing slower as time
of exposure increased. This indicates that various degjrees of protection
existed with a few complexes being extremely resistant to Disinfection.
u.
a.
O
Q_
SYMBOL DOSAGE mq/l
10
10
10
10
0 30 60 90 120 150 180 210 240J 270 300
TIME, seconds
Figure 7. EFFECT OF C!O2 ON CELL-ASSOCIATED POLIVIRUS 1.
(ADAPTED FROM: MOFFA, 13)
113
-------
•Ill
• ";' • " «' 1 'I' '' fTTirWffl 19' Iff? - fIIITT 5 W\'"!!, :'T: ""! '"*; 1|l:!li;Ti " I! 5 V' "! I: ] 'f!'!! "~m
".. Si:
ir •;" i" y ;' i:* i u! •" linn1 »F •,' ,in i' nwi! •' n 11- IT; ,; , i|, i| nniri" • TJ/ ,rn • «T asm
MICROBIOL STANDARDS EVALUATION/C.W. liENDRICKS
1.0
-1
-2
o
o
at
ii
CO
-4
-6
20 30
TIME (MIN.)
40
50
O WASHED £. CPU
WASHED PRIAAARY
EFFLUENT SOLIDS
A 5NTU
D 1NTU
J_
10 20 30
TIME (MIN )
40
50
60
Figure 8. INACTIVATION OF COLIFORMS ASSOCIATED WITH PRIMARY EFFLUENT
SOLIDS BY CHLORINE (pH 6.0, 5.0°.C).
114
ll.lf ,! 'I '',' •, • i !'i i'll<» '4, .* Inl, l,i, l i111!1!1"] ,, I li1"1: < Illlli' ,, ,i, , l'i il'F / >,'! L! .1,,' '
-------
TURBIDITY AND DISINFECTION/J.C. HOFF
Discussion
In these studies, the measure of the presence or absence of viable virions
after exposure to the disinfectant was the ability of the virus-particulate
complex to initiate infection of a cell, thereby resulting in a visible plaque.
Recent studies have shown that enteroviruses associated with j solids remain
infective for cell cultures (14) and for animals (16). The mechanism by which
virus-particulate complexes initiate infection is not known. It is possible that
the complex is broken at the cell surface and the virion enters the cell as an
unadsorbed particle or that the entire complex enters the cell. It is known
that various chemicals, notably serum protein can prevent adsorption of virus
to inorganic particles and also can elute virus previously adsorbed to such
particles. Because of the nature of the assay system Giving cells) it is in-
evitable that some soluble protein will be present at the cell surface. Because
of. this, it is difficult to determine the details of cell infection by adsorbed
virus. It is also possible that viable virions, completely encased in feces, cell
debris, etc. and thereby completely protected from disinfection, may exist
and remain undetected. In a few instances in the studies we conducted, efforts
were made to release protected cell associated virus by freezing and thawing
of samples collected at various intervals during disinfection. The results of
these studies were equivocal. Because of the complex digestive processes
which occur following ingestion, such virus could be freed from encasement
and being viable, initiate infection in the host. Thus, virus which remains
viable but undetectable in cell culture systems could possibly: be detectable
in an animal assay system. - -
While we do not know how well cell associated viruses simulate viruses
as they exist in natural waters, it is likely that this model is more representa-
tive of natural conditions than the model employing freely suspended viruses.
The results of the disinfection studies of naturally occurring solids associated
coliforms show inactivation curves simiilar to those shown by cell associated
viruses. This would indicate that similar effects would be seen with naturally
occurring viruses present in such solids,
In interpreting the results of these studies, particularly the virus disinfec-
tion research, it should be kept in mind that the protective effects shown,
while they well may reflect what actually occurs during water treatment, repre-
sent a tremendous magnification of the virus levels which might be present
under "real world" conditions. Whereas in laboratory experiments, we can
begin studies with tremendous numbers of viruses (~ 10s PFlT/ml), actual
numbers present in some of the worst raw waters used as potable water
sources may contain on the order of 1 PFU/gallon.
The information presented above has been limited to the consideration of
turbidity interferences with disinfection of potable water. Several other very
important aspects of turbidity as related to health hazards associated with pot-
able water have not been addressed. These include the use of turbidity as an in-
dicator of the presence of particulates such as asbestos which may pose a
chemical hazard or as an indicator of the presence of microbiological forms
such as E. histolytica, bacterial spores,, and perhaps Giqrdia lamblia which
have or are believed to have a high innate resistance to disinfection. These
aspects must be taken into account in considering requests! for allowing
turbidity limits higher than 1 NTU.
115
-------
MICROBIOL STANDARDS EVALUATION/C.W. iSNDRICK
Conclusions
I , . j ' » : I''1' ! j' ] , !' , " . [ -: , ' •" '•
As stated previously, turbidity measurement is well established as a useful
operational guide for assessing" the performance of various processes such as
coagulation, sedimentation and filtration in water treatment plant operation.
Unfortunately because of its lack of specificity in determining the nature of the
turbidity particles present, turbidity does not appear to be the ideal indicator of
_•*• •_ J! A _ t • 1! .*._.. _.C __ A4.nl>1A •••*•*.». •*•<•**. n-1-n.nri 4-t-» a. C n+avfckranf+at fff "f llT'Kirllt'XT' "Wit n filCITI-
disinfectability of potable water
since the interference of turbidity with disin-
fection depends much more on the types of turbidity present than the number
of turbidity units present. The Bacteriological data derived from the primary
effluent studies indicate similar [disinfection curves at .turbitities of 5 and 1.
Based on this, one could assume tiSat'water Seatment processes which reduce
this type of turbidity to 1 NTIJT rather than 5 NTU would result in an ap-
proximate 5 fold increase in safety factor. In other words the disinfectant
resistant portion would still be jpresent but at a 5 fold lower concentration.
Although cell-associated virus studies were not conducted at both turbidities,
it seems reasonable that the saiiie observation could be made in the case of
turbidity caused by clays and AJ1PO4. Interference with disinfection efficiency
is either negligible or relatively! minor. It is likely that little in the way of
increased safety factor would be gained from reducing turbidity caused by
these types of particles from 5 NTU to 1 NTU.
While it is evident that turb
:dity is at present the established parameter,
future consideration should b& given to other parameters as a possible
measure of participate interference with disinfection for use as regulatory
criteria. Parameters such as total organic carbon or nitrogen in the suspended
solids present in a particular water might be used to determine whether the
solids are organic or inorganic. Determination of chlorine demand of the
solids present in a particular water might also be used to distinguish in-
organic from organic particles.
Further consideration of the
relevance of coliforms as indicators of viral
safety of disinfected water may also be in order. The results of the studies
on inactivation of naturally occurring coliforms associated with primary
effluent solids indicate that sjich coliforms are far more resistant than
freely suspended coliforms. Thus, the coliform group may be a more
rigorous disinfection indicator 'than is indicated by disinfection studies of
pure cultures. !
CFERENCES
1. American Public Health Association. 19751 Standard-methods for the ex-
amination of water and wast ewater. 14th ed. American Public Health Asso-
ciation, Inc., New York.
2. American Water Works Association. 1971. Water quality and treatment.
3rd ed. McGraw Hill, New York.
3. American Water Works Coriiniittee on Viruses in Water. 1969. Viruses in
', ; ,'; ' water. JA'WWA 61:491-494. \ ^ "[ '_'; ,",„ .' ;
4. Berg, G. 1973. Reassessment I of the virus problem in sewage arid in surface
and renovated waters. Progress in Water Technol. 5:87-94.
5. Bkton, G. 1975. Adsorption of viruses onto surfaces in soil and water.
Water Res. 9:473-484.
6. Boardman, G.D. 1976. Protection of waterborne viruses by virtue of their
affiliation with particulate matter. Ph.D. Thesis. Univ. of Maine.
7. Chang, S.L.,G. Berg, N.A. J Clarke, and P.W. Kabler. 1960. Survival, and
protection against chlorination, of human enteric pathogens in free-living
nematodes isolates from watejr supplies. Am. J. Trop. Med. Hyg. 9:136-142.
-------
TURBIDITY AND DISINFECTION/J.C. HOFF
8. Clarke, N.A., G. Berg, P.W. Kabler, and S.L. Chang. :l 964. Human enteric
viruses in water: source, survival, and reir ovability. International Conference
on Water Pollution Research, Pergamon Press, Oxford.
9. Clarke, N.A. and S.L. Chang. 1959. Enteric viruses in water. JAWWA
57:1299-1317.
10. Cookson, J.T., Jr. 1974. Virus and water s'upply. JAWWA. 66:707-711.
11. Gulp, R.L. 1974. Breakpoint chlorination
for virus inactivation. JAWWA
65:699-703.
12. Hach, C.C. 1972. Understanding turbidity measurement. Indust. Water
Eng. 5:18-22.
13. Moffa, P.E. and J.E. Smith, 1974. Ben,ch-scale high-rate disinfection of
combined sewer overflows with chlorine and chlorine dioxide. Final Report
EPA Project #8-802400. United SCates Environmental Protection Agency,
Washington, D.C.
14. Moore, B.E., B.P. Sagik, and J.F. Molind, Jr. 1974. Viral association with
suspended solids. Water Res. 9:197-203.
15. Neefe, J.R., J.B. Baty, J.G. Reinhold, and J. Stokes, Jr. 1947. Inactivation
of the virus of infectious hepatitis in drinking water. Am. J. Public Health.
37:365-372. '.
16. Schaub, S.A., and B.P. Sagik. 1975. Associition of enteroviruses with natural
and artificially introduced colloidal solids in water and infectivity of solids-asso-
ciated virions. Appl. Microbiol. 30:212-222. '•
17. Stagg, C.H., C. Wallis, and C.H. Ward. 1977. Inactivation of \clay-associated
bacteriophage MS-2 by chlorine. AppL Environ. Microbiol. 33:385-391.
18. Tracy, H.W., V.M. Camarena, and F. Wi|ng. 1966. Coliform persistence in
highly chlorinated waters. JAWWA. 55:1151-1159. '
19. White, G.C. 1972. Handbook ,oi .Chlorination. Van Nostrand Reinhold Co.,
New York. i
117
-------
II 'INiW "II ':,' ' I I' ' I ''lif'l! "'
iiiiiii!j"i» " j mi nil
'I'l'lIU'illl! Ililllllli I I1 i1 '•*> '"
ifi'i
II II
li I in
f": !'!'
I'B i'iv
-------
WASTEWATER REUSE/H.W. WOLF
Wastewater Reuse
ami the
Problems of Water-Borne Disease
Harold W. Wolf
Environmental Engineering
Texas A&M University
College Station, Texas 77843
Introduction
Much has been written about wastewater reuse, a topic that seems to be-
come more popular with time. Since there are many types of reuse applications
made of wastewater, perhaps the first thing we should do is; define the type
of reuse that concerns this presentation: direct, potable reuse. But this still
isn't as sharp a definition as I'd like. Webster (10) says that potable means
"suitable for drinking." I personally believe that there may be other uses of
. water that are more demanding than just drinking (11). For example, a water
could conceivably not present any trouble to a person drinking it if it contained
a substantial population of Staphylococcus aureus, but prepare a salad with it
and allow it to incubate a few hours and the results could be quite different.
Hence, I usually generalize the type of reuse that I have in mind as: for
intimate human usage.
Another definition is essential to our discussion today. Webster (10)
defines disinfection as "the freeing from infection esp. by destroying harmful
microorganisms." The Glossary of Water and Wastewater Control Engineer-
ing (3) defines it as "the art of killing the larger portion of microorganisms
in or on a substance with the probability that all pathogenic bacteria are killed
by the agent used." The Glossary definition thus:
1) introduces the concept of disinfection being a rate function—which it is
' 2) introduces the concept of probability, which is also proper, and
3) excludes viruses.
When applied to drinking water, the Glossary definition must be updated
to include pathogenic viruses. Until it is, we must be careful that we define
the interpretation of disinfection we are talking about.
Organisms of Concern
Since the wastewater reuse under consideration concerns the most intimate
human usages, and since wastewater can conceivably include anything, we
must address virtually all pathogenic agents. It would be inordinately naive
to be concerned only with enteric organisms even though these are expected
to be the most abundant.
From a disinfection point of view, the organisms of concern can be grouped
as follows: .
Bacteria
i Vegetative
Encapsulated
Non-encapsulated
Spores :
.-•-.-. .Fungi & Yeasts
Rickettsiai
119
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
Viruses
Free
Within tissue
Protozoans
Trophozoites (vegetative state)
Cysts (spore state)
Helminths & Flukes
Adults or intermediate forftis
" ' ' " ..... "' " "
. . . . , .
I believe such a grouping of microorganisms is more useful for our purposes
than a long listing of the many genera and species of microorganisms that
can cause disease. For an excellent discussion of waterborne pathogens, see
Geidreich's chapter in Mitchell's book (5). I would be reluctant to make up
a list of pathogens because those I overlooked would surely be the next to be
involved in a waterborne outbreak.
One can look at the grouping presented and ask, "How well does the coli-
form group represent all^of these?" The answer is obviously "Npt nearly well
> enough."
One can also pick out the most resistant forms rather quickly; e.g., the en-
capsulated vegetative bacteria, spores, viruses protected by tissue, cysts, and
ova. A classically durable ovum is that of the roundworm, Ascaris
lumbricoides. -
Lastly, one can look at the list and realize what a formidable task, we're
expecting of any one disinfectant. Viewed from this perspective, hypochlorous
acid (HOC1) looks quite spectacular.
'''Problems . . ' ' .......... '
1) Drug resistance
Any review of water-borne disease problems that might be associated
with potable reuse of wastewater effluents should, in my opinion, start
with the following statement from Grabow, Prozesky, and Smith of
South Africa:
"The therapeutic value of antimicrobial drugs is diminishing due to
the rapid increase of resistant bacteria. A current prominent type of
resistance is mediated by R factors (extra-chromasomal nucleic acid
elements) which may cause high level resistance to many drugs. These
factors may also provide resistance to other antibacterial agents such
as u.v. light, heavy metals, bacteriocins and phages, and may enhance
the virulence and infectivity of pathogens. Intestinal Gram-negative
bacteria like coliforms may act as reservoirs of R factors and transfer
them to pathogens. There is evidence that sewage polluted water may
play an important role in the spread of coliform and other bacteria
carrying R factors. Since coliforms have joined forces with bacteria
increasingly involved in disease, they can no longer be regarded as
harmless indicators of fecal pollution. This calls for a re-evaluation of
water quality standards and for more advanced purification of sewage
prior to discharge into the en viroment" (6).
Drug-resistant coliforms are not the exclusive domain of South
Africa. A few weeks ago, Dr. Sagik (Dean of Science, University of
Texas Health Science Center, San Antonio) mentioned that he was
looking for a coliform to use in tracing studies and searched the Austin
Wastewater Treatment Plant for coliforms resistant to streptomycin +
penicillin + nalidixic acid and found from 1:10 to l:iOO coliforms
„'!' , ' ....... ,, ......... , ..... „ , ;'.;.' .; ....... ; , j ; '
, 120
iiiiiiiiit jsst\.:',tja ..... i:;,.: '.i!1;!; ....... ;„;„ jj ^ ........ ;i; .,f ^..atii ..... iiiiiii.!: j j.. *. ,
-------
WASTEWATER REUSE/H.W. WOLF
resistant to all three (Sagik, B. Telephone communication, Feb. 8,
1977).
Some of us in Texas look at our long border with Mexico with some
trepidation because we know how readily available antibiotics are across
that border, we know that is a large movement of people across that
border, and we read periodically in Morbidity and Mortality—and in
Science (4)—increasingly frequent accounts of illness; caused by anti-
biotic resistant strains of bacteria—many of which occur across that
border.
2) Viruses ;
I think that public health engineers, in general, are fairly comfortable
with the fact that viruses can be inactivated by HOC1. Now and then
we see an article that questions our faith (7). The most recent such
occurrences have been discussed by Akin and Jakubowski (1).
The question of virus-presence in finished drinking waters is going to
be with us a long time. It will be with us until we can cultivate the virus
of infectious hepatitis and it will be with us until we learn a great deal
more than we know now about the slow-growing viroids. I am not
aware, however, that these organisms diifer structurally from the larger
viruses, i.e., they all appear to consist of a nucleic acid in a protein coat.
Consequently, we should not expect the viroids to differ very much from
the larger viruses in their response to HOC1. However, other factors
could contribute to a finding of significant differences, e.g., a small virus
could be protected from the action of HOG1 by being imbedded in a
particulate which would not be; large enough to provide the same pro-
tection to a larger virus.
Unfortunately, HOC1 is a factor in the production of trihalo-
methanes—substances that we would like to minimise the presence
of in our drinking water. Fortunately, we have three other powerful
agents that might be considered. Morris (8) reports that ozone is even
more effective than HOC1, and that viruses—although more resistant
than vegatative bacteria—can also be readily handled by, ozone. Chlorine
dioxide appears capable of handling viruses, and so does one other tech-
nique which is not discussed very often—high pH (12) With all these
techniques available, there appears to be ample methodology for the
handling of viruses in a reuse situation. The concern really boils down
to other qualities of the water and whether these might interfere with
the disinfection process.
3) Pseudomonads
Most of us .are aware of the ubiquity_of the pseudomonads. I think
they can be found almost anywhere that you look. Alexander (2) states
that only a small number of bacterial species can bring,about denitrifica-
tion, and of these; pseudomonads—including the pathogenic Pseudo-
monas aeruginosa—are the more important. Possibly because of then-
relationship to denitrification, pseudomonads are important organisms
in soils and in wastewater treatment.
Even though the pathogenicity of Ps. aeruginosa is determined more
by the patient's state of resistance than by any inherent virulence
(9), the thought of creating conditions for these microorganisms to
proliferate in our distribution and plumbing systems is cause for some
soul-searching in any discussions of reuse for potable purposes.
121
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
Recommendations
Drug-resistant bacteria
1. EPA should support periodic surveys of the drug-resistance of
coliforms isolated from public water supplies. By doing this, a
picture should evolve which will better define the potential of the
problem to our water supplies.
2. EPA should not adopt new coliform standards, but should take
a hard-nosed stance about any infraction of existing standards.
The monitoring program for example, might call for the follow-
ing:
a) If a coliform determination is >1 or >2.2 per 100 ml (de-
pending upon methodology), immediate resampling should be
done. (If it is above 4/100 ml, public notification should be
made)
b) |f the resampling is >1 or 2.2 per 100 ml, both public
notification and immediate study should be done.
Viruses
1. Any studies of virus presence in finished waters must pay more at-
tention to those confounding factors that are known to interfere
with the disinfection processes used. Turbidity, pH, and tempera-
ture are not the only factors; NH3 organic-N, chlorine species, a
measure of organics, and particle-size are also important (in-
tuitively).
2. Routine coliphage determinations at select water treatment plants
should be instituted. If the disinfection process is working as it
should, there should be no viable coliphage in the finished water.
If coliphage are found, why look for pathogenic viruses? The con-
ditions for pathogenic virus survival will have existed.
3. EPA should support basic studies of viruses and basic studies of
• virus disinfection by a variety of disinfecting or inactivating
methods.
Pseudomonads
The .problem which I have referred to as a Pseudomonas problem should
really be addressed as a problem of maintaining water quality in distribu-
tion systems. Recommendations for EPA activity here would include:
1. Basic studies of water quality and organism growth.
2. Strict application of the rules of sanitary surveys with regard to
covered reservoirs, dead ends, adequate pressures, main disinfection,
etc. i_ '
-------
WASTEWATER REUSE/H.W. WOLF
2. Alexander, M. 1961. Introduction to soil microbiology, p. 299. John Wiley &
Sons, Inc., New York.
3. American Public Health Association. 1969. Water and wastewater control
engineering. American Public Health Association, Inc., New York.
4. Calliton, B.J. 1976. Drug resistance growing worse. Science 194:1396.
5. Geldreich, E.E. 1972. Water-borne pathogens, p. 207. In: Mitchell R., (Ed.),
Water pollution microbiology. Wiley-Interscience, New York.
6. Grabow, W.O.K., O.W. Prozesky, and C.S. Smith. 1974., Drug resistant
coliforms call for review of water quality standards. Water Res. 5:1-9.
7. Liu, D.C., H.R. Seraichekas, 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, pp. 171-195. In: Snoeyink, V., and V.
Griffin (Eds.), Virus and water quality: occurrence and control. Proceedings
of the Thirteenth Water Quality Conference. University of Illinois Bulletin,
Vol. 69, No. 1, Urbana, Illinois.
8. Morris, J.C. 1976. The role of ozone in water treatment. Paper presented at
the 96th Annual Conference, AWWA, New Orleans, Louisiana.
9. Smith, D.T. and N.F. Conant. 1960. Zinsser microbiology, p. 425. Appleton-
Century-Crofts, Inc., New York.
10. Webster's New Collegiate Dictionary. 1976. G & C Merriam Company, Spring-
field, Massachusetts.
11. Wolf, H.W. 1972. The coliform count as a measure of water quality, p.
344. In: Mitchell, R. (Ed.), Water pollution microbiology. Wiley-Inter-
science, New York.
12. Wolf, H.W., R.S. Safferman, A.R. Mixson, and C.E. Stringer. 1974. Virus
inactivation during tertiary treatment. In: Malina, J.F., Jr. and B.P. Sagik,
(Eds.), Virus survival in water and wastewater systems. Center for Research
in water resources, The University of Texas, Austin, Texas.
123
-------
-------
ENGINEERING CONTROL PRACTICES/P.D. HANEY
EVALUATION OF MICROBIOLOGICAL STANDARDS FOR
DRINKING WATER :
Engineering Control Practices
.by
Paul D. Haney
Black and Veatch, Consulting Engineers
Kansas City, Missouri 64114
Introduction .
It seems appropriate to hold this symposium at a time when one of the
foremost citizen concerns of at least half the Nation is water, or perhaps,
more accurately, the lack of it. Regional water crises are not new. In my
lifetime, I have acquired some knowledge of several, and have had personal
experience with serious water shortage and water quality problems in the
West and Midwest. Assuming my remarks are of some interest and
significance to symposium participants, please, in assessing them, keep in
mind my western and middlewestern perspective (or provincialism, if you
will).
My experience has been in the fields of public health — state sanitary engi-
neering and the. U.S. Public Health Service — teaching certain sanitary
engineering subjects .and, in more recent years, the practice of engineering
in a professional engineering firm specializing in energy and water projects.
.1 emphasize energy because water and energy are closely intertwined. Each
impacts the other.
With this brief background, I shall now proceed to the business at hand,
engineering control practices, with the full knowledge that I am tackling a
complex matter that cannot be brought into focus in 20 minutes, or even
20 hours or 20 days. Therefore, please bear with me while I report briefly
on what is expected of professional engineering and water utility managers
and operators in the vital fields of public water supply and water quality
control. I shall make occasional reference to Dr. Murphy's law and dts
corollaries.
Drinking Water Standards— Past, Present, and Future
To navigate successfully, one must have good charts, know where he is
going, where he is, and where he has been. There must be a few questions
about where we are and where we are going or we wouldn't be holding this
symposium. Hopefully, at its conclusion, we shall have some better charts
than now. Let's examine briefly where we have been and where we seem to
be.
The first U.S. drinking water standards that enjoyed, official status were
those established by theJJ.S. Public Health Seryicejn 1 9 1.4.-Q&en called the
"Treasury Standards", they covered only the bacteriological aspects of water
quality. A most interesting feature_ of these standards was the inclusion of a
Wnyjjga&Jtlkopped?
What is its jnipOT-tanee? - .
TEe~1914 standards bulletin, which sold for five cents, briefly discussed the
sanitary survey and its importance. The sanitary survey, well known to all
older sanitary engineers, has been a unifying influence in the standards for
over half a century.
125
-------
t*-
TABLE 1. CHANGING STANDARDS FOR DRINKING WATER.
[Tabulated numerical values are in milligrams per liter (mg/1) except for
Bacteria, pH, Taste/Odor, Turbidity, Year, Color]
1914 1925 1942 1946 1962
1975-1977 *
Conform Bacteria/ 100 ml
Total Bacteria/ml
Copper
Iron
-Lead
Manganese
.Magnesium
Zinc
Arsenic
j Chromium
( Selenium
1 Barium
^-\ OJ_Y.2 iim
V^SClIIllUT II
; Mercury
Silver
;pH
Chloride
: Sulf ate
Total Dissolved Solids
Taste/Odor (Threshold Number)
Turbidity (Turbidity Units)
Color Units
<2 <1
100 —
— 0.2
— 0.3
— 0.1
— 0.05
— 100
— 5.0
— —
— —
_ _ _
— —
""•--"— ' T
— —
,„ ^ , -....
— 250
— 250
— 1000
— None
— • "Clear"
— "Colorless"
-------
Hydrogen Sulfide — —
Fuoride —
Phenols __
Cyanide —
Nitrate (N)
Foaming Agents —
Carbon Chloroform Extract . , —
Radioactivity —
; - i i
Pesticides —
i
Sanitary Survey ; ; Yes Yes ' ;
Viruses :
Asbestos . |
Chlorinated Organics ; : j \
What is next? i ; ' '! :
* Federal Register, Dec. 24, 1975; July 9, 1976; March 31, 1977.
1.0
0.001
Yes
1.5
0.001
Yes
0.8-1.67
0.001
.01
10.
0.5
0.2
Yes
Yes
0.05
1.4-2.4
10.
0.5
Yes
.00
o
z
PI
w
o
o
i
b
33
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
Table 1 is an attempt to summarize past Public Health Service (1914-62)
and present Environmental Protection Agency standards. This table reflects
some sixty years of changing culture. It is apparent that bacteriological
quality is still of concern but the bacteriological standards have not changed
greatly since 1914. Probably Ian engineer who could design a plant meeting
the 1914 coliform standards Icould readily design one that could meet cur-
rent coliform standards although much has been learned about water supply
and treatment since 1914. The table also tells us that, over the years, there
has been greatly increased concern over things of a chemical, or better, a
microchemical nature. This upsurge in interest in microchemistry is essentially
a post-World War II happening. This era was marked by an upsurge in in-
dustrial chemistry, notably petrochemicals, and wastes associated with their
manufacturing and use were bound to have an impact on the water resources
of the world. Parenthetically, it should be noted that these new chemical
products were welcomed by consumers who found them effective and con-
venient. Accompanying this industrial revolution was almost overwhelming
progress in analytical chemistry, which is truly the cornerstone of knowledge
about this complex, universal, solvent we call "water". We have, in about 30
years, moved from the era of'[microbiology into microchemistry. Yet, with the
emphasis on chemical quality, we must not forget, not for a moment, the
fundamental and continuing; importance of microbiological water quality.
Speaking very generally, microbiological contamination produces acute health
effects. In a susceptible population, water-borne disease will often manifest
itself in a matter of days or weeks. On the other hand, usual microchemical
pollution; i.e., excluding chemical accidents involving massive contamination,
is chrome m nature, sometimes requiring years to make its effects felt. I
repeat, these effects, while very serious, are usually long-term rather than
short-term. This, of course, makes the investigation of microchemical con-
tamhianfs exceedingly difficult.
Regarding the future of our drinking water standards, I find the crystal ball
quite cloudy. I am certain, however, that they will reflect the inexorable march
of science. They will be more complex, not less. The difference between the
standards of the year 2000 and those of the present probably will be greater
than the difference, noted in the table, between 1914 and 1977. (The pro-
posed upper limit for pH given in the new standards seems unreasonable.)
I would be remiss, if I did hot remind you that at least one very important
facet of the standards does not lend itself well to numerical expression. Con-
sequently, it is not included in the table I have presented. This is the state-
ment, and I quote from the| 1962 standards: "The water supply should be
obtained from the most desirable source which is feasible, and effort'should
be made to prevent or control pollution of the source. If the source is not
adequately protected by natural means, the supply shall be adequately pro-
tected by treatment". ;
The Microbiology Standard
This symposium concentrates on microbiological water quality, which is
quite appropriate. I am sure the organizers recognized the impossibility of dis-
cussing microbiology, per se.jln thinking about the microbiology of drinking
Water, it is essential to think! of many other matters, not the least of which
is disinfection. As is so often the case in this imperfect world, the solution
of one problem can create another. The notable example is chlorination^ for
bacteriological quality improvement which can be attended by potentially
128
-------
ENGINEERING CONTROL PRACTICES/P.D. HANEY
harmful, side reactions. Chlorine has a good. bacterial toxicity but a poor
chemical specificity, and this fact is very impoiitant in many situations.
Over the years the coliform test has served us well and I have little doubt
that it will continue to do so. It is by no means perfect, but 'what is? The
standard test which is the heart of the standard itself may be criticized as
covering too broad a spectrum, too difficult and time consuming, too sensi-
tive in some respects, not sensitive enough in others. True, the standard
test has on occasion caused a false alarm and'much embarrassment but this
has been the result of trying to short-cut the standard procedure and not that
of any inherent defect. Yet, for all of this, the coliform test and the coliform
standard .have certainly been a .major force in our conquest of epidemic
disease. ; i
As a state sanitary engineer, I was once responsible for checking the quality
of some 400 public water supplies. My predecessors had done a good job and
in general these supplies were quite "satisfactory from a sanitary standpoint.
Most were from protected wells. Water from [surface sources, With one ex-
ception, was filtered. Water samples collected by health department personnel
and those shipped to the department laboratory by water utility personnel
were competently examined and the results promptly transmitted to the
utility. "Bad" samples, i.e., those showing coli'forms, were reported by tele-
phone or telegram and confirmed by letter. Quarterly summaries of the bac-
teriological quality of all supplies were prepared and posted. It was dismaying
to note that each quarter, some 100 of the supplies failed to meet the drink-
ing water standards. What was the reason.for this rather alarming state of
affairs? I hate to admit it, but I couldn't answer this question .then and I
can't now. I can only make a few clarifying comments. As I recall, none of
the surface water supplies, all subject to microbiological contamination, and
treating waters containing from about 10 to; far above 100,000 coliform
organisms (MPN) per 100 ml, ever failed to meet the, standards. Neither did
those ground water supplies where fairly extensive treatment systems had been
installed; e.g., iron removal, softening. The principal things in common about
these utilities were filtration and chlorination. O'f these two, I did then and do
now, believe chlorination was of major importance. Invariably, the supplies
that failed to meet the standards Were unchloirinated supplies derived from
wells. In general, the wells were properly located and protected (there were,
of course, exceptions) and they supplied, water that met the bacteriological
standards. However, we encouraged water utility personnel to obtain samples
from various points throughout the distribution system and they complied.
The problem of "bad" samples could, in general, be identified with the fol-
lowing:
1. Poor sampling :
a. bad technique :
b. improper sampling point; e.g., a fire, hydrant.
2. Water main, service line, plumbing repairs; failure to flush and disinfect.
3. Tank painting and repairs; failure to flush! and disinfect. Improper vents
and access covers. ' :
4. Well, well pump, etc. repairs; failure to flush and disinfect.
5. Gross connections; backflow problems.
6. Low or negative pressures in system because of operating defects or in-
adequate hydraulic capacity. ;
7. Ground storage repairs; improper vents, covers, access manholes.
Other than bad sampling, which is a matter of training and judgment, the
problems seemed to arise principally from mindr contamination in the distri-
bution system. Now, there were probably about 4s many defects in the distribu-
129
-------
FI,
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
tion systems of surface sources as there were in ground sources. Yet the
surface supplies had good records while many of the unchlorinated ground
supplies had poor records. The difference apparently was chlorination. The
results of a campaign to secure chlorination of all supplies supported this.
When chlorination was applied, a system that formerly gave an unacceptably
high proportion of bad samples suddenly turned around and consistently met
the standards. Simple, continuous chlorination of all'wells made the difference,
which was dramatic. In fact, the clean-up was so sudden and complete that
the state sanitary engineer was accused of falsifying bacteriological data simply
to prove his point that chlorination would go a long ways toward improving
water safety. The conclusion: the bacteriological standards can be met, in fact
exceeded, by the exercise of reasonable care in system operation coupled with
simple continuous chlorination. To summarize, most of the supplies involved
were acceptable until they entered the distribution system. Here, minor con-
tamination occurred from a multiplicity of sources. Chlorination was gen-
erally adequate to insure compliance with the standards then in force, which
were essentially the same as those published by EPA in 1975. As a brief
epilogue, I can report that the state's program of insistence on chlorination of
all supplies was continued and as of today, all public water supplies, over
700, are chlorinated.
As to surface waters, the program included "selling" free-residual pre-
chlorination to surface water utilities. Generally, this was well received.
Chlorine should be recognized as a water treatment chemical. It is more than
a disinfectant. Prechlorination to a free residual provides values beyond those
associated with excellence in disinfection.
I do not expect to pursue the matter of nematodes in drinking water but they
have been found and they may be more common than we realize in supplies
derived from some surface waters. Their source probably is the raw water,
but the possibility that they may find a "home" in the treatment plant cannot
be excluded. Nematodes are macroscopic, easily detected, and resistant to
chlorine. Their greatest health significance seems to be the fact that they may
ingest pathogenic bacteria and then serve as a carrier and protector for these
pathogens. The water industry is very much aware of nematodes. They are
indeed a "can of worms". Chang and others have conducted extensive in-
vestigations of nematodes in water.
Coliform criteria have been proposed by organizations other than EPA.
The American Water Works Association has established "quality goals" for
drinking water and the AWWA rationale states: "Modern disinfection control
procedures are such that a practical goal can be destruction of all coliform
organisms". The quality goal is "no coliform organisms". This is, I think, an
admirable goal, but proof of compliance probably would be virtually impos-
sible. We cannot deal in absolutes in the water safety business.
The City of Chicago, well known for excellence in the control of water treat-
ment systems, adopted a "Chicago Standard" of not more than 0.5% of 10
ml standard portions positive in contrast to an allowable 10%.
There is some relationship between raw water quality and that of the
finished water. Over the years, guidelines for maximum allowable raw-water
coliform concentrations have been recommended. These have posed problems
and, in large measure, have been ignored, especially on "flashy" midwestern
streams where coliform levels are usually fairly low during low and moderate
flow periods but go out of sight during high flow periods. The latest coliform
criterion for raw waters is that of the National Academy of Science—National
Academy of Engineering published in the "blue book" in 1973. This calls for
. a coliform limit of 20,000/100 ml (2,000/100 ml fecal coliform). It should
130
-------
ENGINEEMNG CONTROL PRACTICES/P.D. HANEY
be understood that this limit is based upon a "defined treatment process", as
specified in the NAS-NAE publication "Water Quality Criteria 1972". With
additional treatment, a much higher coliform level can be handled without
difficulty.
We need to understand that "coliforms" represent a fairly broad spectrum
of organisms derived from many sources, not necessarily disease related.
Studies (1970) by the Department of Health and Environment, State of
Kansas, (Soldier Creek Basin), have shown that the "heaviest bacterial pollu-
tion (coliforms) existed during high flow periods" when dilution was the
greatest. Natural and agricultural sources were obviously significant contrib-
utors. The watershed had an area of 290 square miles and consisted prin-
cipally of some 400 farms plus a 120 square mile Indian Reservation, largely
undeveloped. There were no "point sources" of pollution in the watershed;
i.e., no municipal or industrial outfalls arid no confined animal feeding opera-
tions.
I am especially familiar with water treatment plants located along the Mis-
souri River between Omaha, Nebraska and the Missouri-Mississippi confluence
at St. Louis. These plants at times "see" very high apparent coliform densities
yet have no difficulty meeting the bacterial standards. They are, of course,
operated by competent personnel and provide multistage treatment including
'pre and post chlorination. Many of the plants also include lime softening, a
process long known for its effectiveness in the reduction of bacterial densities.
I believe the coliform test has been and still is a valuable water quality
monitoring tool. It has some shortcomings which are well known to this audi-
ence. Despite these, if I were faced with a decision regarding water quality
and the constraint of being allowed only one test, I believe the coliform test
is the one I would select. In applying the test to treated water, one short-cut
to sure trouble is failure to carry the test through the "completed" phase when
coliforms are. "confirmed" in a finished water in which they are normally not
found. One of my gripes about the standard test is the "confirmed" designa-
tion for the second phase. The word "confirmed" is a strong word and
easily misinterpreted. Coliforms are not "confirmed" until the "completed" test
is conducted—a rather complex and time-consuming procedure. I wish we
could come up with a truly descriptive word for the "confirmed" test.
The relationship between coliform density and pathogens is understood
principally in a qualitative sense. Some thirty years ago a paper was published
in which the authors, Kehr & Butterfield, considered the relationship between
coliforms and enteric pathogens. For a typhoid rate of .01 to 30/1000 popu-
lation/year, Kehr & Butterfield suggested a Salmonella typhi (E. typhosa)
per million coliforms ratio of 3 to 120. They presented evidence that this
ratio might remain relatively constant through natural purification processes,
approximating 99.9%. They concluded that their studies ". . . emphasize the
basic value of the coliform test as an indicator of the possible presence of
pathogens and indicate that a very real danger may exist when coliforms, in
even moderate concentrations, are present. The factor of safety provided by
the ratio of a million or so coliforms present for each E. typhosa would, it is
believed, take care of the usual fluctuations in the ratio of E. typhosa to coli-
forms provided the density of coliforms , . . be kept quite low or eliminated
by methods which reduce the general bacterial population". I commend this
paper to you for its interest and the imagination displayed by the authors, both
of whom I knew and liked as individuals and respected as competent scien-
tists.
In regard to viruses, Berger, et al, .reporting for the American Society of
Civil Engineers in 1970, summarized water treatment practice as follows:
131
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
i
"Ideally, the water treatment plant should be capable of reducing
turbidity to less than 0.1 Jackson Unit and provide an HOC! residual of
1 ppm after contact period of 30 minutes. In practice, it may be assumed
that for waters having a pH less than 8, an acceptable level of protection
will be provided by treatment producing a turbidity less than 5 Jackson
Units and a free .chlorine residual of 0.3-0.4 ppm with a contact period of
301 minutes, "'for waters with higher pH, either the chlorine residual or
the; contact period should."be increased By half. Because the water treat-
ment plant is the major line of defense against introduction of viruses
into the drinking water system, it is necessary to emphasize the absolute
necessity of constant vigilance in careful plant operation."
I believe these recommendations still provide sound guidance to the engi-
neering profession.
My conclusions: Chlorine is a proven disinfectant. Prechlorination to a
free residual provides excellence in disinfection and is a valuable treatment
adjunct. Ground water supplies need at least marginal chlorination as a
safety factor. Such supplies, if not chlorinated, will have difficulty complying
1 with existing bacterial standards, principally because of "problems" associated
with distribution. As of now and assuming we can continue to use chlorine as
in the,past, I see no serious problems in complying with the intent of the
, •„,- s ; . ," • ;; standards. • ... , ,
Supplementing these remarks, I do foresee some problems for the small
utilities who collect only one sample per month. Chance contamination re-
sulting from faulty collection, poor judgment in selecting the sampling point,
and laboratory errors can easily throw a small utility out of compliance. This
sets up a chain reaction of notification that could prove extremely burdensome
and possibly result in loss of public confidence if repeated frequently. Public
notification has both good and bad features. It is a delicate matter. Judgment
is involved in this entire matter of standards compliance and we all know the
futility of trying to write judgment into any sort of law, rule, regulation, or
guideline. Some better publications about the standards would be helpful. I
commend Region VII, EPA, for their booklet titled "Answers to Questions
About the Safe Drinking Water Act". Hopefully, the standards themselves will
.be pubished in better format than that provided by the Federal Register,
which is poor. .
Water Distribution
Here is a real problem. The distribution system, the most costly portion of a
public water supply, has in full measure been treated like a "poor relative", in
the past. .This attitude is now changing. There can be many troubles in the
distribution system. I have already mentioned some of them. Dr. Murphy's
law, which tells us that: "In any field of scientific endeavor, that which can
go wrong will go. wrong" certainly finds application in the water distribution
system.
The distribution system presents a large water-pipe interface. I once roughly
.estimated this at 0-6 to 0.7 acre per 1,000 population. Strange things can and
do happen at this very large interface. Corrosion is one, but hi keeping with the
theme of this conference, I shall emphasize biological effects. These effects,
which produce dramatic water quality changes, are identified in the literature
and reach their peak in distribution zones of poor circulation, notably "dead
ends". Here chlorine disappears and remaining bacteria make the most of
their opportunity to grow and produce changes, all of which are displeasing to
the water customer. Perhaps there is no health problem but the customers
•• 132
-------
ENGINEERING CONTROL PRACTICES/P.D. HANEY
don't like turbid, smelly water and they won't suffer it in silence either. On the
other hand, perhaps there are health problems. These bacteria are seldom
studied to any extent. Possibly they have greater direct or indirect health sig-
nificance than we realize. An indirect effect that has been identified is inter-
ference with the standard coliform test. This reminds me that the present and
past standards, the 1914 edition excepted, are silent on the matter of the
total numbers of bacteria allowable. I know there are good reasons for this
omission; but, nevertheless, consider it unfortunate. As an alternative, I hope
EPA will strengthen its statements and recommendaions as to the value of the
"standard plate count" and will also recommend a guide number. If I were
running a water system, I would certainly include the plate count as a routine
test (9).
Another problem in water distribution, the significance of which has not
been generally appreciated, is that of bacterial slime formation on the interior
of water transmission mains. Th|s is serious for at least two reasons; the hy-
draulic capacity of the pipe is significantly reduced; energy requirements for
pumping are significantly increased. It is almost unbelievable what a thin, rip-
pling deposit of bacterial slime will do to a pipe's flow characteristics. A very
thin, rippling, slime layer averaging no more than about 1/32 inch in thick-
ness can reduce the capacity of a fairly large pipe as much as 50% without
appreciably changing its internal diameter. The slime growth is microbial in
nature and has high adsorptive capacity. It often contains considerable iron
and manganese. The slime forming organisms can live and thrive on trace
amounts of organics and minerals. The film builds insidiously over a period of
years by extracting nutrients from the water and excreting waste products into
it.
Consider a 12-inch water pipe, connected to a pumping station. The pipe,
when installed,,has. a Hazen-Williams "C" value of 135. If, in five years, this
"C" value is reduced to, say, 80, the capacity loss for the same pressure drop
is 100(135-80)7135 — 41%. To maintain the original flow in the pipe, the
pressure at the pumping station must be increased to the extent of! [135/80]1-85
= 2.6 times its initial value or an increase of 160%. Considering that energy
and energy costs, associated with pumping, vary directly as the pressure
against which the pumps are working, it is evident that bacterial slime growths
in pipelines are an extremely serious matter. These slime, or so-called
"nuisance", organisms may or may not have health significance. (I doubt that
we can answer this question entirely satisfactorily.) On the other hand, they
do have practical operating and economic significance. The alternative to
higher pumping pressures is to clean the line and restore the "C" value. This
is expensive and difficult, and, obviously, the pipe must be out ^ of service
during cleaning. Furthermore, cleaning is, at best, a temporary remedy. Un-
less the organisms are controlled somehow, the whole process starts over the
day the clean pipe is put back into service. Pipe slime brings the water-energy
relationship into sharp focus.
Another objectionable "occurrence that can-result from pipe slime is a
sloughing off of quantities of the slime. This can produce an offensive and
probably unwholesome condition at what the U.S. Public Health Service, and
now the EPA, call the "free flowing outlet of the consumer".
The distribution system is indeed a labyrinth, replete with traps for the
unwary. It of ten illustrates a well kifown corollary of Dr. Murphy's law: "Left
to themselves, things always go from bad to worse".
133
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
The Anatomy of a Water-Borae Epidemic
Let's look at what can happen in distribution. This ''happening", or one like
it, might well form the topic of another symposium. My treatment of it will
be necessarily brief. This city, located in the western Midwest, had a population
of about 11,000 and was an important railroad center. (The latter was im-
portant in at least two respects.) The water supply was obtained from eight
wells having a depth of somewhat more than 100 feet. They were located
several miles from the city and tapped a major aquifer known to be of good
mineral and biological quality. The well water was pumped to a concrete
storage reservoir, covered, as I recall, from which it was repumped through
about 8 miles of 12 inch and 14 inch transmission pipes to the city's water
distribution system. Early in September, the state health department was
notified by a practicing physician that a severe outbreak of intestinal disorder
was occurring in the city. Samples of water were collected and coliform
tests performed. Samples from the wells were "clear". Samples from the
distribution system all contained fairly large numbers of coliform bacteria.
The water system did not include chlorination, and emergency chlorination
equipment was installed. "Boil-the-water" advice was distributed. The rail-
road, a major line, was ordered to cease taking on city water. World War II
was in progress and war industries and members of the armed forces traveling
on troop trains were involved, as well as the city population. Corrective ef-
forts were top late to avert an epidemic and there were more than 3,000
,'U JM1:'1 , 'HI. I
HYDRANT
SEWAGE OVERFLOWED
i. *- VALVE HANDLE
GROUND LINE ELEV.,100 5
VALVE
Figure 1. CROSS- CONNECTION SCHEMATIC DRAWING.
(American Journal of Public Health, 1944)
134
-------
HYPOCHLORITE
SOLUTION
CALIBRATED PUMP
R HYPOCHLORINATOR
CHLORINE RESIDUAL
TESTING
FLOW MEASUREMENT
INJECTION TUBE AND
CORPORATION COOK VALVE CLOSED
TYPICAL SET UP FOR WATER DISINFECTION WITH A HYPOCHLORINATOR.
Figure 2. DISINFECTION OF NEW OR REPAIRED WATER MAINS (Office of
Civilian Defense, 1943}
W
§
P3
sa
O
8
O
w
CO
r«
b
in
-------
i ..... it'
I"
.ii;:11*';
III
I,:. illU! . i . , H
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
RAW WATER FROM CLINTON
REISEIRVOIR & PUMPING STATION
CARBON FACILITIES
FUTURE FLOW SPLITTER
. T'
. fcOAGULANT
NPOLYMER
/PERMANGANATE
^ (__ I
LJ
FUTURE
RAPID MIX
NOJ
/FUTURE
/ PRESED-V
1
WASH
WATER
RECOVERY
I ilMENTATION!
i \ BASIN J
I Xs. ^**^
iJYPASSJ^j-^' V
f FUTURE*.
PRIMARY)
yBASINy
' ""* FUTURE
RAPID MIX
II;. innr i .|l '!'
Tilitll111 i H . i 'li!.l<
RAPID MIX
NO.
PRESED-
IMENTATION
BASIN
(FLOC.-SED^
BYPASS
PRIMAR
BASIN
POLYMER
CARBON
SODA ASH
COAGULANT
PERMANGANATE
.RAPID MIX
NO. 2
ECONDARY
BASIN
(FLOC.-SED.1
SECONDARY
. BASIN
FLUORIDE
POLYPHOSPHATE
CHLORINE
SLUDGE
CONTROL
BUILDING
POLYPHOSPHATE
TRANSFER,
PUMP
WEST HILLS
SERVICE
CLEARWELL
CENTRAL
SERVICE
•FUTURE
RESERVOIR
t'HMi:
Figure 3. LAWRENCE, KANSAS - CLINTON WATER TREATJyiENT PLANT FLOW
DIAGRAM AND POINTS OF CHEMICAL APPLICATION.
cases of dysentery. Public health laboratory tests showed that the predominant
organism was tShigelia paradysenteriae. How did this organism enter the
syslem? Figure 1 gives essential details. A 14 inch supply main was out of
service for repairs. Pressure was "off" for 12 hours or more. During this
period there was a sewer stoppage in a small railroad village where railroad
maintenance employees were quartered. The sewage backed up, overflowed
frost-proof nydraritpits and flowed backward into the 14 inch transmission
main. The head available to cause this flow of sewage into the empty trans-
mission main was at least six feet. How much sewage entered the water main
is unknown, but it was enough. When the transmission "main was restored to
service, it was given only a minimum flush-out and was not disinfected. Re-
calling that this was a transmission main, when service was restored, what-
136
;
' ..... i tri
1 111 lililiiiij J Is ;;, ,^
ill ilill
^ ii JLAiiiL'i-&. !,'..!• I:
-------
ENGINEERING CONTROL PRACTICES/P.D. HANEY
ever sewage was stored in the main, probably quite a bit, was distributed
throughout the city. Here is a thing that should not have happened and the
probability of its occurrence was very low. Nevertheless, it did happen. Here
is a distribution system disaster that chlorination of the supply would not have
averted. One lesson derived from this experience is that pipes Out of service
for repair must be flushed thoroughly and disinfected before1 they can be
assumed to furnish safe water. How often is this procedure followed? Often?
Seldom? Not at all? The answer is, I believe, "often" but not "always". Re-
member, a water main out of service is a serious matter and there is great
pressure to restore service for public fire safety as well as convenience. It is
the duty of responsible professionals to guide repair crews as to the correct
procedure. Disinfection of a new or repaired main is no mean feat. It often
requires mechanical ingenuity and technical competence of a high order
(Figure 2). Another lesson in "preventive medicine" is that if "frost-proof"
facilities are installed, they should not fee drained to a sewer.
Referring again to Murphy's Law, this disaster seems to be an illustration
of the important corollary: "If there is a possibility of things going wrong, the
one that will go wrong is the one that will do the most damage". The distri-
bution system often seems to be a living, thinking thing, well aware of all of
Murphy's laws and how to apply them. ;
If you are not familiar with water works practice and want tp know more
about it, I recommend to you an old publication, produced by :the Office of
Civilian Defense during World War II (1943)—OCD Publication 2022,
titled "Waterworks Engineering in Disaster";
The Lawrence, Kansas Clinton Reservoir Water Treatment Plant
Figure 3 illustrates this water treatment plant, recently designed by the
author's firm. It will have a capacity of 10 million gallons daily and includes
softening, as well as other standard treatment systems.
The raw water will be obtained from CJinton Reservoir, a 7,000 acre lake,
now being constructed by the Corps of Engineers, four miles southwest of
Lawrence, Kansas. This is a multi-purpose reservoir. Its functions include:
flood control, water quality control, recreation, public water supply, and bio-
support. The dam will control 367 square miles of the watershed of the
Wakarusa River. Storage allocations are as follows:
Flood Control 267,800 acre-feet
Multipurpose 129,400
Sediment 28,500 ~
Water Supply :i 10,400 (36 billion gallons)
Land use in the watershed is principally crop land and pasture. Two small,
sewered communities are located about 15 miles above the reservoir. Their
combined population is about 2500. Both have sewage treatment plants.
There are also a few rural mobile home parks. There are no significant agrir
cultural or industrial waste point sources. Estimated mineral characteristics
of the impounded water are as follows:
137
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
Long-Term Mean Values
Cations Anions
mg/1 meq/1 rng/1 meq/1
Ca«* 88- 4.39 HC03- 280 4.59
Mg** 14 1.15 S042- 69 1.44
Na+ 14 0.61
K+ 3.4 0.09 Cl- 10 0.28
6.24 | 6.31
Total hardness is expected to range from 100 to 382 mg/1, carbonate hard-
ness, 100 to 332 mg/1. Coliform and fecal coliform densities will vary widely
with runoff. Coliform MPN values in the range of 200,000 to 300,000/100
ml are likely during peak inflow periods. The estimated long-term fecal-
coliform, goemetric-mean density is expected to be about 50/100 ml. Profe-
ably the most difficult water treatment conditions will be those associated
with algal blooms and related taste-and-odor, coagulation and filtration prob-
lems. Taste-and-odor control will, as usual, be the most difficult and the
most unpredictable.
Note from the schematic that free residual chlorination will be employed,
with provision for alternative chlorine feed points. The water will be am-
moniated after filtration and the free chlorine converted to chloramine.
" Detention time in the various basins at 10 mgd will be as follows:
Presedimentation
II I 111 , II 'i- . ' in, •
Primary *
Secondary
* Based on 20% "split" treatment bypass
This is a symposium on microbiology and is, therefore, concerned with all
those things that contribute to rendering the water "safe". I have no feeling
of uncertainty about the bacterial quality of the water this plant will deliver
to public service. We should have no difficulty meeting the coliform "MCLs .
In fact, I would be surprised if we did not do much better than, the drinking
water standards require. For viruses, we have no standards, but I believe
that if we did have, the plant effluent would readily meet them. The plant
will employ excess lime treatment and high lime treatment has been shown
to be quite effective in virus reduction. This may be due to the combination
of high pH plus the excellent coagulation achieved by lime softening of a
fairly hard water. In addition, we have free-residual chlorination and con-
siderable contact time. We also have, as back-up, dual media filters. If there
are any viruses present, we should remove them, destroy them, or at least
inactivate them to the point that they are incapable of producing disease.
While I am no authority on viruses, my judgment, based on experience with
water plants and literature reports, tells me this plant will produce "safe"
water. 'The water should also be esthetically pleasing and soft enough for
general domestic use. The turbidity standard also can be met without diffi-
culty. Continuous turbidity monitoring equipment has been provided.
•••. :••' • •; •••••"' . ,;• . • '';•,'.,. : li1" • M:I-| !•. ••'• :,;!'' - • •••" :i i
Impact o£ the Standards on Engineering Practice
Turbidity, a light scattering phenomenon, is closely related to micro-
biological quality. When turbidity is measured, we are indirectly measuring
138
-------
ENGINEERING CONTROL PRACTICES/P.D. HANEY
participates in the water and particulatess are important. Turbidity has long
been viewed as a sign of potential danger. Certainly, turbid water, whatever
the cause, is unappealing, if not actually repulsive. Good turbidity removal
at a treatment plant is reassuring insofar as bacterial and virus hazards are
-concerned. Turbidity is appropriately classed as a health hazard and while
not all turbidity is related to microbiological quality, often it is closely related.
Turbidity in finished water may interfere with chlorine disinfection, prevent
maintenance of a residual during distribution, interfere with bacteriological
testing, and be a sign of haphazard plant operation and consequent danger of
water-borne disease.
I do not believe there is anything in the current EPA microbiological drink-
ing water standards that changed the design of the Lawrence, Kansas plant*
If, in the future, a virus standard is set, monitoring for viruses will be re-
quired, and this will call for additional effort by laboratory personnel. There
is, however, looming on the horizon the matter of microchemical effects of
chlorination. Organic chemists have long known of chlorine's affinity for
organics and its ability to enter organic molecules, forming a variety of chloro-
substitution products. Only recently have health authorities and the water
industry begun to delve into these rather subtle, attenuated chlorine-organic
reactions in water. We had some experience in the past with chlorophenols
but this was considered principally a taste-odor problem without health sig-
nificance. Regrettably, we didn't look farther, but the analytical equipment
and the competence to run it have not been generally available for very many
years.
We are now confronted with the real possibility of a standard for halo-
genated compounds in water, such as: carbon tetrachlofide, : chloroform,
dichloromethane, bromoform, dibromomethane, and probably others. This
standard doubtlessly will set limits in terms of micrograms rather than milli-
grams per liter, or-possibly micromoles per liter. If this standard cannot be
met by conventional chlorination practice, then modifications will be re-
quired; however, in making these modifications, we must be ever mindful of
their effects on microbiological quality of the finished water. The new micro-
chemical standards involving halogenated organics are closely related to the
microbiological standards because they may influence, arid possibly drastically
alter, chlorination practice. Engineers ami the water industry have had a lot
of experience with chlorine and we are not going to cast it aside without
a great deal of careful consideration. It is in this area that engineering prac-
tice and drinking water standards are likely to collide head-on. Remember, I
am speculating about a probable standard of unknown magnitude, not an
existing standard. If we do, in fact, have to curtail chlorination somehow, or
use untried substitutes, there will be great professional concern about micro-
biological quality. None of us want to solve a microchemical water problem
and at the same time create a microbiological hazard. Surely, this would be
the negation of sound professional practice.
What can be done? Are there chlorine substitutes that avoid the micro-
chemical problem? There apparently are and they are being studied. Examples
are ozone and chlorine dioxide. Both are powerful disinfectants that insofar
as we know avoid the possibly harmful organic reactions that result when
we use chlorine. Both of these substitutes have engineering and possibly other
disadvantages. First, although, we have had some experience with chlorine
dioxide, we have not had much, and even less with ozone. The latter has
been used for many years in Europe and elsewhere, but there has been little
real engineering experience with its generation and application in U.S. water
works practice. Other potential problems with ozone are lack of a residual
139
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
I •• '
and the inability to store reserve supplies. Ozone must be generated at the
rate used. Furthermore, we know very little about the reactions of either
ozone or chlorine dioxide with trace substances, notably organics, in water.
They are both powerful oxidizing agents, probably fully capable of reacting
with many trace organics. What are the reaction products? Are they innocuous
or harmful? I don't believe we can answer this question today, and the
answers may be several years away.
What other means are available for dealing with the problems that arise
from present chlorination practice? Remove the organics. This can be done
by granular carbon filtration but only at an almost staggering cost and only
after application of considerable engineering effort. While we know a good
deal about carbon filtration, there is probably as much to be learned about
practical day-to-day, reliable operation, as has been learned.
We can reduce the reactivity of the chlorine added to water by converting
it in the water, to chloramine. Here, again, we encounter a penalty. Reduced
reactivity means reduced disinfecting power. This means higher residuals and
a requirement for much longer reaction time. The problem of higher residuals
would be easyto handle, but providing the necessary long reaction time can
be very difficult. Also, chloramines are not spectacularly effective for de-
struction or inactivation of viruses.
j
Summary
The current EPA microbiological standards are essentially those of past
years. They are proven standards that can readily be met by sound engineer-
ing design and good operating practice. The standards call for considerable
paperwork and record keeping, to "which many small utilities are unaccus-
tomed. The required paperwork may be a nuisance but it is not all bad.
It is dismaying and alarming to learn by experience how many small water
utilities, and some not so small, keep poor or practically no water use, quality,
and equipment records and have no up-to-date maps of their water systems.
The matter of public notification has good and bad features and, doubtlessly,
there will be considerable controversy about this requirement.
Present bacteriological standards have not had a major impact on engineer-
ing design. With the aid of coagulation, sedimentation, filtration, and chlorina-
tion, we can meet or exceed these standards without great difficulty. We can
probably also meet any reasonable virus standards proposed. Achieving the
turbidity standard should not be difficult, except for some utilities using un-
filtered surface supplies. For those, the technology, and engineering expertise
to apply it, are available. The problem is, therefore, one of management,
finance, and public support. The job can be done, but consumers must ex-
pect to pay more for water than in the past. One of the problems in the water
industry is that water, a truly precious substance, has always, hi the past,
been plentiful and cheap—probably too cheap. This is changing.
Any alteration required in chlorination practice, as, for example, curtail-
ment or elimination of prechlorination, will have a major impact on engineer-
ing design. We are ready, willing, and able to apply proven new systems, as
required, but we can't do the whole job alone. I am pleased that the scientific
community is, at long last, showing much more interest in water. Engineers
must look to the scientists to define basic water quality needs and we invite
them to work with us in charting a true course. Scientists need the help of
engineers and engineers certainly need the help of scientists. Working alone,
neither group can be completely effective, but working together, there is very
il 4 ! 'little we can't do.
140
-------
ENGINEERING CONTROL PRACTICES/P.D. HANEY
REFERENCES
1. Berger, Bernard B. et al. 1970, Engineering evaluation of virus hazard in
water. J. Sanit. Eng. Div., Proc. Am. Soc. Civ. Eng. 96:111-161.
2. Board of Directors, AWWA. Quality goals for potable water. JAWWA
€0:1317-1322.
3. Chang, Shih Lu, G. Berg, N.A. Clarke:, and P.W. Kabler. 1960. Survival and
protection against chlorination of human enteric pathogens in free-living
nematodes isolated from water supplies. Am. J. Trop. Med. Hyg. 9:136-142.
4. Kehr, Robert W., and C.T. Butterfield. 1943. Notes on the relation between
coliforms and enteric pathogens. Public Health Reports 55:589-607.
5. Kinnaman, C.H., and F.C. Beelman. 1944. An epidemic of 3000 cases of
bacillary dysentery involving a war industry, and members of the armed
forces. Am. J. Public Health 34:948-954.
6. Stoltenberg, G.A. 1970. Water quality in an agricultural watershed. In:
Transactions of the 20th Annual Conference on Sanitary Engineering, Bulle-
tin No. 62. School of Engineering and Architecture, Lawrence, Kansas.
7. The Committee on Water Quality Criteria. 1972. Water quality criteria—
1972, Environmental Studies Board, National Academy of Sciences, National
Academy of Engineering, Washington, D.C.
8. United States Pubjic Health Service. 1962. Public health service drinking
water standards. Public Health Service Publication No. 956. U.S. Department
of Health, Education and Welfare, Washington, D.C.
9. Water Research Centre. 1976. Deterioration of bacteriological quality of
water during distribution. In: Notes on water research. Water Research
Centre, London.
141
-------
1 ' ii
-------
Water Treatment Plant, Jackson, Mississippi
SESSION 3
Compliance with the Coliform Standard
Chairman: Dr. Charles W. Hendricks, Criteria & Standards Division, Office
of Drinking Water, U.S. Environmental I'rotection Agency, Washington, D.C.
143
-------
,
'if
-------
DRINKING WATER REGULATIONS/C.W, HENDRICKS
Compliance with Drinking Water Regulations
Charles W. Ilendricks
Office of Drinking Water
U.S. Environmental Protection Agency
Washington, D.C. 20460 '.
In 1970, the Department of Health, Education and Welfare undertook a
study of public water supply systems (4). The results of this study indicated:
1. 36% of the tap samples contained one or more bacteriological or chemi-
cal constituents exceeding the limits in the 1962 Public Health Service
Drinking Water Standards.
2. 85% of the systems did not analyze a sufficient number of bacteriologi-
cal samples.
3. 77% of the plant operators were inadequately trained in fundamental
water microbiology.
4. 46% of the plant operators were deficient in knowledge of the chemistry
relating to their plant operation.
5. 79% of the systems were not inspected by State or county authorities
in the year prior to the study.
For these and other compelling reasons, Congress enacted the Safe Drink-
ing Water Act (2) (PL 93-523) in 1974, "to assure that the water supplied
to the public is safe to drink." The basic structure of the Act, as it pertains
to public water systems, calls for .the establishment of: maximum contaminant
levels for most contaminants in drinking water; and criteria and procedures,
including quality control measures, to assure compliance with such maximum
contaminant levels.
Provisions. of the Safe Drinking Water Act and The Interim Primary
Drinking Water Regulations
The Safe Drinking Water Act applies to each "Public Water System,"
defined in Section 1401 (4) of the Act as "a system for the provision to the
public of piped water for human consumption, if such system has at least
fifteen service connections or regularly serves at least twenty-five individuals."
Both privately and publicly owned water systems are covered, and the U.S.
Environmental Protection Agency has interpreted service "to the public" to
include factories, churches, schools, and private housing developments (3).
The definition of a "public water system" in the National Interim Primary
Regulations (3) seeks to explain the meaning of the statutory reference to
"regular" service. It was originally proposed to interpret the term to include
service for three months out of the year, but this definition would have ex-
cluded many public accommodations such as campgrounds, lodges and sea-
sonal recreation areas which serve large numbers of tourists but are in
operation for perhaps less than three months each year. The definition in
the final version covers systems serving at least twenty-five individuals at least
60 days out of the year.
Two types of "public water systems" are recognized in the Interim Pri-
mary Regulations—those serving community residents and those serving
transients for intermittent users. This differentiation is based on whether the
water served is an individual's major source of drinking water. Although
this concept has more applicability for chemical than microbiological con-
taminants, it is, never the less, based upon the fact that the possible health
effects of a contaminant in drinking water in many cases are quite different
145
-------
Ill1 11111 1 '1 III 1' ii '•: " " Ml '1
1 111 1 1 ; i 1 i
in Hill i ' i ','.••:'• • 'i1 iii
1 i iiiih i 1 ' i
'1 I
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
•
for a person drinking the water for a long period of time than for a person
• drinking the water only briefly or intermittently. For microbiological pur-
poses, however, the risk of infectious disease is dependent on whether the
organism is present in the water, and little distinction is made in the Interim
Primary Regulations for the microbiological monitoring of very small com-
munity water systems serving 1000 or -less residents as compared to the non-
community supplies which serve the more transient users.
The Maximum Contaminant Levels (MCL), for coliform bacteria con-
^i'/'tamed in the Interim Primary Regulations (Table 1) are based upon the
analytical technique used. A laboratory may employ the membrane filter
procedure or either the ten or one-hundred milliliter fermentation tube proce-
dure. Any of these procedures may be employed at the discretion of the
State, but only one procedure can be used during any compliance period
since the limits prescribed for the fermentation tube methods are based upon
tubes containing gas and not values taken from the "most probable number"
table.
Limits for coliform bacteria, whether the sample has been analyzed by
the membrane filter oir the fermentation tube technique, are expressed in two
parts. The first part of the MCL establishes the arithmetic mean concentra-
tion of coliform bacteria that can be allowed in water from the system per
month. In essence this number represents a limit for the general water quality
of the system as a whole! Tfie• s'econH portion of the MCL seeks to place a
i ; i II II ill! ^ i iII j || i i i ,, . •! " • n|j|r.
TABLE 1. MAXIMUM MICROBIOLOGICAL CONTAMINANT
" . '" LEVELS*
Jilllli 1| ill'Mi, Illil, ' '
UK „ i'illil
A. A mean of 1 coliform per 100 ml of water per month.
B. Four coliforms per 100 ml of water in more than one sample when less
than, 20 samples are examined per month; or
Four coUforms per 100 ml of water in more than five percent of'the
samples when 20 or more samples are examined per month.
FERMENTATION TUBE TECHNIQUE |
10 ML Procedure
A. No coliform bacteria in more than 10 percent of the portions examined
per month.
B. No coliform bacteria in three or more portions in more than one
sample when less than 20 samples are examined per month; or
No coliform bacteria in three or more portions in more than five per-
cent of the samples when 20 or more samples are examined per month.
100 ML Procedure
A. No coliform bacteria in 60 percent of the portions examined per month.
B. No coliform bacteria in five portions in more than one sample
:y?hen less than five samples are examined per month; or
No coliform In five portions in more than 20 percent of the samples
when five or more samples are examined per month.
* National Interim Primary Drinking Water Regulations (40 CFR 141.14)
146" ' ' "• ' ' " '- ' ' '
-------
DRINKING WATER REGULATIONS/C.W. HENDRICKS
limit on the number of coliform bacteria that would be acceptable in any
individual sample and represents a limit for water quality at a specific
sampling point within the distribution system. Failure of any portion of the
limit is sufficient cause to fail the MCL.
Table 2 is a summary of the major sampling and analytical requirements
for coliform bacteria that are contained in the Interim Primary Regulations.
These provisions instruct the utility in how many samples to take, the analyti-
cal procedure to use and under what conditions should follow-up (check
samples) be taken. Also included is a provision for the substitution of chlorine
residual determinations for some of the required coliform determinations
should the States desire to employ it. ',
The sampling and analytical requirements of the Regulations determine
the extent to which compliance with the maximum contaminant levels will
be maintained. It has been and continues to be the Office of Water Supply
policy to keep the Primary Regulations flexible and to allow for State discre-
tion where possible. For this reason the analytical procedures other than
those cited in Standard Methods may be employed if it can be shown that
the alternative analytical technique is substantially equivalent to the prescribed
test in precision and accuracy as it relates to the determination of compliance
with the MCL. Likewise, some monitoring requirements may be reduced by
the State for small community and non-community systems if local condi-
tions warrant the reduction.
Problem Areas in the Interim Primary Regulations
The Maximum Contaminant Levels for coliform bacteria and their sam-
pling and analytical provisions as contained in the Interim Primary Regula-
TABLE 2. MICROBIOLOGICAL CONTAMINANT SAMPLING AND
ANALYTICAL REQUIREMENTS.*
MAJOR PROVISIONS
A. Analytical procedures are from Standard Methods, APHA, 13th
Edition.
B. Sampling frequency is based on population and is specified.
C. Sampling frequency for small community systems serving persons 25-
1000 may.be reduced to 1 per quarter at State discretion.
D. Sampling frequency for non-community systems of 1 per quarter may
be reduced on the basis of a sanitary survey.
E. Two consecutive, daily check samples are to be taken if:
—The MF count exceeds four coliforms per 100 ml of water.
—-Coliforms occur in three or more .10 ml portions.
—Coliforms occur in five 100 ml portions.
F. Chlorine residual analyses may be substituted for up to 75 percent of
the coliform determinations at a rate of four chlorine residual determi-
nations taken for each coliform test substituted. At least one chlorine
residual test must be taken per day. _
G. Public notification is required if a MCL or a monitoring frequency is
violated.
* National Interim Primary Drinking Water Regulations (40 CFR 141.21)
147
-------
I!1!'' I I'!1' "Bit!1! .'• .aillllt; ' ' nUIIIi ""•!• "'"' i .,- ''In'1!- •,' ';'!'; .1? H't'iBR PI lib I'll! SI! I rv'ffl ; IP1;' t''»!iif'5i'l!iri« |i|t".ii!|i!''»l™!i|'!"!|" .,'Kfjyw;!' i ':'l
i;,
!.
liiiiii " i •'! '•! " "
wr : •;.,! • ' :
!' jilt
i \
MICROBIOL STANDARDS EVALUATlON/C.. HENDRICKS
11 ,11' , , i,', '.. !!! .• "! ' ' " i i" ;" .1. •' •' i ,„ , • , • ij i' :• ,31 "I , , ,.''' .,•• II :'!!
tions are certainly not the last wdrd in developing drinking water regulations.
Fqr example, the promulgated MCL's for coliform bacteria are basically the
1962 Public Health Service Standards (5) with, minor refinements and clari-
fications. However, further changes may be desirable before Revised Regula-
tions are developed. For example
cal methods do not resolve the r
assumed to be present in a singl
tories assume an upper limit of
individual bacteria to a level of
upper limit assumed will affect
determine compliance with the
, the MCL's for the membrane filter analyti-
uestion of how many coliform bacteria are
; highly contaminated sample. Some labora-
50, while others seek to continue to count
100 or even higher in a single sample. The
he monthly average which is calculated to
MCL's. Another question relating to the
As the regulations are written,
coliform bacteria MCL's is the ^natter of possible spurious positive samples.
all routine samples must be counted even
though subsequent check samples may not indicate contamination. The
reason for this is that bacterial contamination is often intermittent or transient.
Negative follow-up samples taken a day or more afer a positive routine
sample has been analyzed cannot demonstrate that the positive result was in
error. It may be possible, however, to prescribe a means of dealing with
spurious positive results without compromising the integrity of MCL's.
A third question concerning the MCl's for coliform. bacteria is the rela-
tionship of monthly averages J>f coliform bacteria levels to monthly per-
centages of positive samples. For example, the monthly average MCL for
the membrane filter Method is violated if the monthly average exceeds one
coliform bacterium per sample. However, for purposes of determining whether
the month-percentage-of-positivejsamples MCL is violated, a sample is counted
as positive only if it contains nkore than four coliform bacteria. Thus, it is
possible, particularly when a relatively small number of samples is taken, for
a system to fail the monthly Average MCL even when no single sample
taken during the month is out i»f compliance with the limit
In addition to the questions listed above, this Office has been asked to con-
sider an amendment to the Interim Primary Regulations that would specifically
address the case where one positive sample would cause the supply to violate
the MCL. These requests cite th 2 following.
1. For small utilities (perhaps those less than 10,000 persons served
although the cited examples are often those serving 2,400 or less),
the collection of one sample requiring the collection of check samples
may cause a supply to :;ail to mee.t the MCL for that compliance
period, thus triggering tb.4 public notification requirements.
2. The collection of one such sample followed by two negative check
samples should not be counted a failure of the MCL because it could
have resulted from improper sampling technique, from accidental
contamination in the laboratory, or the lack of precision in the culturing
procedures. ' •
3. Public notification by way of the news media would serve no purpose
because it would be too long after the fact for consumers to take self-
protective actions (even if such action had been justified at the time
of collection of the sample in question). The notice may also unneces-
sarily upset and alarm consumers, and would ultimately be self-defeating
as consumers might develop a "no-hum" attitude when it became
apparent no waterborne
tamination notices.
Certainly these comments are
is not addressed in the Interim
148
disease transmission accompanied the con-
significant because they point to an area that
Primary Regulations. I would hope that dis-
t • w
-------
DRINKING WATER REGULATIONS/C.W. HENDRICKS
cussion at this symposium would focus on
results especially for small systems taking 1
what to do about false positive
!ss than ten samples per month.
In addition to this question, the following also warrant discussion and
comment.
1. Validity of the Coliform Indicator Syjstem
—Are we using the proper indicator system?
—Do the present limits protect health?
—Is it possible to calculate risk from these limits?
-^-Can non-biological tests be sulbstitutjed? ;
2. Sampling Frequency and the Arithmet c Mean |
—Is the calculation of a mean appro] >riate for small systems?
—How many samples are enough?
—How firm is the technical basis; for t le present monitoring frequency?
3. Compliance With the Coliform Bacteia MCL
—Is the two part limit practicable?
—Should follow-up (check) samples be used in calculation of the mean?
—What is the significance of two negative follow-up samples?
—Is the MCL properly established on I a technical basis?
—Can the limit be met for all systems?
In conclusion, the Interim Primary Regulations for microbiology are
basically the same as those contained in I he 1962 Public Health Service
Standards with certain minor clarifications. These Regulations are adequate,
in my opinion, to protect the public health providing the water comes from a
protected, uncontaminated source. There ar; certain desirable modifications
that can be made by amendment, others, hotoever, must await further study.
Hopefully this symposium will provide guidance to the Agency on how these
modifications should be made.
REFERENCES
1. House of Representatives. 1974. Safes Driri
No. 93-1185. 93rd Congress, 2nd session,
2. The safe drinking water act, title XIV-Sa:
ding Water Act, pp. 16-17. Report
iVashington, D.C.
rety of public water systems, (PL
States, Washington, D.C.
Agency. 1975. National
interim
93-523). 1974. Congress of the United
3. United States Environmental Protection
primary drinking water regulations. Fed. Register 40:59566-59588.
4. United States Public Health Service. 1970.1 Community water supply study-
significance of national findings. United States Department of Health, Educa-
tion, and -Welfare, Washington, D.C.
5. United States Public Health Service. 1962.
States Department of Health, Education, and Welfare, Washington, D.C.
QUESTION AND ANSWER SESSION
W.B. Jeffcoat, Alabama Department of Public Health, Montgomery, Ala-
bama.
Your .statement...!!! session^ 3, concerning
the regulation is quite true: We in Alabama have relayed our concern to
Region IV on the same matters. You onlj
however. The numerical value of '"TNTC"
mentioned. You stated that help had been
Drinking water standards. United
the three problem areas with
addressed one of these areas,
and the dual MCL's were not
sought to solve these problems.
Help from where? EPA? States? Universities? Water Systems? Did the help
149
-------
IMF! BISEilB1!"
HIM '• iipj'ii ..... p, j"s "" > ....... j ...... if ' .: ......... "P • ..... ; 'K" ; :T! ..... ..... t ....... ] ...... 'Wi ,s ..... 7i:i«K w ..... ' ? :"i ...... in ..... ' f >. ' ...... • m f'-
t ...... iiRiisf iiiisif am>" iiiifE ...... «' :VKV ..... f .• ...... •'• ..... i ....... :^
• • w: ?! -SK
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
come from people who are familiar with the operation of a water supply
system or. did it come from EPA officials and University personnel whose
only contact with a water system is an occasional visit.
I'm sure you realize that federal agencies especially EPA have low levels
of credibility. This is something we wish to avoid on the State level. Forcing
a water supply to comply with unrealistic requirements will place the State
in the same category of EPA. The goal of the SDWA is "safe drinking water
for all" and to accomplish this goal cooperation is needed between the
State and water system: There is more to operating a water system than
merely complying with the standards. I agree with Mr. Taylor and Mr.
Harrison that the mean should not be used in determining public notifica-
tions. Check samples should also be counted as they do with other parame-
ters hi the National Interim Primary Drinking Water Regulations. My main
question is will EPA consider the proposals of the utilities and States or
will EPA rely on information provided by those unfamiliar with the actual,
daily operation of public water systems and with the actual, daily operation
of a state public water supply supervision program?
C.W. Hendricks, Office of Drinking Water, ILS. Environmental Protection
Agency, Washington, D.C.
This meeting was designed to provide the U.S. Environmental Protection
Agency with the best information that experts in Water Supply from EPA,
the States, Universities and water utilities can provide. I do not believe that
any one person, or profession for that matter, has all of the information at
hand needed to correct the deficiencies that we see in the colifonn maxi-
mum contaminant levels, I would anticipate that any amendments to the
National Interim Primary Regulations would be thoroughly reviewed by
State and utility personnel before they are promulgated.
i
G.W. Fuhs, New York Department of Health, Albany, New York.
How did you determine there was no danger connected with coliform
'"'si standard?
!|i,,!l!lll! ' '' ' '. '"''I' ill.'1. I1 ". II, ,' ,, , i.11" J . fill '» .ill1 I I 'I i1 I'P ' I ':
C.W. Hendricks.
I would refer you to Mr. Floyd Taylor's paper elsewhere in this manual for
a description of how our present colifonn standard originated. I suspect
that the standards were empirically derived without any study to determine
what the risk to health might be. I might add that experience has shown us
that the present coliform standards adequately protect the public health pro-
viding a high quality uncontaminated source water is used. Any substantive
changes m the MCL will require careful evaluation before they are pro-
mulgated.
Comment by E.E. Geldreich, Water Supply Research Laboratory/MERL,
U.S. Environmental Protection Agency, Cincinnati, Ohio.
If I may, I would like to respond to this presentation by Dr. Hendricks in
a series of short comments on the many aspects of the subject.
a. Two Classes of Water Quality
The proposed amendments in the bacteriological regulations are drifting
into a direction of establishing two classes of public water quality in the
Nation with a disregard for identifying poor quality waters in smaller sys-
tems. While this may ease the burden of small water supply surveillance on
the State agencies, the proposed changes will weaken the goal of achieving
"! 150 ' '" " ':''' ' ' !":" '': "' ' "' ' "''"
-------
DRINKING WATER REGULATIONS/C.W. HENDRICKS
uniform, high quality Water through minimal regulations for all public water
supply systems. Furthermore, these relaxed regulations would be applied to
the very water systems that our 1969 national community water supply
survey indicated were most subject to producing marginal quality potable
water.
b. Variance in MF-MPN Test Application
As the regulations now stand there is a variance in the application of
rules related to the membrane filter versus the multiple tube test and an
uneven interpretation of bacteriological results when using either of these
two alternative bacteriological procedures.
Since either of these procedures (MF or MPN) may be used in the ex-
amination of potable water, all results should be reporting MF results in
terms of coliforms per 100 ml. The: present practice of reporting MF re-
sults in terms of coliforms per 100 ml while restricting MPN reports to only
the number of positive tubes, results in some inequities of interpretation.
Every laboratory that is capable of performing the multiple tube test is also
capable of computing the proper MPN value. The use of a system that re-
ports only the number of positive tubes per test is archaic, and easily subject
to manipulation to meet a monthly minimum and should be discarded.
c. Recognition of Low-Level Contamination Occurrences
The definition of an unsatisfactory sample should be "1 per 100 ml"
for the membrane filter test and "2.2 per 100 ml" for the multiple tube test.
These limits are equivalent since the MPN test value is a statistical estimate
and uses a table of numbers elevated by incorporation of a positive bias.
This revision in the definition of an unsatisfactory sample would direct at-
tention to the occurrence of persistent low level contamination problems
that have been ignored in the past. This change would not alter the intent of
the proposed amendment as it relates to repeat sampling since the term
"negative" is interpreted to mean that no coliforms were detected in the
official test portions by either the MPN or MF procedure. The change
would require mandatory check samples for all detected occurrences of
coliforms in public water supplies.
d. Sample Frequency for Small Supplies
The proposed reduction of sampling frequency for small community
systems serving 25 to 1000 persons to one sample per quarter- and further
reducing for non-community systems of one per quarter on the basis of a
sanitary survey should not be permitted. Since the number of supplies
serving 25 to 100 persons is largely unknown and these supplies are mostly _
comprised of small trailer parks and roadside facilities that serve a transient
population, sampling should only .be reduced to once each quarter during
non-recreational periods of the year, however, increased to once per month
during the normal recreation period.
Since the burden of increased monitoring and corrective action will be
on public water supplies serving populations between 25 and 9,400, guide-
line dates should be established to intensify program direction first to sup-.
plies serving 1,000 to 9,400 people, .then to supplies in the 100 to 1,000
population category, leaving the smallest supplies (25 to 100 population
served) as the last group to be phased into the program.
e. Check Sample Incentives
Check samples should be included in the monthly average of routine
samples for all water supplies, regardless of size. This would; offer incentive
to water plant management to increase their sampling program beyond the
minimum two consecutive daily check samples required following an un-
satisfactory sample report. Because of the urgency to detect contamination
151
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
'in'1'1'! 'I " " „, ' ! ; .. ,,', ..... •.''''. '' '": ' 'i';1 '• • '? 'ij;/1' i ""I"!1,: ' ,:'*•) ''' li '
f" :'
S' >: , , • • , • ' , , ',',' i • ,i .' i " ' ''. , , • i , , • . , i<,; ,.,11 .M,: • , :• ,• >!„
breakthroughs into the potable water, requests for a repeat sampling following
unsatisfactory results should be made i. promptly and the sampling initiated
within .no more than 48 hours. The answer to the perennial concern about
the integrity of the sampling during collection could best be resolved by
stating a requirement that sample collectors must be certified by the ap-
propriate state agency following successful completion of a 1-2 day training
course. Thereafter, recertification could be related to observed performance
of sample collectors during the recertification of the laboratory.
f. Public Notification
The public notification requirements should be restructured to include
two levels of response. The first level of response occurs when one to four
cbliforms are detected by the MF method or 2.2 to 5.1 coliforms are re-
ported in the MPN procedure. In this case, the water system must institute
aft immediate review of treatment practices, including examination and clean-
ing (if necessary) of the distribution system. Daily check samples are manda-
tory during both the sanitary survey and the period of corrective action
until two consecutive satisfactory samples are obtained. A written record
of the survey report, corrective action(s) and bacteriological results are
maintained for five years. The second level of response involves any oc-
currences of 5 or more coliforms per 100 ml on the MF or above 5.1 per
100 ml as measured by the MPN test in one or more of the daily check
samples. In this later situation, public notification would be mandatory, in-
cluding a "boil water order," if necessary, immediate increases in disinfec-
tion to achieve a disinfection residual in. the most remote sections of the
system, hourly measurements of disinfectant residuals and daily bacteriologi-
cal, sampling at these remote sections. As soon as the monitoring data on
the distribution network demonstrates a disinfectant residual for 24 continu-
ous hours, and two consecutive days of satisfactory bacteriological results, a
public notice cancelling the "boil water order" should be issued.
g. Program Approach
In essence, while portions of the regulations need to be changed to re-
solve problems in monitoring and interpretation of laboratory results, these
modifications must not drift toward establishing two classes of public water
quality in the Nation with a disregard for identifying and responding to
poor quality waters in smaller systems. A systematic approach to attacking
the increased program responsibility should be geared to solving current con-
straints in state and local water supply staffing and monitoring costs, not
in minimizing objectives defined by intent of the law.
152
-------
COLIFORM STANEiARD: INTERSTATE CARRIERS/F. TAYLOR
Experiences With the Coliform Standard Under the
Interstate Carrier Program
Floyd Taylor
Chief, Water Supply Branch
U.S. Environmental Protection Agency
Boston, Massachusetts 02203
Introduction
The past often provides the record of experience which is useful in
guiding decisions on present and future courses of action. History is also
littered with the wreckage of men and systems who ignored its lessons and
this of course is what We seek to avoid by the intensive examinations we are
giving this "Evaluation of the Microbiology Standards for Drinking Water."
Our particular part of this examination deals with experience with the coli-
form standard under the interstate carrier certification program conducted
for 56 years by the U.S. Public Health Service and for seven; by the Environ-
mental Protection Agency. I believe it will be helpful to briefly review the
development of the coliform standard as it appeared in successive issues of
•the U.S.P.H.S. Drinking Water Standards.
The Coliform Standard
In 1914 the entire document, commonly referred to as the Treasury
Standards, was entitled "Bacteriological Standard for Drinking Water," under
which Secretary McAdoo informed the Surgeon General that "In the future
common carriers will be required to furnish water for passengers in inter-
state traffic which will conform to this standard." (5) The sections of that
standard pertinent to today's discussion read as follows: "Not more than
one out of five lOcc portions of any sample examined shall show the presence
of organisms of the Bacillus coli group when tested as follows:" and "The
total number of bacteria developing on standard agar plates, incubated 24
hours at 37°C shall not exceed 100 per cubic centimeter." (6) It should be
noted that no mention was made of the number of samples to be taken nor
of the deviations from the above stated requirements which would appear in
later reyisions._The next Treasury Department document was entitled "Drink-
ing Water Standards" and we see the beginning of the present day practice
for it read, in part, as follows:
"II. AS TO BACTERIOLOGICAL QUALITY
"1. Of all the standard (lOcc)-portions examined in accordance with the
procedure specified below, not more than 10 percent shall show the presence
of organisms of the> JB. coli group. .
"2. Occasionally three or more of the five equal (lOcc) portions consti-
tuting a single standard sample may show the presence of E. '• coli. This shall
not be allowable if it occurs in more than
(a) Five percent of the standard samples when 20 or more samples have
- been examined;
(b) One standard sample when less than 20 samples have been examined.
Note—It is to be understood that in the examination of any water supply the
series of samples must conform to both the above requirements 1 and 2. For
example, where the total number of samples is less than six, the occurrence of
153
-------
lIlliKf" f ! ! 1 I!11!"1!:,,
Pi!!' '' , ".
!('»:;• ' i .
1 IIPK 1 i Sfil'i'i!' 'i1'"1! •III!1!!1111"' lillill11'!'!!1!';'!:'- If '• Ti !' ' I1:';
" ' si;!, :;;,, .'-yv f 'H'f fia,,"!'
in;; i I'1!, J ;•,!" Ij
- !? "i !;' "'i " ! i: &<'• 'ijj !• "CM" ' " fjf'i .«•. . • i",,1";- "«' •); !i;i W'i ("iji •!? '!" ' ' Hjf i I'-faf!
: .; ! ."'• ..'..', '' t:- .i*'''' ''!' f-'W
;l . . . .•:,'..' ' t I' J
llliriiliTlLiniljiiliil I'lli'llT ; r .."!:» '.ii;. .IIIUII1 1. ::i :' .1 n I.'.)1;.. II II inn II null
' !!!-.' "'- I;-1' ' Si1'!!:,'!"'' ' ' :i 1 ' 'I" I1 1 1 1
'" . . : '' "'::' '"; ' '" ' " ', "'"";,' ! ''''i"11' !':"' • 'l'"l':!|;!.li
")! I'1;; , I fl '
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
t 111! ! .;! 1
til'
positive tests in three or more of the five portions of any single sample, al-
though it would be permitted under requirement 2, would constitute a failure
to meet requirement 1."
In the 1942 and 1946 document by then called the Public Health Service
Drinking Water Standards — standard portions and standard samples were
defined and the 100 ml standard sample was introduced as were the first
requirements as to the number of samples which should be collected. The
\yell known graph showing the relation between population density and num-
ber of samples to be collected per month appeared in the 1942 standards.
It will be helpful to note the slight changes in the allowed limits as they
have bearing on today's discussion. The requirement was as follows:
"12.203 (b) (1) of all the standard ten milliliter (10 ml) portions examined
per month in accordance with the specified procedure, not more than ten
(10) percent shall show the presence of organisms of the coliform group."
"12.203 (b) (2) occasionally three (3) or more of the five (5) equal ten
milliliter (10 ml) portions constituting a single standard sample may show
the presence of organisms of the coliform group, provided that this shall not
be allowable if it occurs in consecutive samples or in more than
(i) Five (3) percent of the standard samples when twenty (20) or more
samples have been examined per month.
(ii) One (1) standard sample when less than twenty (20) samples have
been examined per month.
Provided further that when three or more of the five equal ten milliliter
(10 ml) portions constituting a single standard sample show the presence
of organisms of the coliform group, daily samples from the same sampling
point shall be collected promptly and examined until the results obtained
from at least two consecutive samples show the water to be of satisfactory
quality." It is not necessary to emphasize the requirements for 100 ml
sampling as this system has not been widely adopted even though it has been
carried over into the EPA Interim Primary Drinking Water Regulations. It
will be realized that the 10' ml multiple tube series for the 1925 and 1942
standards were specifying an average of less than 2.2 coliform organisms per
100 ml with occasional allowances as high as over 16 organisms per 100
ml, but, beginning m 1942, only if check samples (two consecutive) were
negative. In other words a high sample was the signal to take corrective
action.
In the last (1962) PHS Standards the general wording of the 1946 edi-
tion was continued but with the further restriction that "the presence of the
coliform group in three or more 10 ml portions shall not be allowable if
this occurs in two consecutive samples" (2). Also in 1962 the membrane
filter technique was recognized with the requirements not to exceed an average
of one colony per 100 ml (monthly arithmetic) and on occasion not more
than 4/100 ml. And that is how we arrived where we are today with the
bacteriological requirements of the Interim Primary Drinking Water Standards
wnick are essentially those of the 1962 PHS Standards. I would however
point out that the 1914 document and its successors, or rather the applica-
tion of them together with the practice of filtration and disinfection were
responsible for the dramatic reduction of enteric disease in this country
during the early 1900's.
Coverage of the Standards
I
When the first federal drinking water standards were promulgated there
were in 1916, 4077 U.S. public water supplies which came under their juris-
154
-------
COLIFORM STANDARD: INTERSTATE CARRIERS/F. TAYLOR
diction. It is estimated that they then served about 35% of Americans who
utilized public water supplies. In 1975 the number of Interstate Carrier Wa-
ter Supplies was 689 serving 90,000,000 people or about 55% of the popula-
tion on public water supplies. I would also cite the historic policy by which
the standards were applied namely through a cooperative, federal-state pro-
gram. Even more importantly, the standards were adapted by nearly all
the states as applicable to public water supplies generally within their jurisdic-
tions. However, judging from experience in New England, this application
was administratively and not legally enforced. Now a word about the gen-
eral procedures for applying the bacteriological standards.
Application of Standards
Records of the precise methods of applying the first drinking water
standards were difficult to find but Treasury Department Circular No. 48,
October 5, 1912 directed PHS commissioned officers traveling on common
carriers to make observations regarding sanitary conditions and submit ap-
propriate reports (1). More importantly, the states were required to make
reports to .the Service on the quality of the sources and presumably gave
warning when the quality of a supply failed to meet the requirements. Let
me here state for the benefit of those who are not aware of how the Inter-
state Carrier Water System worked, the enforcement phase has always been
an action between the federal government and the common: carrier not be-
tween the federal government and the water purveyor. Furthermore, since
the early days of the system the states made recommendations as to the
type of action which should be taken.
Sometime between 1925 and 1938, when my personal experience with the
system began, a three level method for approving carrier water supplies was
adopted. First there -was - full approval, - second provisional approval and
third, prohibited. The first meant that all was well, the second that there was
some question about the capability of the supply to continuously provide safe
water and prohibited meant what it said. What the classification would be
was somewhat arbitrarily determined by discussions between the state health
engineer and the PHS district engineer.
Beginning in 1942 the concept of numbers of samples per month and
allowable ratios of coliforms were adopted and an annual report form was
introduced which required the State to show for each month what the bac-
teriological record had been. However these reports were submitted only
once a year to PHS and_"bad" results (which by that time could be a year
or more old) were usually explained by a footnote. Invariably the supply was,
at worst, "provisionally approved." With the great shift during this period
of personnel and respurces to water pollution control activities at both state
and federal levels, interest in water supply declined almost proportionately.
By 1959 many state and Federal water supply programs had fallen into a
state of low effectiveness. In that year the Public Health Service water supply
program was placed under new management.
In 1962 Public Health Service Headquarters Water Supply Program ad-
vised all the regional offices that the bacteriological requirements * of the
standards meant what they said and would be enforced. In the ensuing years
the numbers of provisional and prohibited actions increased markedly. It
was also during those years that the Service began to keep track of the wa-
* HEW General Counsel determined that only the bacteriological provisions of
the standards were enforceable.
155
-------
i inijjjii,!!1 ' , r :h;>i'»,, i ill" ,i, if, |
i!'1 iiiiiis i i ',i:' iiS, • \ i'1"1
111,1''. "i
Ik1 ill
(ill
II!!1,
II Illlli
MICROBIOL STANDAR 3S EVALUATION/C.W. HENDRICKS
ter supply improvementsthat resulted from all enforcement actions and
although I cannot quote you exact data I can tell you that the rate of new
water supply construction resulting from the program was of the order of
several millions of d
In 1971 when the
to EPA the first sy
liars per year.
interstate Carrier Water Supply Program had gone over
tematic guide to the Interstate Carrier Water Supply
Certification Program appeared entitled "A Guide to the Interstate Carrier
Water Supply Certification Program (7). It provided valuable criteria for
certifying Interstate Carrier Water Supplies including both bacteriological and
chemical. Again the
and the basic elemen s were as follows:
':'"!, ' " "!'!•!' ''"l; iJit'",""" 'i" ' ": iii"
For Approved:
;, ,|,|i,i ,;,, ,,,,; • ,, , „, ,,,, ,,„„ , „ , ,„ , , ,, „,,,,, l , ,, , , ....
r quality must meet all the limits of Section 3.2 (sections as they
" 2 standards) and must meet the sampling ratios of
appear in the 19
Section 3.1 at leas
For Provisional A
Water quality
12 month reportir
for 2 or more months
iproval:
f iled to meet Section 3.2 requirements for one of any
period and/or monitoring requirements were failed
Use Prohibited:
Water quality :
months of any 1
violated by failure
samples for 3 mori
To get back to a "p
ply must have a recc
bacteriological resultr
I can not speak fo:
Regjon I but I can si
grading of monitorin
tions. In one state w
I Hill :*•• .'Hi.jvwjjs., a i 'i i
the public water supp
equipments
treatment,
'50% did meet them.
I would also like
bacteriological' 'n
hibited actions which
field, Massachusetts
1969 Springfield, a ci
tion as the only
the PHS, standards
adequate watershed j
$125,000 automatic,
teriological record ha
In, the summer of
of its distribution 53
100,000 inhabitants.
refused to change tl
'I'"' Today'Fall" River "is
million most of whid
156
"i
I
ormer are of particular interest to us in this discussion
11 out of each 12 month period.
iled to meet Section 3.2 requirements for 2 or more
month period and/or monitoring requirements were
to obtain less than 50% of the required number of
hs of any 12 month period.
rovisional" from a "prohibited" classification the sup-
rd of at least three consecutive months of satisfactory
and sampling frequency.
the impact of the systematized requirements outside of
y that there it has resulted in better reporting, and up-
frequency and the improvement of plants and opera-
have seen an improvement from 1970 when 70% of
ies did not meet bacteriological standards to 1976 when
to cite two cases in point where failure to meet the
of the Drinking Water Standards resulted in pro-
encouraged needed improvements. The first was Spring-
und the second Pall River, Massachusetts. Until about
of some 160,000 provided water with slow sand filtra-
. Frequent instances of coliform levels in excess of
d been noted and a sanitary survey revealed less than
rotection. The city was requested to, and did, install a
chlorination station since which installation the bac-
been excellent.
969 Fall River experienced widespread contamination
stem and the state issued boiled water orders to its
The PHS issued a "Prohibited Use" certification and
e classification until adequate treatment was assured.
using a modern water filter plant which cost some $4.5
was paid by a HUD grant.
vilillt'-V
,,=4*1 -H
-------
COLIFORM STANDARD: INTERSTATE CARRIERS/F. TAYLOR
Conclusions
I believe I have complied with my instructions to tell you how the bac-
teriological limits of the interstate carrier standards were implemented from
Public Health Service Headquarters and regional offices and what were some
of the results. The third question is more difficult. "Should we continue to
implement the limit for communities in the same way as indicated in the
interim regulations: my recommendation is "yes" as far as the numbers of
allowable coliforms are concerned and; "no" as concerning the monitoring
and public notification requirements.
Let me next say that some of you would have recognized a similarity
between the "provisional approval" of the old ICWS system and the variance
and exemption provisions of the SDWA. These allowed room or time in
which to negotiate. I firmly believe this room to negotiate: plus the Use Pro-
hibited action clause, where needed, accounted for our
ICWS program. These principles should continue to govern our program.
REFERENCES
1. Kent, F.S. 1959. Administration of the interstate quarantine regulations
governing drinking water. Interstate Carrier Section, Creneral Engineering
Program, United States Public Health Service, Washington, D.C. (Available
from Tloyd Taylor, United States Environmental Protection Agency, Boston,
Massachusetts)
success under the
2. United States Public Health Service. 1962. Drinking wat
;r standards. United
States Department of Health, Education, and Welfare, Washington, D.C.
3. United States Public Health Service. 1925. Drinking water
standards. Treasury
Department. Public Health Reports 40:696.
4. United States- Public Health-Service. 1947. New interstate quarantine regu-
lations. Federal Security Agency. Fed. Register 12:3189-3197.
5. United States Public Health Service. 1914. Bacteriolc
drinking water. Public Health Reports 29:2959.
6. Ibid, p. 2960.
7. Water Supply Division. 1973. A guide to the interstate
certification, p.
ington, D.C.
igical standard for
carrier water supply
4, 5. United States Environmental Protection Agency, Wash-
157
-------
ii!1:
.11" t
m
I;1 ..; "I'l, ; !" .a;.; [ *MI'
':,-:.-:; i:, I; IP luf ,,,!<:,,v, ,
'!'!*' iiat i'1-;
"
-f V
'I ! "..|il.ir!! lill
.•'hi!' 1 : i "li iiiifl
":|{I;I!"1 !!" !' • iilflli ;
\' • Ills: 'lif
••' ,; Hi:*-! f:
ll|>!llll !> "Mlliv
i • , • ji'ij, ;,;!
lull ill
1 *
-------
COLIFORM STANDARD: REGIONS/J. HARRISON
Experiences with the Coliform Standard
Joseph Harrison
Water Supply Branch
U.S. Environmental Protection Agency
Chicago, Illinois 60604
Introduction
When Dr. Charles Hendricks invited me—almost two months ago—to
participate in this symposium regarding standards for bacteriological quality
of drinking water, I in turn asked each of the nine other EPA Regions and a
number of States to help me gather the facts and figures that would docu-
ment experiences with the historic coliform standard. I have received re-
sponses from most Regions and about eleven different State agencies.
I will describe some typical experiences that public water systems have in
finding coliform bacteria, and how various State agencies and EPA regions
have treated these findings in the past. There has been more diversity than
consistency in the supervision of bacteriological quality. But I think you
will agree that no State or EPA Region has ever been as stringent as the
National Interim Primary Drinking Water Regulations (NIPDWR) are now
asking us to be.
It is not fair for me to take the entire credit for this paper as it is almost
entirely pieced together from the inputs of others". A number of people have
written me ease history descriptions of bacteriological problems including
follow-up findings and actions. I will quote a number of them directly. The
cases illustrate some important points and I am sure you will enjoy them as
I did. ; ; - -
I guess one reason-that I am here toda.y is that I have some very strong
feeling about the current microbiological standard. It is not logical; water
systems cannot comply with it; and it cannot be enforced as currently written
in 40 CFR 141.14 and 141.21.
Coliform bacteria in drinking water represent an acute rather than a long
term health effect. A single swallow of coliform is just as dangerous as a
monthly average exposure. The bacteriological safety of drinking water should
not be defined by an arithmetic mean of all samples or percent of positive
portions over a month's period. The individual sample limits define the safety
of drinking water, whereas a monthly or yearly average coliform limit per-
tains to the reliability of drinking water systems. By applying both limits, as
we have in the December 24 NIPDWR, we end up with a double standard.
Public water system consumers can, according to our current standards,
safely drink water with four coliforms per 100 milliliters every day of the
month until the last day when this same water suddenly becomes unsafe be-
cause the computed monthly average exceeds one coliform per 100 ml.
It is a good idea to use long term averages or percentages of bacteriological
data to judge the adequacy of water system facilities and operations. But I
am absolutely convinced we must use only the individual sample limits to
define safe drinking water. Monthly summaries of bacteriological data should
do no more than trigger investigative action to ascetrain that a serious health
hazard is not lurking on the sideline arid waiting to emerge. It is not proper
to define water with four coliforms per 100 milliliters as safe to drink and at
the same time say water with more than one coliform is unsafe.
The second major flaw regards public notification.The December 24
regulations require that the public be notified on the basis of unverified bac-
159
-------
Il1' lil1 !,,,
iiui ni'M
;':*
(I )
I,Bill
:,: sit
mini
81
MICROBIOL STANI >ARDS EVALUATION/CJW. HENDRICKS
teriological data.
ICJLlUJliJKIwCU. vicii-ci. LI..L juiAWfcj «u MV——.~-- — - - ^ f j ^
unsatisfactory bacteriological sample result will force nearly all public water
systems in this country to give public notice.
P.L. 93-523 pu
,, i* | f_ y j~*j£*j I-"-*- -*nw .i,LWfc.*i*wt*fcAv/AJ. **p*»jjf *"*»*»*Q —— —— —
such as bacteria and nitrates is .not meant as a warning for consumers to take
__i: :* lt.~:_',»«,'„. V.S.a1f Qtofpo will still TiaVfi to 1SSU6 dlSCre-
emergency action
pendent of SDWA
.n fact, as Section 141.21 (f) is currently written, a single
ilic notification respecting acute health effect contaminants
their own behalf. States will still have to issue discre-
cmcr&cLiw y O.UUX/.LI m niv-*.* w T? ** ^ w ».*«.....•-• ~«~ -—- •• — - _
tionary boil orders and other emergency advisories for this purpose, inde-
requirements.
LJCJL1UWUL \Ji. \JL-f YY.I . JH^V*l.4.wi**«.*..«.
The Safe Drinling Water Act (SDWA) notice, on the other hand, will be
included in the'^ater bill and is rather a more casual news letter and
torical record of
and what is bein,
that a State qua
a past water quality problem, the problem's significance,
done about it. The State implementation regulations say
mat it ocaLC i^fying for primary enforcement responsibility need only
require notice in the water bills or, if the bills are less frequent than quarterly,
"of direct mail. The water bill announcements will be at
icast a wc«. - e rhaps months later than the occurrence of a bacteriological
problem. This no1 ice should further include "a balanced explanation of the
significance or seriousness to the public health, a fair explanation of ^ steps
taken to correct anv problem, and the results of any additional sampling." This
**. - ... • . . J • TT T* M_.A*.* O1__1 1 Q*» TI7T*l/-»Vl
is all in keeping
with the intent expressed in House Report 93-1185, which
A guide to the
i^ OIL 111. j\.w^j^/JLri£f »»it** *.i..i.w **^fcv**- ^ — *.
states: "The purpose of this notice requirement is to educate the public as to
the extent to which public water systems serving them are performing . . .,
to develop public awareness 'of the problems facing public water systems, and
to encourage a willingness to support greater expenditure at all levels of gov-
ernment to assist in solving these problems."
It £ important therefore, that these information notices be factual and
valid We,.must bi certain they are not triggered by spurious results that would
unduly alarm the public. There is, in fact, good argument for public notice to
be determined oaly after the results of a comprehensive investigation. But.
with the inheren: error we will always have in microbiological testing, the
SDWA violationf and public notice should NEVER be based on the result
of a single colifo rm test. If we ever start enforcing on the basis of non-valid
results and no actual health hazard, water system operators will soon begin
to bias samples toward negative results rather than search for water quality
problems. .
I will explain my thoughts and recommendations for correcting these
anomalies in a wiile. But first I would like to relate some specific experiences
with the coliform standard that a number of people have helped me document
for this symposiv m. •
EPA Regions' Experience With the Coliform Standard
Interstate Carrier Water Supply Certification Program was
written in 1971 by EPA's water supply program. It was, perhaps, the first
attempt to establish uniform criteria and procedures for interpreting the
drinking water standards (DWS) and for classifying interstate carrier water
supplies (ICWS). But even with the use of this uniform guide, the 20% of
ICWS recorded ia the latest tally as "Provisionally Approved" does not reflect
the total number of bacteriological discrepancies that were found nor the
number that would have failed the present microbiological MCL (maximum
contaminant level). Professional judgment can never be completely eliminated
from evaluation [of the bacteriological safety and reliability of public water
systems.
160
-------
COLIFORM STANDARD: REGIONS/J. HARRISON
Region III gave two good examples of where judgmenjt
to override a classification that the numbers alone would
was wisely .used
lave dictated.
gical samples per
7 analysis. In July
sample was found
sample made the
100 ml and thus
:ollected and they
was a sampling
to a "Provisional"
present and no
latest changes we
ny—900,000
in the following
Chesapeake City, Virginia—50,000 population served
This supply collects between 70 and 100 bacteriolc
month from its distribution system and performs the M'
of 1976 they collected 92 samples. That month, one bad
with a coliform density of 520 per 100 ml. This one
mean density for that month calculate out to 5.7 per
exceed the standard. Appropriate check samples .were i
came out negative. It is a very good possibility that theire
error. This should have resulted in an ICWS downgrade
classification. In our judgment there was no health hazatd
action was taken to downgrade the system.
The December 24 NIPDWR are written—even with the
heard about today—to take this latitude away on June 24
Another case involves the Philadelphia Surbah Water Compa:
population served. Region III accepted the explanation givejn
letter without any downgrade in classification.
Mr. Walter Stanley ' :
Pennsylvania Dept. of Environmental Resources
1875 New Hope Street
Norristown, PA 19401
Dear Mr. Stanley:
This will explain, for the record, why we reported a mean coliform
density of 2.3 organisms/100 ml for the week ending July 28, 1973.
During that week a number of samples from our system showed the
presence of coliform organisms. However, we began to doubt the veracity
of the results since they reflected points widely scattered throughout our
several divisions (300 square mile area).
A thorough investigation made early in August showed that the "bad"
samples were not due to the source—but to one particular sample man.
We established this fact by having a second man follow the first one and
sample the same points for several days. Without exception, the first man's
samples showed coliforms, whereas those collected by the follow-up man
were negative.
Even a thorough medical examination failed to show how the first
sampler's samples became contaminated. Nevertheless, we took this man
.,, _,rt* ' j.1. ._ —-- TI -_».. =- — .^. -3 - -«.«•< r,*. 1+ lvvt -- -J-1-. rx-*«/^ii fVti 4"fn imnrt if» r^O^*'f O1*1J"\lrt
-------
!? ;;; ?: f j if >wi i
MICROBIOL STANDARDS EVALUATION/C.W. HMDRICKS
, . i , . . , . ,
determining the degree of health risk and appropriate follow-up when bac-
teriological monitoring raises a question about the reliability of water system
facilities or operation.
Region IV uses a case history concerning Florence, South Carolina, to
point out two key weaknesses in the microbiological standard:
On February 24, 1972, the Region was advised by the State that the
Florence System had had eleven highly contaminated samples on February
14 and 15, such that they would surely fail for the month. At that time,
it also became known that the system had failed in October 1971, based
on one sample out of 36 unsatisfactory with a count of 47 for a mean coli-
form density of 1.3. Subsequent sampling (not rigorous check sampling)
showed no problems. After a joint USEPA-State investigation, the system
was classified USE PROHIBITED on February, with appropriate telegrams
and a press release.
This instance illustrates two problems with the Microbiological MCL:
(1) One single unsatisfactory sample in October caused the system to fail
and would have necessitated public notice under the Interim Primary
Drinking Water Regulations; (2) Official action on the flagrantly bad
quality detected in mid-February was not warranted until the end of the
month, when problems had disappeared.
It should be -noted in this instance that although the Florence system was
classified "Use Prohibited", the State gave public assurances of no health
danger. State agencies in the past have not rigorously determined compliance
with bacteriological standards. Even though State standards are very similar
or identical to the PHS standards, enforcement actions are only contemplated
. when contamanition is substantially more serious than that minimally neces-
sary to fail the MCL. Many States feel the real value of the bacteriological
test is to paint a monthly arid yearly picture of water quality which can be
considered, along with knowledge of the system and its operation, to reach
a conclusion as to whether that system is adequate. Some States do require
resamples whenever routine results indicate that coliform is present. But the
most effective follow-up is that provided through direct one-on-one contact
between the field engineer and the water supply. This follow-up is by and
large at the discretion of the engineer. One such person in Region V who, I
know, is doing an outstanding job gave me this description of his operation:
It is my policy that when one sample has a count of 10/100 ml or more,
or two or more samples show contamination, to call the results to the
supply operator. At this time we attempt to determine if anyhing has hap-
pened which might have caused the contamination. Based on our telephone
conversations, we make specific recommendations such as: boil orders, in-
creased chlorine residuals, collect samples from proper taps, resampling,
etc. Recommendations obviously vary considerably depending upon infor-
mation obtained.
I would not want to see this work interrupted for the purpose of enforcing
additional monitoring, reporting, and public notification where water quality
is not impaired or a health hazard does not exist.
One of the most striking statements I received was written by Leroy Stratton
of the Illinois Department of Public Health. He noted that facilities which
would be regulated by his Department "are primarily non-community sup-
plies with none exceeding 1,000 population." These systems will be required
to sample on a quarterly basis, therefore "requiring public notice any time a
contaminated sample is received, regardless of the level of contamination." I
would add that two-thirds or more of the nation's community water systems
are also less than 1,000 population, likely to be sampling once every three
162
-------
COLIFOBJM STANDARD: REGIONS/J. HARRISON
months, and therefore require public notice every time a contaminated sample
is received.
Don Taylor, in EPA's Region IV office, examined 12 months of records
for 122 public water supplies and reached these conclusions based on the
December 24 regulations:
"1. The "average" public water supply will fail the monitoring require-
ments almost 5 (4.77) months each year.
2. The "average" public water supply will fail the bacteriological qual-
ity provisions approximately one month in each 21 months of
record.
3. The "average" public water supply would go to public notice ap-
proximately 5 or 6 times per year."
One final example again from Region IV points out the fallacy of the
monthly average limit for dictating water quality safety.
An interstate carrier water system, Ingalls Shipbuilding-West in Pasca-
goula, Mississippi, showed chronic friability to consistently maintain bac-
teriological quality in 1972 and 1973. Strict application of the PHS Drink-
ing Water Standards and the "Guide to the Interstate Water Supply Certifi-
cation Program" resulted in the following sequence of actions:
3/10/72—System classified USE PROHIBITED (with press release)
based on failure to meet standards 11/71 and 2/72.
5/11/72—System reclassified PROVISIONALLY APPROVED (with
press release) based on 3 months good record.
,.2/15/73—-System classified USE iPROHIBITED (with press release)
based on failure in 1/73.
8/ 1/73—System reclassified PROVISIONALLY APPROVED based
on 4 months good record. (System also failed in 2/73).
9/14/73—System classified USE PROHIBITED (with press release)
- :- based on failure in 8/73'. •
""" 1/17/74—SystemT reciassified PROVMONALLY APPROVED based
on 4 months good record.
9/19/74—System reclassified APPROVED based on 12 months good
record.
The classification actions outlined above resulted from five individual
months of failure to meet standards. In only one of these months was the
failure the result of a single unsatisfactory sample (1 unsatisfactory out of
4). In the other four instances, at least two samples were unsatisfactory.
This example points out the problems of small systems meeting the
standards, even with onus of unfavorable publicity.
? The official record "includes' another noteworthy month. In August 1972,
83 samples were collected, one was unsatisfactory with a count of 80.
During the previous and subsequent months four samples had been col-
lected monthly. The unsatisfactory sample would have meant another USE
PROHIBITED situation, and evidently approximately 79 additional sam-
ples were taken to bring the system into compliance.
The best way to eliminate the monthly average dilemma is to eliminate the
monthly average limit. A better monthly summary is simply the number or
percentage of positive coliform results in «ny one month. But an important
distinction must be made between any monthly summary coliform limit and
an individual sample limit. The first pertains tothe reliability of water systems,
and the second pertains to the safety of drinking water.
The occurrence or recurrence of any coliform in drinking water should be
investigated to ascertain and correct potential health hazards. But only one
limit should actually define safe drinking water. We must remember that un-
163
-------
:wir ""; r^i
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
I!'1 i;i!*iil >, ' I <:!l ni
reliable water systems can produce safe drinking water, and discrepancies
respecting a low monthly summary limit may be only a potential rather than
an actual or existing health hazard.
'•• • '"" ';»''" • j ''"Proposals
d! | , : ri,jij: iiiiili! i-1''(it '" . 'Li. 1'.. • S ' ;. ,;•' r! -i;/vt •'•( "iv: ,""11-, ".•. * ;m-" folOi "*! 1».: ;.pt) li'th " 'Wi,!' ; ::",,!n,i!i ',:•:«;*»(
As stated in the preamble to the December 24 National Interim Primary
Drinking Water Regulations, "further changes in the microbiological regula-
tion may be desirable." A number of people other than myself have given
this considerable thought, and have had input into the following two pro-
posals. These two alternatives I feel are logical. Either one, if chosen, is prac-
ticalenough to allow implementation.
The text of Section 141.14 and parts (d), (e), (f), and (g) of Section
'iiji':; •• ^//yE; = 'ii 14121 in the December 24 regulations should be deleted. Part (h) of Section
*:'•.!'.' ••'•<:*'-:)*!• *': ,": M S'l""should be labeled (g). The following then inserted as replacement
lliiiiii" ' I'iiii'll't:, "viis
III i I P 3 ' ; r-: B sit
and change:
;:: Section 141.14 . " ' "
The following are the maximum contaminant levels tor colirorm
bacteria applicable to community water systems and non-community
'••V.''w||er systems'; Compliance with maximum contaminant levels for
cbliform bacteria is determined pursuant to Section 141.21.
The last sentence is consistent with the MCL. regulations for organic and
inorganic chemicals. It allows compliance to be defined by follow-up activi-
ties," an.3it prevents causing violations and public notice on the basis of a
single sample result.
;; ; " ;';;;;;, ' "" ' " -. PROPOSAL i ,' . , '"|
Section 141.14 (Cont'd)
(A) When the membrane filter technique pursuant to Section
141.21 (a) is used, the number of coliform bacteria shall be
less than one per 100 milliliters in any sample collected and
analyzed pursuant to 141.2l(b) or (c).
(B) When the fermentation tube method and either ten milliliter
standard portions or 100 milliliter standard portions pursuant
f to Section 141.21 (a) are used, coliform bacteria shall not be
present in any portions in any sample collected and analyzed
pursuant to Section 141.21 (b) or (c), ...'..'..
Bacteria in drinking water represent an acute health effect and the presence
of any coliform organism in finished drinking water suggests either inade-
quate treatment or access of undesirable materials to thejwater after treat-
ment. Coliform also may be indicative of dangerous contamination and of
Water that is not safe to drink. The limit therefore simply pertains to the
presence of any coliform in drinking water". It eliminates the need for an
arbitrary allowable number and it is easily understood.
;:."•!| :',;•r*!i "i -,...;.,-Section 141.21
(d)(ij When a sample exceeds;"a maximum contaminant level set
forth in Section 141.41 (a) or (b), one check sample shall
be collected from the same sampling point and examined.
If a subsequent sample has. already been taken from the
sampling point, it shall be considered the check sample.
If a supply monitors regularly as prescribed in (b) and (c) of this section
and there does not appear to be serious contamination, then one check sample
should be enough. However, paragraph (3) will allow the State to require
"I", '"',', :" '164 ' " " '
-------
COLIFORM STANDARD: REGIONS/J. HARRISON
additional check samples if a health hazard is suspected. Counting the next
subsequent sample as the check sample makes a great deal of sense but wasn't
spelled out in the December 24 regulations. Many water systems will have
already resampled from the same tap before the results of the previous
sample are known.
(2) When the examination of the check sample required in Sec-
tion 141.21(d)(l) shows the presence of coliform organisms,
the supplier of water shall:
(I) Report to the State within 48 hours;
(II) Initiate an investigation, including the collection and
examination of additional check samples from the same
point and other sampling points in the area, to define
the extent of the problem; and
(III) Notify the public in the area affected by the indicated
contamination as prescribed in Section 141.32 or the
State public notification requirements, unless the State
determines that no health hazard has actually existed.
Based on the results of the check sample, the supplier of water must
1) report to the State, 2) initiate an investigation, and 3) notify the public.
An. important provision regards public notice requirements only to the area
affected by the indicated contamination. For large regional water systems it
is folly to require notification to perhaps one-half the State's population when
the problem may be limited to one section of one town, one waiter main, one
building, or one faucet. Of course, the utilities would have the option to
notify all consumers if they so choose.
(3) The State, at its discretion, may require that additional check
samples be collected at a specified frequency from the same
sampling point and other sampling points in the area and ex-
., amined to identify and eliminate suspected health hazards
when a sample exceeds a maximum contaminant level pursuant
to Section 141.14(a) or (b), even if the check sample required
in Section 141.21(d)(l) does not indicate the presence of
coliform bacteria.
Some sample results or bacteriological histories may raise a question about
the reliability of the system to continuously supply safe drinking water even
though no unsafe water has yet been found. In this case the State may re-
quire additional samples to identify and eliminate suspected health hazards.
(4) When the cause of the indicated contamination has been deter-
mined and corrected, additional check samples shall be col-
— lected at a frequency directed by the State. :
Just to confirm that the correction "has been effected, additional samples
may be wanted for a specified time period.
(5) The location at which the check sample was taken pursuant to
Paragraph (d)(l) of this Section shall not be eliminated from
future sampling without approval of the State.
Same requirement as in the December 24 regulation.
(E) The State may determine that unreliable examination results
for a sample collected in a discreet monitoring period pursuant
to Sections 141.21(b) or (c) were caused by factors beyond
„_„„::__„ the control of the water" supplier. Such factors could be ex-
cessive transit time between collection and examination of the
sample, samples being broken in transit, or interference test
""results When the membrane filter technique is used. If this is
the case, another sample collected immediately upon learning
165
-------
III"!) i1 I ! ' I'll!' 'I'liillllllllllllll'llli
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
ill
111!! I
llni '
»'ii1'!'!1:'!!!1 ' 'I
!|i:!!!:'<::! L
I'll*.' •'•! in tii;"
o
I
ii "-nil. Mil : t •;{:''t1"'!1 ;l""!
w
j"!!;.:,, ,l!"'.!l,'!.f :\if-if, .Ji'i1.':"1'!,'"' {3
.; :: : I
:: ;: • -'i '.
'Sill ''''Hiiil !'•. . l;-»>! :::., 'if
i ....... sill' ...... Illiallll ,' r", ' ( -i
" 1
'•(• ::j .-!!,: ! t
!'• Tail1' i! !!!,!"' , ' ''iJilim
1
A
0.0
fi S
y <
11
•B8J8
•a •* o o
aE*
33
t- OOVD C
-------
COLIFORM STANDARD: REGIONS/J. HARRISON
of these results may be used in determining compliance with
sampling requirements in Section 141.21(b) or (c). How-
ever, a single sample may not be attributed to more than one
monitoring period.
Overgrowths, "too long in transits"—over 48 hours not 30 hours, and
broken bottles frequently account for 10% of the samples being invalid or
the results unreliable. These samples should be disregarded, but unless the
operator is then allowed to collect and count-another sample, he will be in
violation of the monitoring requirement through no fault of his own.
(F) Check samples, samples with unreliable examination results,
and special purpose samples,, such as those taken to determine
whether disinfection practices following water main place-
ment, replacement, or repair have been sufficient, shall not
be used to determine compliance with Section 141.21 (b) or
(c).
Same as in the December 24 regulation, except samples with unreliable
examination results are not counted as well.
Table 1 was compiled with data from 10 different Regional and State.
agencies. It shows that ten percent of public water systems experience one
to four coliform organisms per 100 milliliters at least once per year. Seldom
have these occurrences been found to present a health risk. Another ten per-
cent of public water systems experience over four coliforms per 100 milli-
liters at least once per year and twenty percent of those occurrences have
been found on follow-up to be of signiJicance.
These data point out that a limit of one coliform may be more regulation
and more expense than is needed to define safe drinking water. Federal
drinking water standards have for over 50 years defined water with four
coliforms per 100 milliliters as safe to drink and there is no evidence now
saying it is not. Some State agencies from a practical ^standpoint may want
to or have to prioritize follow-up regarding bacteriological results. This lati-
tude to four has historically existed in the Public Health Service standards
and may be needed in the future when SDWA requirements are applied to
hundreds of thousands of small public water systems. Proposal 2 divides the
standard into two parts—141.14(a) and (b) addresses the safety of drinking
water and 141.14(c) addresses the reliability of drinking water systems. The
first paragraph of Section 141.14 and changes to Section 141.21 would re-
main as I have discussed with Propsal 1.
PROPOSAL 2
(Section 141.14 cont'd)
(A) When the membrane filter technique is used, the number
of coliform bacteria shall not exceed four per 100 milli-
liters in any sample collected and analyzed pursuant to
Section 141.21 (b) or (c).
(B)(l) When the fermentation tube method and ten milliliter
standard portions are used, coliform bacteria shall not
be present in three or more portions in any sample col-
~ . • . lected and analyzed pursuant to Section 141.21 (b) or (c).
(B)(2) When the fermentation tube method and 100 milliliter
standard portions are used, coliform bacteria shall not be
present in five portions in any sample collected and
analyzed pursuant to Section 141.21 (b) or (c).
167
-------
iMIfiii!!
1P1
II
fliii! Nli'K'i
! s
Ilin;,.!!!!,. i 'ill
l;i i'illtt ',! I ,. » . Illli
iiK M:n;i ' I!!5
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
(C) The State and public water system shall initiate definitive
••',- action to identify the cause of the positive bacteriological
sample results and to eliminate potential health hazards
which migh exist in "the system when monitoring pur-
suant to Section 141.21 (b) or (c) shows the presence of
any coliform organisms in either of the following:
(C)(l) more than ten percent of the samples when 20 or more
samples are examined per month; or
(C)(2) more than two samples when less than 20 samples are
examined per month.
" ' ''"'i ,„;' in !!!!,i, I J : 'tli ' i * j " 'B Ii1" iV'lii1!1 ''''liilv/ffliNH"!1!;11!! ,:: /I"." i H» 4 ii.dl I'1*'1 "' 111, iiJI'iT •' i, ;:•'; '!',.' T "iiv
.QUESTION AND ANSWER SESSION
- ,''. ; •'••.. ;;s •';,;." •:• < ;;sf 4f
C.W. Hendricks, Office of Drinking Water, U.S. Environmental Protection
Agency, Washington, D.C.
Do you have any recommendations for a simpler standard of sample col-
lecting to eliminate the confusion caused by terminology of "less" or "more
than 20 samples per month"?
II III III! II ,' I'llV "iii'i'ii iiii!1!!1 i,' " i:jj; ' '? " i ill;1 »!|l ,» "'iii!!!,1!;!'1 ',!'!'lii!li?''''iii'ii liiiii* «"''ii 1 • '!">i: ^nii' itrA Km 'i'lj!' r, N v ' , "!!ij|,iii,!"!'', ,. M •',' ," i £
J. Harrison, Water Supply Branch, U.S. Environmental Protection Agency,
Chicago, Illinois.
My proposal 1 and the first parts of proposal 2 simply refer to the result
of any sample, and eliminate confusion caused by different standards .for
those systems collecting less or more than 20 samples per month. This ade-
quately covers a definition of safe drinking water. However, if addressing
the reliability of water systems, one must be concerned about repetitive sample
results. For those systems collecting 20 or more samples per month, it is best
to regulate repetitive results by a percentage of all samples collected; for
those collecting less than 20 samples, by a simple number.
MJ. Allen, U.S. Environmental Protection Agency, Cincinnati, Ohio.
Are you defining a "spurious" water sample as one sample which results
in raising the monthly average greater than maximum MCL? If not, what is
your definition of a "spurious" water sample?
J. Harrison
I used the word "spurious" in the context that Safe Drinking Water Act
notices to the public should not be "triggered by spurious results that would
unduly alarm the public." By a spurious result I mean a false or illegitimate
analytical result that does not reflect quality of the water but is rather caused
by an error in analysis, sample collection, transit, or other anomalies not
directly related to the water supply system. Such anomalies and false results
do occur and they are the reason that nitrate as well as all other maximum
contaminant levels, except coliform bacteria, are never violated on the basis
of a single unverified sample result. It is not proper that the December 24
Regulations allow spurious results to raise the monthly average greater than
the MCL, and require over 90 percent of all public water systems—those
required to collect only one sample each reporting period—to notify the
public every time a positive bacteriological sample result is obtained. My pro-
posals 1 and 2 would both define a microbiological violation and attendant
notification on the basis of two—the check sample also being unsatisfactory—
analytical results.' ' '
168
-------
IMPLEMENTATION: COMMUNITIES/RJ. BECKER
Implementation of the Coliform Standard
by a Community Water Utility
Robert J. Becker
Indianapolis Water Company
Indianapolis, Indiana 46202
As the sole representative of the public water supply industry, I feel it
is appropriate to begin with a disclaimer statement. Namely, that the views
and opinions expressed by the speaker do not necessarily represent those
of all community water supplies governed by the Safe Drinking Water
Act.
I must assume my presence here today reflects the desire of the Environ-
mental Protection Agency to obtain and evaluate the bacteriological operating
data from a representative and responsible water utility. Thanks to a number
of illustrious predecessors, the Indianapolis Water Company has a continuous
record of bacteriological data on raw, treated, and distribution system samples
dating back over fifty years.
The Indianapolis Water Company is an investor owned utility serving ap-
proximately 700,000 water consumers in Central Indiana. Currently 95%
of our annual supply comes from two surface streams which during low
flow periods become rather scrawny. Each supply is supplemented from
upstream storage impoundments during these low flow periods. These
streams receive a significant proportion of municipal wastewater discharge
as well as agricultural runoff. While they are undoubtedly not ideal supplies,
I feel certain they are far from the worst.
Our treatment facilities would be considered fairly conventional. Coagu-
lation is with alum, and the mixing and settling basins provide a designed
retention time of between two and onie-half to four hours. Disinfection is
achieved with chlorine application prior to mixing and settling at a rate
sufficient to maintain a free chlorine residual through the filters. All filters
are of the rapid sand type with a six inch anthracite coal cap.
As is the common practice of almost all water utilities in this country
we rely almost totally on chlorine to provide a bacteriologically safe water
to the consumer's tap. During the winter months chlorine residuals leav-
ing our treatment facilities are maintained slightly in excess of 1.00 ppm.
This residual is increased to about 1.50 ppm during the hot summer months.
Normally, we apply anhydrous ammonia to the filter effluent to convert to
a chloramine residual prior to delivery to the distribution system. We
realize that a chloramine residual is" a less potent bacteriocide than a free
chlorine residual. Therefore, for about a four week period each spring and
fall we discontinue ammonia application, and treat the entire system with
a free chlorine residual. It is important that this occur while water tempera-
tures are not much greater than 50 °F.
As a system supplying a population of approximately 700,000, we are
required to collect 250 samples monthly. Actually, we will run about 375
samples per month. Our distribution system sampling points are selected for
their strategic location, general sanitary conditions, and availability for our
regular entrance. At .the present time, we have 53 sample sites located
throughout "our service areaT Due to distances involved to the extremities of
the system, we have four samples trips. Each trip involves 15 to 20 samples.
Obviously, some locations are sampled, more frequently than others. It Is
our intention to have the preparation of all samples completed within
four hours of collection. All of our outlying booster stations and elevated
169
-------
!r Tf '""S*!!!!*
Li "If!11,
I1- ran1
• If I!"* "I i! in
.• • '. ••: ;:,;i .I . ''">;,: ' .,.• iiiiiF1 m "!;. •
!' ''I.1' ' ' ' ; i ""If U,1' 111! 1 •!'"' '|l"
i " " ;! i, ;; . "'.ii,,!' : ' ijliii'1!,,,, I'ly,:1 "''
; ' III!1)!:,] !« ' .Uj'i,
'. ' I „ ' • " , I si: I" kft'ni
: , •; , 'i i i • P; i1 ,„ 1 i , iV" " i ji!'S;," If:,"!, :, ' '
, 'I' 111''! II'1 , . „ : i, ^l < . . ', llllll'llilj III!" '•':
j iiil
'|L VI
.'•••! 1 "
„: i ,:;)!]
1 ' •:•
i1' i1
i) •: i . ; : '-jiii iiii:':
'.jii i,i' i n iiiii,
" :•'! ' I ' !' ' '.ill I!!1
•ifi1 ! '„(;,; I,,;! "Eir|ii
i, i „ ; '"''Si1 I-111
i ii : ii i , iiii"1
n^llt i , '. iiji'liiii'
-AVD/NoixYmvAa SCTHVONVXS loiacraoiw
si!
;"i'i '. il i iPil'.f
li <
IIIII III I 111 i!ii!l!,;Liill'i!'!ii:lllllll»ailr ,
-------
IMPLEMENTATION: COMMUNITIES/!?..!. BECKER
tanks serve as sampling sites. Other sites include fire stations, filling stations,
and the like. We purposely have a few sajmpling sites on or near dead
ends. We are trying to 'maintain a knowledge of water quality throughout
the system, and not just obtain a good recor'd to impress the public or the
public health authorities. We do not now have nor ever had during our
period of record any open or uncovered finished water reservoirs.
There are many factors which influence t|he water quality which is de-
livered to the consumer's tap. One of the grejat limitations in achieving uni-
form quality throughout an entire distribution system lies in our inability to
maintain adequate levels of chlorine rissiduals indefinitely. We have docu-
mented this occurrence in a series of laboratory experiments. Sterile gallon
containers of the White River treatment plant effluent were stored at 68°F
and 80°F. Identical containers were used jfor both waters with a free
chlorine residual and a chloramine or cornbined residual. Figure 1 il-
lustrates the persistence of chlorine residuals in relation to both time and
form. Chlorine residuals were determined by starch-iodide titration as speci-
fied by Standard Methods. As you can readily see, the water temperature
and the form of residual make a substantial idifference in the persistence of
the residual. This work was done under fairly ideal laboratory conditions.
I doubt that chlorine residuals in our distribution piping would persist as
well. |
For comparative purposes, we ran a similar test on our best well water
supply. Figure 2 illustrates the difference in the persistence of free chlorine
residuals between our surface water treatment plant effluent and this particu-
lar well water supply. These findings correlate perfectly with the work done
in Detroit by Mike Taras on the chlorihation of nitrogenous compounds.
Another major obstacle for the maintenance] of high water quality through-
out the distribution system is the. design criteria. Pipe sizing is designed
usually to meet fire flows ~6n fop of normal (system demands. This unques-
tionably is a deterrent to desirable displacement, particularly in a remote
residential or low water use area. The location of sizeable elevated or
ground storage facilities on this piping network adds still another obstacle
as positive circulation is not normally achievjed. In a system such as ours,
which includes 15 different pressure districts, the division valving creates
additional dead ends to those normally encountered in a system that is
subject to substantial growth in all directions. We have a total of over
1750 dead end situations in our system. I mention these factors which
adversely affect circulation within the system only to emphasize the com-
plexity and magnitude of the problem we face.
Despite our. utilization of a chloramine residual during the summer months,
we still lose our residuals in distribution system storage reservoirs and in
remote portions of the system. In recognition! of this problem, we have in-
stalled supplemental chlorination equipment in all booster pumping stations
and elevated storage tanks that have been constructed on the system since.
1956. These installations have become more sophisticated as our experience
and confidence have grown with unmanneli automatic chlorine feeding
systems. They are normally in operation from about May through November.
We do perform all of our own bacteriological analysis generally as pre-
scribed in Standard Methods. The laboratory personnel, equipment, and pro-
cedures are all inspected and. approved annuaHy by the Indiana; State Board
of Health. Following an .extensive study conducted in bur laboratory in the
mid 1940's we selected Lauryl Ttryptose Broth for all sample tubes on the
basis that it resulted in less false presumptives than Lactose Broth on our
particular waters. We do make practice of carrying to completion the testing
of gas formers from finished water samples.1 For over ten years we have
171
-------
sili I!,!!!
ill! ilsM
!
;j II I ;|! i!i |! ii
15
DAYS
Figure i PERSISTENCE OF FREE CHLORINE RESIDUALS AFTER 5 HOUR CONTACT TIME.
§
I
WJ
I
I
kH
i
"n
^
a
o
1
GO-
-------
IMPLEMENTATION: COMMUNITIES/RJ. BECKER
run parallel comparisons between the membrane filter technique and
multiple tube for various selective locations. As a result of this study we
have elected to stay with the .multiple tube procedure for all of our bac-
teriological analyses. We believe it is the more sensitive for our particular
situation. This should become apparent as we examine the data which
will follow.
Certainly one consideration in the bacterial quality of the finished water
is .the bacterial quality of the sources. Since the White River provides ap-
proximately two-thirds of our total production, let's examine this supply
from a bacteriological standpoint. Table 1 enumerates the monthly average
MPN for the 10-year period 1967-76. From the table it is fairly obvious
that the quality is subject to considerable variation. Generally the months
with the greater numbers reflect a period of higher surface runoff. The
last two months of 1976 correspond to an extremely dry period with neglible
surface runoff. By some standards this would not be considered a particu-
larly desirable source. However, it happens to be the major waterway
through the central portion of Indiana,
In recent years we have also been testing our raw waters for the more
specific type fecal coliform organisms. Table 2 illustrates the results of this
particular series based on twice a week sampling for the past year. Monthly
average numerical comparisons are shown for the MPN as determined by
the multiple tube procedure, the MPN as determined by the membrane
filter technique utilizing M-Endo Broth, the fecal coliform count utilizing
M-FC Broth and incubation at 44.5-° C, and fecal strep, counts utilizing
both M-Enterococcus and KF media. As you can readily see the fecal
coliform numbers are significantly below those obtained for the total coli-
form. For the membrane filter, fecal coliform, and fecal strep, determina-
tions between .5 and 5 ml of raw water are used depending upon raw
water turbidity and bacteriological quality. --.-—?•.-
Table 3 illustrates the bacteriological results of the treatment plant efflu-
ent and distribution system samples over a 50 year period at 5 year inter-
vals. WR indicates the effluent samples from the White River treatment
plant, and FC indicates the effluent samples from the Fall Creek treatment
plant. The Fall Creek rapid sand treatment plant was placed in service as
a secondary supply in 1941. At the: White River plant the initial rapid
sand filters were placed in service in 1926. Prior to that all filtration was
through slow sand filters since about 1905. Additional rapid'sand filter ex-
pansion at this plant occurred in 1953, 1959, and 1972. Back in 1925 and
1930 the chlorine residuals leaving the plant were in the range of .10
ppm. This would be considered an extremely low range- according to our
current standards. I would sincerely doubt that it was a breakpointed or
free chlorine residual. A good understanding of this chlorination phenomena
did not occur until the mid to late 1930's. In reviewing this data I feel it
is apparent the conventional treatment process with responsible operators
can provide an exceptionally safe bacteriological product regardless of varia-
tions in the quality of the source. Likewise the data indicates that this
quality does tend to deteriorate slightly in the distribution system. While we
are able to show a significant and consistent improvement in the coliform
recovery from the system samples, we.have not shown the same degree of
success with regard to the total (standard) plate count. Undoubtedly the im-
provement in coliform recovery is largely the result of leaving the treatment
plants with higher levels of chlorine residuals over the years.
173
-------
inn 'iii I in mi iiiiiiiiiiiiii'iiii iiiiiiiiii
I '(I I'1!!! I I I i I1"!'!! I I i Ml l|i|i||l I id
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
1 lillil|li inn
O
00
§
vo
vo vo
ON r-
W
•*
r~
J> r-
&
, ..... .1 1 •:,* , .. ..... .;»', . , af • aiiis • •.
[[[
•.•f»:v
iii ' ;>
"'::
,
sill "E t-i
" ~
"1 O
i ,,« O
' :'
IV • RliJi , ! , !' n,i M""
.i!'!1!!!:1 i i-i {;.:,
i!1 ..... "H i. "luiiifl ' .
f ..... ; -it • ;,>y:
.._.
&
, g
i
13
l
VO VO
'*"
CS
JO vi oo
ro O\ CN
10
Cl
oo
r,«oCvo-o
10 O 00 t-l «0 CO O t- O O, 00^ r-^VO^
«o" ON ON r-T T-? CO" rf vT rt" •* C<1 O JO
_J vo ra 1-1 r~- CN
X^ ^H" ^ 1/-T CO" U-T VO VO . V
1-1 10 i-l ^H
?2 18
cQvJvOVOOi-iCNVDVO OO,1'"!
eo" r-" i-n" vT oo" c-f i-T t-H T-<
C^i "^1" CO C^ ^^
s;
fe- fe
^? ^ >5
o
'cf
CO
B
p
O i-l CO ^ JO
oo ON ON cN O "^r
» oo vo •* V)
"i co r-- o
VO CO V> CO VO CO
oo fN vo co
CM c^l ON ^5 JO
. ON ON r>J O '*
co o\ 1-1 '"fr f^
O t-- -t
O •—I O\
•^- OJ •"*
&
<
S
fe
fe
t-. r-- O oo
o c»i r~
Tj- •* CO
CO ^ 0
00 CO 04 O
co co
vi
0
V)
co oo O »n ON vo
oo tN ON ON 1-3 '^r
vo co fN TH t^
t«. in vi o 10
r—I C-l t~ O 'T
VO fN . C*
5 oo m vo ON oo
t-.. t-» ^j- co vo
cTof I-H"
co^
vo
ON_ ON vj_ oo (N
*j
a, g
o o
U (U
OH PH
iiiiiuiii ( ..... iii iiiiiii i ii in i iiiiiiiiiiL i, iii I iiiiiii .......... 'in 1 1 iiiii
174
i ..... i ....... mi , i ii iiiiti .......... iiii
am im, ..... \IU:M liii ; s ...... IIIIMII v^ ..... [in.;. :iii:ii:ii£! ...... '
-------
IMPLEMENTATION: COMMUNITIES/R J. BECKER
TABLE 3. ANNUAL AVERAGES.
35° Plate Count
Plant Effluent
WR FC
1975
1970
1965
1960
1955
1950
1945
1940
1935
1930
1925
AVG.
1
2
2
2
1
2
2
3
5
4
. 10
3
3
1
2
2
1
1
2
2 .
% Coliiform
Plant Effluent
WR FC
.06
.01
.02
.01
.01
.05
.18
.36
.00
.00
.00
.07
.28
.01
.17
.00
.01
.03
.03
.08
35° Plate Count % Coliform
Distribution System
System Distribution
75
300
18
100
300
110
60
85
11
210
325
140
.14
.29
.29
.32
.44
.67
.80
: 1.48
1.40
1.50
2.32
.88
Table 4 examines the average monthly distribution of system bacteriologi-
cal samples for the 25-year period (1951-1976). We examine approximately
375 samples or 1875 tubes per month from the distribution system. This is
125 samples over that required by the Drinking Water Standards on the
basis of population served. The numbers under the coliform heading
represent the average number of tubes that completed as coliforms for
that particular month. Naturally you recall that a total plate count limi-
tation was proposed in the Interim Primary Drinking Water Standard.. We
objected rather strenuously on the basts that our_historical record^, indicated
we regularly recovered" counts in excess of this number, and we were un-
aware of any health related significance. The second row of numbers indi-
cates the monthly average total plate counts. With a few individual samples
reaching 10,000 colonies it is not difficult to reach a high monthly aver-
age. The lower row of numbers indicates the number of plates or samples
which would have exceeded for the month the proposed 500 plate count.
Quite obviously the deterioration in bacteriological quality is a seasonal
one for us peaking during the hot summer months. Water temperature plays
a part in a number of influential parameters.
Table 5 provides an analysis of all distribution system samples on a monthly
basis for the past ten years (1967-76). The seasonal trend which^appeared.
so traditionally in the previous slide is not so apparent in this one. However,
we can determine that during this particular period that gas formers were
recovered in .29% of the tubes examined. Likewise that the number of coli-
forms which completed averaged .19% of the tubes examined or almost
exactly two-thirds of the gas formers recovered. The irregular pattern shown
in this slide perhaps introduces still another variable. Certainly the mis-
handling of the sample in either collection or preparation can lead to en-
tirely erroneous results which nevertheless influence the quantitative measure
of water quality or safety. While this does have a slight impact on our
monthly averages, it might be critical for a small utility with a limited num-
ber of samples.
Table 6 analyzes each sample which resulted in gas formation in three or
more 10 ml portions for the years 1970-76. For each sample is shown the
chlorine residual, total place count, number of tubes* showing gas "formation,
and the number of tubes completing as coliforms. Almost identical data is
175
-------
TABLE 4 25-YEAR MONTHLY AVERAGE (1951-76) DISTRIBUTION SYSTEM SAMPLES.
Jan. Feb.
COLIFORMS
TOTAL PLATE
NO. PLATES
— EXCESS 500
Jan.
Feb.
Mar.
Apr.
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
Total
Avg.
COUNT
TABLE
No.
SAMPLES
3,782
3,596
3,841
3,690
3,779
3,769
3,765
3,882
3,623
3,961
3,440
3,698
44,826
3,736
3 4
8 8
1 1
Mar Apr.
.3 4
6 6
0 1
May June
6 9
41 78
5 10
5. DISTRIBUTION SYSTEM SAMPLE
No.
TUBES
18,910
17,980
19,205
18,450
; 18,895
18,845
18,825
19,410
18,115
19,805
17,200
i 18,490
224,130
18,680
GAS
FORM
26
44
22
53
64
27
98
77
110
53
40
46
660
55
.29%
0/5
3,764
3,575
3,831
3,667
3,756
3,760
3,733
3,853
3,596
3,950
3,423
3,667
44,575
3,715
July
10
318
25
ANALYSIS
1/5
13
11
7
13
12
7
21
17
17
8
13
28
167
14
Aug. Sept.
12 9
439 262
38 31
(1967-1976).
No. of Samples
Oct.
7
133
19
2/5 3/5 4/5
5
4
1
3
3
0
5
5
3
2
2
2
35
3
0
3
0
3
2
0
3
1
3
0
0
0
15
1.3
0
1
0
1
2
0
2
4
4
0
2
0
16
1.3
Nov.
3
52
7
5/5
0
2
2
3
4
2
1
1
0
1
0
1
17
1.4
Dec.
3
18
6
TOTAL
COLI
23
• 42
19
47
52
17
53
51
48
17
25
37
431
36
.19%
o
w
s
f
§
§
o
G ; :•
H
o • i "
55 ' "_'
p ; . ^:
s • • "
55 ' ";"
l? V
o . i1.
on ; 't\'
s
j ,, -„-
-------
TABLE 6. DISTRIBUTION SYSTEM COHFORM RECOVERY (3 OR MORE—10 ML) ANALYSIS.
Date
5-20-76
8-18-76
9- 9-76
4-25-75
5- 1-75
9- 2-75
5- 9-74
5-31-75
9-23-74
11-18-74
12-11-74
5- 9-73 '
2-22-72
I J.-7A-T).
: 4-17-72
!j 7-17-72 , •
j 4-29-71
4-30-71:
6-15-71
'• 7-12-71 '
8-31-71 |
2- 9-70 •
— : 5-14-79-
I 5-15-70
6-19-70 '.
7- 8-70
8-21-70
8-25-70
11-30-70
Sample '.
No. |
52 , ! '
140 J •
146 !
151
139
302
139 i :
140
302
302
' 42 : i i.
33 i '< ;
52 ! • !'
139 ; ;
139 P. ;
148 ;
145
43 .
302
58
;56 '.'
148 !
302
'61 i-. f
150 i ' !
150 i ; !
146 •' ;
62
146
C12
Res.
.80
.36
.12
1.09
.92
.76
.88
.40
..76
.72
1.00
1.04
1.04
=80
.64
1.08
.28
.40
.64
.08
.04
.76
n j
.20
.72
.84
0.00
.12
.40
SAMPLE
TPC1 GF2
35 5
11 3
5,400 ••'•' 4
1 : 3
110 5
1 i 3
240 ' 1 5
70 ' 4
1 i 5
1 4
115 j 5
2 ,
6
19
650 :
1
1 ;
10 ';
1 •
5,400
55 ':
i .';
1 ,.
55
55 C
4
•3
5
5
3
5
: 5
5
4
3
5
1 3
5(
... . '5
4 :i j 3
14,000 ' 4
6 ! 4
1 4
COLI
5
3
4
3
5
3
5
3
4
4
5
4
3
5
5
3
5
5
5
4
3
5
3
3
5
3
4
4
4
CHECK SAMPLE
TPC1
0
1,200
85
2
1
0
1
1
1
1
1
1
5
5
1
1
1
1
1
11,000
600
1
1
18
N.A.
•N.A.
45
1
GF2
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0~,"
1
0
2
b
0
0
0
COLI
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
••—'"•" /\ " •
u
0
1
0
0
0
0
§
a
g
. &
§
H
a
o
5?
o
o
1
h3
M
tn
t/1
p
J-1
s
1. Total Plate Count
2. Gas Formation
O
-------
> - ^, :. t - , s - i ^*-* • ----.
TABLE 7. DEAD END SAMPLE COMPARISON WITH TIME.
IWC LAB (2 HR.)
EPA LABORATORY
Date
6-24-75
7-16-75
No.
1
2
3
4
5
6
7
8
9
Gl, Res.
0.00
.54
.20
.12
.12
.12
0.00
0.00
.12
TPC*
4,900
1
0
2
0
1
44
300
300
GF
0
0
0
0
0
0
0
1
0
2 COLI
0
0
0
0
0
0
0
1 '
0
TPC
110,000
540
730
650
240
480
21,000
760
30,000
GF
5
0
0
3
0
0
2
0
0
COLI
3
0
0
2
0
0
2
0
0
1. Total Plate -Count
2. Gas Formation
TABLE 8.
- Temp-F
_Turb.-Raw
MPN-Raw
Algae-Raw
Alum-$/MG
CMor-#/MG
Turb.-Sett
Algae-Sett
Turb.-PE
Algae-PE
t Coli-D.S.
:TPC-D.S.
:TPC>500
Jan.
38
46
49,231
1,252
224
50
1.5
212
.36
25
4
12
1
Feb.
38
90
38,383
934
260
43
1,6
139
.35
14
1
9
2
Mar.
43
55
17,504
1,485
133
44
1.9
62
.23
9
0
3
0
Apr.
51
85
23,488
5,779
193
46
1.5
450
.26
25
7
11
3
WHITE
May
66
90
6,415
13,167
215
60
1.4
1,232
.34
94
9
28
4
RIVER SUPPLY— 1975.
June
73
105
3,151
7,632
182
64
1.7
551
.25
54
0
3
0
July
77
90
5,709
10,953
249
81
1.9
1,497
.57
165
0
180
13-
Aug.
79
70
7,010
12,431
403
94
2.1
1,744
.85
158
1
450
39
Sept.
68
45
2,451
15,503
185
72
1.9
1,463
.37
99
4
90
16
Oct.
58
35
2,579
7,358
116
60
1.5
1,274
.30
90
0
31
7
Nov.
52
25
5,018
3,982
114
46
1.6
272
.26
23
o
65
4
MF §
03
0 §
0 CO
•H
0 &
u o
0 >
0 o
0 n
0 >
|
|
*; ,33-
Dec. §
40 rl
30-— &
57,085
3,263
172
59
2.2
328
.35
26,
4' :
0
-------
IMPLEMENTATION: COMMUNITIES/R.J. BECKER
also shown on each check sample. As you examine this information, it be-
comes apparent that regardless of your theory or beliefs there exists some
exceptions. My own interpretation is that a majority of these unsatisfactory
results are due to mishandling either during sample collection or sample
preparation. This impression primarily centers around the relatively high.
chlorine residuals occurring in some of the samples at the time of collection,
and the conviction that this range of residual is not conducive to significant
bacterial growths. Likewise the generally satisfactory results recovered in the
check samples.
During the summer of 1975 an EPA, Cincinnati Laboratory team was at
our utility making some virus studies. The bacteriological laboratory back in
Cincinnati requested that we collect several dead end water samples, and re-
turn them to Cincinnati with the virus team. Two sets of samples were col-
lected from nine separate locations. Each sample was collected from the
nozzle flow from a dry barrel fire hydrant as soon as it was anticipated that
the branch line had been displaced. This is obviously an undesirable sample
point from a sanitary standpoint. Sample preparation in our lab was within
two hours. I have no definite knowledge of the time interval prior to prepa-'
ration in the EPA lab, but it could have been close to 24 hours after col-
lection. Table 7 illustrates the results from both labs with the primary variable
being time. This is still another obstacle which a small utility faces in getting
a water sample to a central laboratory for analysis, and according to some
authorities presents a major handicap in obtaining representative quality
results.
'
Table 8 is a brief monthly summary looking af a number of raw water
parameters which may have an effect on finished water and distribution sys-
tem water quality. The year of 1975 was not a particularly bad year relative
to water quality, :butit does rillustr ate one previously unmentioned parameter
which reportedly has a significant impact upon bacteriological quality. The
parameter that I would like to focus on ,is that of turbidity. Turbidity is one
of the most nebulous for quantitative comparison. Turbidity comes in all
sizes, shapes, and charges. Although cold water temperatures do interfere
with coagulation, silt particles accompanying soil runoff, are normally relaT
lively easy to remove. However the smaller the particle size, the more diffi-
cult it is to remove. This is demonstrated by comparing the algae counts
for raw, settled, and filter effluents. For our supply the most normally pre-
dominant organisms are diatoms such as Cyclotella, Synedra, and Navicula.
As you can see from the data presented we fairly consistently remove 88%
of the algae by coagulation and settling. The datar also indicates that 92% of
the remainder are removed:by filtration. Raw water algae counts as high as
50,000 to 100,000 organisms per ml have been recorded in the past. Our
record over several years on raw, settled, and , filter effluent algae samples,
indicates the percent removal statistics hold fairly uniform. Naturally, the
greater the raw water concentration the higher the concentration will be in
the finished water resulting in higher turbidity readings. If you will look closely
at the month of August, you will. see it w.as the month with the highest
effluent turbidity even with a twice average alum dosage for about an average
raw water turbidity. In the statement of Basis and Purpose for the Proposed
National Interim Primary Drinking Water Standards it states "A properly
operated water treatment plant employing coagulants and granular filtration
should have no difficulty in consistently producing a finished water conforming
to this limit." We are one utility that can meet the limitation of one (1) Tur-
bidity Unit on an annual basis, but from our historical data we know we will
179
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
exceed that limitation on a monthly basis under certain raw water quality
conditions.
:o confuse the issue by the presentation of
to the subject of bacteriological quality.
2X one requiring a knowledgeable interpreta-
It has not been my intention
limited operating data pertinent
Rather that the subject is -a compl
tion of perhaps misleading results. The Indianapolic Water Company histori-
cally has met and expects to contmue to meet the proposed coliform standard
in the Primary Drinking Water Regulations. Please remember that PL 93-523
is known as the "Safe Drinking Water Act" and not the "Pure Drinking
Water Act."
The most distasteful portion of the Act to me and many other water
utility operators is that of public notification. No responsible operator would
hesitate to forewarn their consumers when a known hazard exists, but to do
so based upon an erroneous test does nothing but discredit the utility or the
industry. It is my personal observ ition that public notification should be made
only after a careful review of dai:a indicating an MCL has been exceeded by
a knowledgeable public health authority, and that a hazard does or did exist.
Another area of the Act that [ question is the responsibility of the utility
to maintain water quality to the tap. This really is beyond the limits of any
control by the utility, and actually falls on private property. The water utility
does not establish plumbing codei, approve building plans, or perform build-
ing inspections during or following construction. Likewise it has absolutely
no control over the remodeling or repair of existing plumbing within the
consumers property. We know from first hand experience that these actions
can result in unsatisfactory and anrepresentative bacteriological quality.
From the standpoint of elapse :d time and our operating data the second
daily check sample would appear superfluous. In the event of plumbing repair
contamination introduced from within the sample structure, samples col-
lected from adjacent sites may prove to be much more representative of actual
water quality in the system. If a bad sample is recovered in a small system
late in the month, there is no reasonable way an operator can collect enough
samples to improve his monthly average! If the bad sample occurs early in
the month, he may have a change. If there is a suspected cause of sample
contamination in the original sample, it makes more sense to me to disregard
it in the record and retain the results of a satisfactory check sample. Rather
than retain the results of an unrspresentative sample, and not even consider
the results of the check sample. I realize this all boils down to the numbers
game in trying to meet a hard and fast regulation. I think it would be well to
point out that most responsible w ater utilities operate according to well docu-
mented and proven practices ani principles evaluated to fit their own par-
ticular situation. As long as all these systems perform properly, satisfactory
bacteriological results will generslly follow. However, if a fault occurs in the
system and unsatisfactory bacteriological quality results, no number of addi-
tional samples will correct the fault.
I sincerely appreciate the opportunity to express my obviously prejudiced
views on this subject to this group today. I am equally grateful that it is not
my responsibility to establish bacteriological quality standards to protect the
public's health under all the known variable conditions that exist throughout
the nation.
180
-------
IMPLEMENTATION: STATE/HJ. ONGERTH
Compliance with PL 93-523 Coliform Standards
Henry J. Ongeith
California Department of Health
Berkeley, California 94704 :
Many people have taken the coliform standards for granted. The advent
of PL 93-523 and its mandatory compliance! requirements have now begun
to raise questions in the minds of some about problems of attainability and
compliance with these standards. This presentation will briefly review the
development of certain aspects of the coliform standards and discuss the
program of compliance in California. j
In 1925 Treasury Department standards for drinking water were revised
and the coliform standards as we know them] were first presented. The mem-
bership of the advisory committee reads like a Who's Who of Sanitary
Engineering including: George C. Whipple C. A. Emerson, George W.
Fuller, Charles G. Hyde, Edwin O. Jordan, Harry E. Jordan, Milton J.
Rosenau, Robert S. Weston, C.-E. A. Winslqw, and Abel Wolman—a most
impressive group.
The report of this committee is brief and exceedinly interesting, and is
presented in full as follows:
"The task referred to this committee by the Surgeon General of
the Public Health Service is to forrmlate definite specifications
which may be used by the Public Health Service in the administra-
tive action which it is required to take upon the supplies of drink-
ing water offered by common carriers
for the use of passengers
carried in interstate traffic. The recommendations submitted apply,
therefore, only to this special case, and
general application. :
"Since the purpose of the supervisior
ire not"proposed for more
.which the Public Health
Service exercises over these water supplies is to safeguard the health
of the public, the examinations and specific requirements herein
proposed have reference chiefly to forming a judgment of safety,
and are designed especially to afford protection against the most
serious danger which is associated with water supplies, namely, that
of infection with typhoid fever and other diseases of similar origin
and transmission. Less emphasis has been placed upon physical arid
chemical characteristics affecting the acceptability of water with
respect of appearance, taste, and odor, because these are matters
of less fundamental importance and becpuse, in actual experience,
the water supplies which come under consideration, if satisfactory
from the standpoint of safety, will usually be found satisfactory
with respect to physical and chemical characteristics.
"The first step toward the establishment of standards which will
insure the safety of water supplies conforming to themris--to agree
upon some criterion of safety. This is necessary because "safety"
•_ j. —....——I!—... ~* „ *!• AW f, Mn f\ A't-Tinl f i* f*t*r**3ii /•»£»/-! tr« t*cfc1»s-friTrja
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
safety of a water supply, it is well established that the water supplies
of many of our large cities are safe in the sense stated above, since
the large populations using them continuously have, in recent years,
suffered only a minimal incidence of typhoid fever and other poten-
tially water-borne infections. Whether or not these water supplies
have had any part whatsoever in the conveyance of such infections
during the period referred to is a question that can not be answered
will full certainty; but the total incidence of the diseases has been
so low that even though the water supplies be charged with responsi-
bility for the maximum share which may reasonably be suggested,
the risk of infection through them is still very small compared to
the ordinary hazards of everyday life.1
"The committee has, therefore, taken this better class of municipal
water supplies as its standard of comparison with respect to safety
and proposes, as a fair objective, that the water supplies furnished
by common carriers to passengers in interstate traffic be of com-
parable safety. As regards protection of the traveling public, such a
standard is fair, since it implies that the use of the water supplied
to them in travel shall not add to the almost negligible risk which
is ordinarily incurred at home by those who habitually use water
supplies of somewhat better than average quality. From the stand-
point of the carriers also, this standard is believed to be fair and
reasonable,' since it refers to water supplies which .are actually
obtainable in all sections of the country and from a great variety
of sources,
'The next and principal task of the committee has been to set
up objective requirements which will conform to this general
standard of safety; that is, requirements which will ordinarily be
fulfilled by the municipal supplies of epidemiologically demon-
strated safety which constitute the standard of comparison, but
will exclude supplies of less assured safety. Since there is no single
and measurable characteristic of water supplies which bears any
known and constant relation to actual safety, the standard recom-
mended is composite, including certain requirements relative to the
source and protection of the water supplies in question as indicated
by a careful sanitary survey, and certain other requirements relative
to bacterial content as shown by standard tests.
"It is anticipated that little objection will be raised to the re-
quirements laid down as to source and protection, at least to then-
general intent, because they are based upon well recognized prin-
ciples of sanitary engineering, and because they are necessarily
stated in general terms which imply a rather broad consideration
of each supply from all angles and the exercise of discretion in
forming an ultimate judgment of its fitness. The bacteriological
standard, on the other hand, is stated in. definite quantitative terms.
This is unavoidable if such a standard be included at all, since the
methods of bacteriological examinations are quantitative and yield
1This evidence actually proves only that the water supplies in question have
been generally "safe" in the past during the period of low prevalence of infection.
The likelihbcxi that they will continue to be equally or more safe .in the future must,
of course, be reckoned from "other considerations, such as the probability of
future change in the pollution of their watershed, the character and consistency of
their protection, etc. ••
182
-------
IMPLEMENTATION: STATE/H.J. ONGERTH
. results in the definite terms used in the standard. However, in view
of the well-recognized principle that the significance of bacteriologi-
cal examinations is variable, and must be interpreted with due
regard to all other facts known about the particular water supply
in question, the objection may be raised that a rigid application of
this standard will arbitrarily exclude a considerable number of
water supplies which conform to all other requirements and which
competent opinion will consider to be quite safe. The validity of
this criticism is recognized, but it is not considered of sufficient
force to require or justify the lowering of the bacteriological stand-
ard proposed. This viewpoint appears proper when it is recognized
that the definite terms of bacteriological quality in which this
standard is expressed represent only agreement as to safety, and
not as to limiting values beyond which demonstrable or even pre-
sumptive danger lies. Between the point on which the committee
is in agreement as to the assured safety of water supplies and the
point at which agreement could be reached as to their dangerous
quality is a wide zone. Within this zone lie many water supplies
which, if considered in the light of available evidence from all
angles, are believed to be as safe as other supplies which conform
to all the bacteriological requirements.
"The committee, therefore, considers it preferable to recommend
that in actual practice the bacteriological standard be applied, as
are other requirements, with some latitude; in other words, that
supplies which, on rigid inspection are found to be satisfactory
" in other respects but fail to meet the bacteriological standard,
may be accepted in the discretion of the certifying authority. In view
of the character of the personnel entrusted with the responsibility for
;-:-.investigation arid administr:ative action; the committee feels assured
this procedure is preferable to the alternative of rigid and auto-
matic application."
In one of the appendices of the report, the committee's philosophy on
attainability of the standards is presented as follows:
"As to the reasons for specifying these particular limiting values
rather than some other values, either higher or lower, it is obvious
that the assignment of any definite limits of bacterial content as a
criterion of the safety of water supplies of diverse origin and history
must necessarily be an arbitrary procedure, because the relation
which the ^terminable bacterial content bears to the actual safety
....'_ of._a._water_.supply_is^ variable and to some extent indeterminate.
Therefore, all that may be claimed for the standards proposed is
that, hi the judgment of this committee, they are reasonable; that is,
are consistent with the other requirements specified as to source and
protection of the water supplies in question, afford an ample guar-
antee of safety, and.can be met without too costly and burdensome
effort. In connection with this hist consideration, the committee
has analyzed the records of daily ^examinations of a considerable
number of municipal water supplies for the years 1919: to 1922
and finds .that!.a substantial majority of,them conform to both re-
quirements of the standard."
Whether or not, these statements reflected general professional opinion
on the subject or whether they molded! professional opinion cannot be said,
but most certainly as I saw the standards applied in the decade of the 30's
183
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
and on, practices in application of the standards were consistent with these
views.
The drinking water standards, from the very beginning, were adopted
under tlie authority of the Interstate Quarantine Regulations and were the
standards for the water to be used on common carriers, railroads, vessels,
etc.; carrying passengers iri interstate traffic. In addition, these coliform
standards (of course, the organisms were called B. coli in the 20's and early
30's) were generally recognized as the nation's drinking water standards and
applied by the states. I do not know the details of the history of the use of
these standards state-by-state. I believe that gradually more states formally
adopted these standards, as their own, by reference. As to compliance and
enforcement, undoubtedly this was a mixed bag around the country. First,
speaking just about compliance for waters used on interstate carriers: Until
1942, sampling for compliance with the standards was done partly at the
source and partly hi distribution systems. Unquestionably, practices varied
from place to place. It is likely that, as with places studied for the 1925
standards, most of the sampling was done at the point of production of the
water, at treatment facilities or at ground water sources. It is significant that
as to compliance, water systems will more readily comply if sampling is done
at sources than if sampling is done in distribution systems.
The 1925 standards contain no directions or limitations with relation to
sampling location. It is likely that the practices at the time were to sample
at the point of water production and to a lesser degree in distribution sysems.
The 1942 standards, in recognition of the Chicago amoebic dysentery out-
break attributed to cross connections, required that the bacteriological exami-
nation of water be of samples collected throughout the distribution system.
In other words, source samples could not be utilized in considering com-
pliance. Thus, even though the coliform numbers for the standards were
unchanged from the 1925 standards, in fact these standards were more
difficult to attain.
At this tune compliance was based on annual results—a summary of all
samples collected during the reporting year. Sometime in the late 1940's the
Public Health Service administratively changed the ground rules and there-
upon required compliance on a month-by-month basis. This also increased
the difficulty of compliance.
The remainder of this discussion concerns the conduct of the coliform
fnonitoring-compliance program in California through the years. The Bureau
of Sanitary Engineering was created in 1915. At that time little water was
receiving treatment. The city of Sacramento obtained water without treat-
ment from the polluted Sacramento River. According to the July 1915
monthly bulletin of the California State Board of Health, Sacramento had
a death rate from typhoid averaging about 53^ per 100,000 annually for
each five year period since 1895 although some of these were "imported"
cases. The case rate, as contrasted to the death rate, was about ten times
as high. The typhoid death rate for the entire State was 13.6 per 100,000
population. The program emphasis for the Bureau of Sanitary Engineering
for the next 15 years was to get sewage out of water supply sources, to get
adequate treatment for surface water sources, and to eliminate hazardous
wells. The first statewide comprehensive sanitary engineering survey of
community water systems was made in 1930. Beginning in the mid-30's with
the advent of WPA and PWA projects, program emphasis was transferred
to development of sewage treatment facilities to replace raw sewage dis-
charges. When I came to work in the State Health Department in 1938, the
water supply program was focused on (1) dealing with "problem" systems
184
-------
MPLEMENTATION: STATE/H.J. ONGERTH
to get correction of major defects, and (2) annual inspections of the interstate
carrier systems. I estimate that at that time the Department had probably
four or five hundred systems under surveillance.
Efforts were made to bring systems into compliance with the coliform
standards when they were not met, but. the major emphasis was on improving
physical works and operating procedures that were likely to result in, or had
the potential for, production of poor quality water. Compliance with the
bacteriological standards was (and still is) a secondary matter. Public health
is founded on the solid rock of prevention. Our philosophy is that the safety
of public water supply is assured through attention to the development of
satisfactory facilities and existence of capable management and operator per-
sonnel. Sampling and analysis is only one part of the surveillance process and
provides after-the-fact information, analogous to watching the scenery from
the rear observation platform of a train. In 1938 a few utilities—the larger
muncipally owned ones, such as Sacramento, San Francisco, East Bay MUD,
Los Angeles, and San Diego—and 2 or 3 major investor-owned utilities such
as California Water Service, carried out routine bacteriological monitoring
programs with samples analyzed in their own laboratories. For the rest of
the systems, the State Health Department did infrequent sampling. Samples
were collected at least once a year at the interstate carrier systems, which
numbered 100 or so. For the rest, water systems were visited at least on a
decennial basis, and more often where the problems were known to exist. At
the time of the decennial visit, a thorough investigation was made of the
physical works, facilities operation, and relevant operating records to the
extent that any existed. Bacteriological and chemical sampling also was done.
The water at some of the places met the coliform standards, and at other
places, it did not.
, .As the years went on, and = additional manpower became available—and
this was always a hard struggle—the program was tightened up. Gradually
a policy evolved of requiring all water utilities to collect water samples in
compliance with the Drinking Water Standards frequency requirements, and
to have those samples analyzed in "approved" laboratories. Beginning in
the early 60's, we initiated an aggressive program of requiring compliance
with the frequency standards. As more; places complied, we began collecting
and organizing statewide . statistics with relation to compliance with the
bacteriological standards. As of about 1968 approximately 30% of the
water systems under our jurisdiction—1100 with 200 or more service con-
nections—failed to meet the coliform standards one or more months out of
the year: Perhaps ~a third of those-not in compliance failed multiple months,
some failing as much as 6 or~more months of the year. For the past 8 years,
we have worked very aggressively to get compliance, following up month-by-
month with those places which failed to meet the coliform standards the
previous month. Now, after about 8 years of quite energetic effort, though
with no efforts at legal action, the statistics for 1976 show about 8.9% (97)
of approximately 1100 systems failing to meet the coliform standards one or
more months. Last year 2 places failed to meet the standards for 5 months;
2 failed for 3 months; 14 failed 2 months; the remainder (79) for only 1
month. These statistics are low since compliance has been calculated from
running series of the last 20 samples.
As I have indicated, the "state" program deals with water utilities with
200 or more service connections. Local health departments deal with the
less than 200 service connection systems, and the current inventory records
about 4400 community water systems under local health department juris-
diction. These systems are only partly under routine bacteriological sur-
185
-------
']! "•''» III1 '"I1 In VMrnii Illlllltl 1'i il'llll^^ llllllillllp1 "' i! '» IT I 'II1 'I'' ', 'j'l , i .nil .iiilllllpllliinllili „ ': ' ,,i'i| rl Ti1,,," Ill , 'illil",!'v ,i ' iiil'l" i|i|i|ll!ll""i" IH.11111"1 IT'1; ,,'!' "IIII I' J""i1 Tl'iiii!" • '', ' ' '' :l '"'i'llll'j', IK1,11» I1,'1!!'! 1'l,''ll,ll!lll!ili!ll'
I . .
" , , , '. . , I ' . Ill'1
!! ' i" ,
11 ; , ' • ' IF , I ., I , :
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
veillance, and statistics are available for calendar 1976 for 2136 of these
systems. The results show that 32 systems failed the coliform standards 6
or more months; 209 systems failed 2-5 months; and 162 systems failed 1
month. In summary, 403 systems failed 1 or more months, amounting to
approximately 981 system/month failures. This would have produced that
many public notifications of failure to meet coliform standards. As noted
above, these statistics are on the low side because compliance was cal-
culated from running series of the last 20 samples. In addition, it is likely
that these statistics in part represent the better small water systems. If all
systems were under bacteriological surveillance, the proportion of failure
to comply would be higher. Most certainly there is a big problem of bringing
these systems into compliance with the standards.
A final comment—and an important one—what are the causes for failure?
Of course, they are varied. Almost no failures occur at the point of produc-
tion at water filtration facilities. Some failures are traced to malfunctioning
of cUormation facilities at small places with surface water sources and no
filtration facilities. A few failures are traced to coliform "contamination"
of wells. (We will not accept chlorination of coliform-producing well water
as a method of compliance. Wells must produce coliform-free water or be
cut off the system.) Other causes for "high" coliform are storage facility
defects; for example, birds gaining entry, work on distribution lines (main
breakers, etc.)—coupled with unsuccessful main disinfection—and, finally,
the elusive "sampling error." ..--..
Follow-up samples are required at all places failing to meet the standards.
In a substantial fraction of places, the follow-up samples are free from coli-
form, and the original positives are attributed.to "sampling error"—or to some
transitory condition of water quality. If the cause of the positive sample
is a real thing, further positives may result—though sometimes sporadically.
As to cross connections producing coliform contamination, few if any cross
connections have been discovered through bacteriological sampling. This is
not to say that cross connections have not resulted in contamination of water
systems—-the reverse is our experience.
Since 1970 California statutes have contained a public notification section
as follows:
"When it is determined by the department that it is in the public
interest, -the department shall notify a person, who has not met the
standards of the department for water quality of such noncom-
pliance, and shall require such person to notify in writing each of
his customers of the department's determination that the quality
pf tfte water fails to comply with the standards, requirements,
' or conditions established by the department arid to include any
comments of the department regarding the possible dangers because
of such noncompiiance. The content of such statement shall be
approved by the department prior to distribution. Notification by
such person shall be repeated at intervals as required by the de-
partment or until the department determines that there is compliance
with the standards or requirements of the department."
In 1973 the Legislature added the following with relation to public
notification:
"Upon determination by any person who furnishes or supplies
water to a user for domestic purposes that a significant rise in the
bacterial count of water has occurred in water which is being sup-
plied to a user for domestic purposes within the State of California,
the person who furnishes or supplies such water shall provide, at
;„; -:., ;;;, . ,;;:; , , • • j '
186
-------
IMPLEMENTATION: STATE/H.J. ONGERTH
such person's expense, an analysis of such water within 24 hours
to the county health officer and the State Department of Health.
As used in this subdivision 'significant rise in the bacterial count
of water' means such rise in the bacterial count of water as the de-
partment determines, by regulation, represents an immediate danger
to the health of the water users."
Regulations adopted pursuant to this section define significant rise as
follows:
"(a) Significant rise in bacterial count representing an immediate
danger to the health of water users is defined as an increase in
coliform bacteria, when associated with a suspected waterborne
illness or disruption of physical works or operating procedures.
"(b) The following criteria shall be used as an indication of a
significant rise in bacterial count, and is applicable to routine bac-
teriological samples . . .:
"(1) Whenever all five portions of more than one sample
collected during a week show the presence of coliform or-
ganisms; or
"(2) Whenever 50 percent or more of the portions of a
group of two or more samples collected during a day show the
presence of coliform organism; or
"(3) Whenever the daily samples, collected following un-
. •— satisfactory: samples . . ., show the presence of coliform
organisms in three or more portions in each sample if more
; than one sample is taken, or in all portions if only a single
sample is taken."
Upon a finding of significant rise, notification of both the Department
of'Heaith and the water users is required, as follows:-
"(a) When the coliform "limits specified above are reached or
exceeded, the water utility shall immediately notify the department
by telephone of the bacteriological findings and, at the same tune,
shall furnish information on the current status of physical works
and operating procedures which may have caused the elevated
bacteriological findings, or any information on community illness
suspected of being waterborne.
"(b) Upon receipt of informatiion specified in (a), the depart-
ment, taking cognizance of all available information shall deter-
mine whether or not continued delivery of water would constitute
anTimminent danger "Ma the health of the users.
--"--:----*'(1)-Upon- finding an imminent danger, the department
shall immediately notify the water utility of this finding and
of the necessity to implement the emergency notification
plans ..."
_ These sections have recently been, modified where necessary to bring Cali-
fornia statutes into compliance with PL-93-523.
In conclusion, an opinion on the attainability of the current bacteriological
standards: These standards are stringent—only very minor defects can
produce failures to meet.the., standards-—but they are attainable when every-
thing is in good order. Unless there is universal chlorination with chlorine
residuals carried to the end of all distribution systems, a fraction of water
systems will always fail to meet the standards. I consider that this eventuality
does not represent a "public health problem of any significance. The problem
with compliance is that with current waterworks practices, a great deal of
187
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
regulatory effort is necessary in order to get a reasonable degree of com-
pliance. Meanwhile, I am sure some states have a long way to go.
;t
188
-------
CANADIAN DRINKING WATER STANDARDS/T.P. S1JBRAHMANYAN
Canadian Drinking Water Standards
T. P. Subrahmanyan, D. A. ScMemann, and A. S. Rhodes
Laboratory Services Branch
Ontario Ministry of Health
Toronto, Ontario, Canada
3. Henderson
Canadian Public Health Association
Ottawa, Ontario, Canada
E. Bowmer
Division of Laboratories
British Columbia Department of Health
Vancouver, British Columbia, Canada
K. R. Rozee
Department of Microbiology
Dalhousie University
Halifax, Nova Scotia, Canada
P. Payment
Institut Armand Frappier
Cite de Laval
. Quebec, Canada
^desirability of having uniform drinking water standards throughout
Canada has long been recognized. In early attempts to achieve this, the
considerable amount of data and expertise available in the United States
was utilized. This was done mainly through cooperation between national
agencies and by consultation with many scientists active in the relevant
fields. Our present attempt to coordinate U.S. and Canadian efforts follows
this tradition.
We do not claim to represent the opinions of every agency involved in
the regulation and enforcement of drinking water standards in Canada.
However, as members of a major committee (formed by the Canadian Public
Health Association) involved in the revision of Canadian Drinking Water
Standards,-we-hope-to present a-Canadian consensus arrived at by this
committee : in its presentation to the Department of-National Health and
Welfare, Canada.
The history .of drinking water standards in Canada dates back to 1923,
when the Federal Cabinet passed an order-in-council which provided a
standard of bacteriological quality applicable to water used for drinking and
culinary purposes in vessels using Canadian inland waters including the
Great Lakes.-In-1930rthese-regulations v?ere extended to all types of common
carriers crossing international borders into Canada and inter-provincial
borders within Canada. Canadian coastal shipping was included in 1937.
, .Drinking water standards became more generally applicable when the
Department of. National Health and Welfare, with the approval of the
Dominion Council of Health, began to apply the 1943 United States Public
Health Seryicet.JStandards, pending an examination of the.situation in Canada.
In the, interim, Canadian authorities referred to the U.S. standards of 1943
and 1962 as well as those published by the World Health Organization in
189
-------
,11 '!' i; ,
li! i, t'i!1
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
1958 and 1963. The Canadian Public Health Association (CPHA) also
recognized the need for standards suitable for Canadian conditions which are
not always the same as those in. the United States and other parts of the
developed world. It formed a Drinking Water Committee in 1966 to develop
appropriate guidelines. A year later, the Department of National Health and
Welfare convened an Advisory Committee on Public Health Engineering.
The task of drawing up drinking water standards was delegated to a sub-
committee which decided to combine forces with the earlier group formed
by CPHA With the able assistance of Mr. S. K. Rrishnaswami, who was
appointed principal investigator, the joint committee prepared preliminary
reports and drafts for desirable standards.
Three levels of drinking water quality were designated on the basis ot tne
premise that water for domestic use should be free from pathogenic or-
ganisms and their indicators as well as from other deleterious substances
besides being palatable and free from objectionable odours, tastes and
colours. The poorest of these quality levels was called 'maximum per-
missible" the next "acceptable" and the most desirable level, which should
be the goal was called "objective". It was emphasized that the poorest quality
should not be permitted for any length of time. On the basis of the detailed
recommendations of the committee and taking into account other national
and international standards, the 1968 Canadian Drinking Water Standards
and Objectives was drawn up and published by the Department of National
Health and Welfare. .„
The original committee emphasized the need for continuous surveillance
of changes in water quality and for studies to develop basic information on
Canadian conditions and requirements. In 1975, the Department of National
Health and Welfare established a Federal-Provincial working group for re-
vising the Canadian Drinking Water Standards, a task which is expected to
be completed by 1978." As part of this process, a number of submissions
have been made including the report by the CPHA advisory committee of
which the authors are members. One of the authors (J. H.) prepared a draft
for a Criterion Document entitled "Microbiological Quality of Drinking
Water" based on current knowledge drawn from national and international
sources. The criteria document with recommendations was finalized and
Submitted to the Department of National Health and Welfare.
Regulation and enforcement of water quality standards in Canada are
traditionally vested in provincial agencies with the territories in the north
falling under federal jurisdiction. Differences can and do exist between
provinces. The situation in Ontario, for example, is complex. The Ontario
Ministry of the Environment (formerly Ontario Water Resources Commis-
sion) is responsible for the overall control of regulation and enforcement of
water quality in public supplies. They work in cooperation with the Min-
istry of Health, especially when potential health problems are encountered.
Both Ministries maintain large water microbiology laboratories centrally in
Toronto and smaller ones in various regions of the province. To avoid
duplication of efforts, the Ministry of the Environment laboratories limit
their activities to public piped systems which are under their control. The
Ministry of Health laboratories deal with specimens collected by local and
provincial health agencies. In addition, major municipalities such as Toronto
have their own testing facilities and in these cases, the Ministries play con-
sultative roles. .
Similar jurisdictional delineations with some unavoidable overlaps exist
in other parts of the country. The regulatory agencies at provincial and
municipal levels follow the current Canadian Drinking Water Standards with
1 I9° """ .! '',', '
iii,;i'Jikiiil ili.'4iii'i!!,,f'.i Jl'illnld n; •;.! •./1,.., i,.',i. .viii'ililJuiK..\,„,. _,,,L.... ..ViL'.-_I'«\L_
|.ilji!B||ihhijlli! lLhiiillJii.1,1 .ilL,.,,, IIIIIIIBIill iiliii iJiiii....,/,,, ' •, iliii;!, iiiillliLililllllllll" I. ,1 Jliniilillllllllili., ,!!• •, i III jil 4,1 Jlllllill II .«• ill liliillliililiiil n, n If1!. I,. „, ;l L,, ^ L _. .4. ^i1' k. jyjiiu :«. i ..* .J'l,'.!._..!_,.. iL ;..,' j, ,_,*_,_ ^l _ju jlJUi, iMiill, • Ji.iiilniii.i n I 'i .MI .' • 'i, Jllli'! &. •iiil>;! i'JmlJiil Ml Mitilli idd «: •;.! ' /1, n ^ ivHliI i.il»i i ,„... _,, iL.... ..VillL'.- < '<> i.i. J •• '.I. > iHO® •• 'ii i1' '"4 r. iil'.. .•.!; u!i il.1 jil
-------
CANADIAN DRINKING WATER STANDARDS/T.P. SUBRAHMANYAN
modifications designed to suit perceived local needs and to improve the
service. Some, such as the Ontario Ministry of the Environment, have
developed their own objectives, which conform to the national standards but
specify details of improved sample collection or additional examinations.
Thus the results of the attempts to attain a measure of uniformity are
encouraging.
The 1968 Canadian Drinking Water Standards specify that methods given
in the current edition of "Standard Methods for the Examination of Water
and Wastewater (APHA)" should be employed in the routine examination
of water. It is also recommended that siampling frequencies should conform
to the U.S. Public Health Services recommendations of 1962 with the proviso
that frequency and location of sampling be established by the control agency
or Medical Officer of Health after adequate investigation of the source,
method of treatment and protection of the water supply. Raw waterjs con-
of the samples STany consecuUve
30-day B^nod^have^je^l^oJiforja^ejasjities of less than IQOjref 100 ml and
an^aTcolifo~ ~
^^
colifonn~density of IRSS than 1 nn(riper~TUu ml. although the organisms
relationship.
•"FmisHeddrinking water is considered "acceptable" if 95% or more of the
samples tested in a 30-day period are negative for total coliform organisms
when enumerated as a Most Probable Number (MPN) index by the multiple
tube fermentation method. None of the positive samples should have MPN
indices greater than 4 per 100 ml. Waters with MPN indices of 4 to 10 per
100 ml in more than three consecutive examinations should be immediately
investigated in respect of source, treatment and other parameters. According
to the 1968 standards, the membrane filter (MF) coliform test should be
used to determine the potability of a given water supply only after adequate
parallel testing has demonstrated that the MPN and MF methods yield
equivalent information-relative to the sanitary quality of the water supply
system. The MF method should be carried out on a minimum sample
volume of 200 ml. Drinking water is considered "acceptable" if at least 95%
of the samples hi a 30-day period are negative and none of the positive
samples show1 MF counts greater than 4 per 200 ml or 10 per 500 ml.
Any water with MF counts greater than 6 per 200 ml or 15 per 500 ml
or approaching these limits in three consecutive tests should be immediately
investigated. A special series of samples and the commencement of suitable
and effective remedial measures must follow any deviation from 'MPN or
MF standards. In such an event, all samples must be carried to the "com-
pleted" test" procedures. In addition, the demonstration of any pathogenic
organism in finished water must lead to the rejection of the supply until effec-
tive remedial treatment and disinfection have been completed. The Medical
Officer of Health or the control agency may issue an order requiring boiling
of water or other corrective action in appropriate situations.
There is no published information on whether water quality standards are
in fact being met. However, we were able to draw upon the experience of
Mr. L. Vlassoff, Chief, Microbiology Section, Ontario Ministry of the En-
vironment. In his experience of many years involving over 500 water treatment
plants, enforcement 0* standards has mot presented serious problems. In
these examinations, a simpler sampling system is used. A minimum of 8
samples per month are collected at weekly intervals from small communi-
ties but for large centres, the number of samples is at least the same as in the
1968 standards. MPN, MF or Presence-Absence tests are carried out on 100
ml samples. About 5% of the specimens require detailed testing, which
includes tests for several individual bacteria. As a result of the good rapport
191
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
established between microbiologists and engineers managing public water
systems in Ontario, there is good cooperation in the enforcement of stand-
ards and in taking rapid remedial action as well as in carrying out additional
examinations when indicated. The Drinking Water Standards specified for
public supplies may not be applicable to other water supplies which'are also
likely to have enforcement problems. In revising standards, this aspect should
*^££±±-«f the recent CPHA committee on "Microbiologi-
cal Quality of Drinking water" are summarized below and the more im-
portant sections compared with the corresponding secttons in the 1975 V*.
National Interim Primary Drinking Water Regulations. The detailed recom-
mendations of the CPHA Committee are available from the Canadian De-
partment of National Health and Welfare.*
1. Since modern technology is capable of purifying water even from
heavily polluted sources, it was felt that the major concerns should
be adequacy of treatment and safety of the finished water.
2. The total conform group is preferred as an indicator of treatment
adequacy and fecal coliform. measurement is preferred for monitor-
ing raw water quality and to indicate the potential presence ol
pathogens at source. The present total coliform standards should be
retained but a minimal sample volume of 200 ml should be stipulated.
The more detailed U.S. regulations differ from this in that the recom-
mended sample volumes are not the same and that maximum per-
missible contaminant levels (MCLs) are given in detail. When the
new Canadian Standards are drawn up, the final recommendations
are likely to show, a greater similarity to the U.S. regulations.
3 The general bacterial population should be monitored regularly. The
"acceptable" level should be limited to 500 organisms per ml since
at higher levels, coliform measurements may not be reliable. An
additional recommendation is that the definition of coliforms should
include the cytochrome oxidase negative characteristic in order to
exclude members of the genus Aerompnas.
" 4 A free chlorine residual of 0.3 mg per litre should be maintained
' throughout the system, the turbidity of water entering the distri-
bution system should be limited to one Jackson turbidity unit. The
pH should not exceed 7.5. If higher pH levels have to be maintained,
it would be necessary to specify higher chlorine residuals for more
alkaline conditions. When chlorine residual measurements are
substituted for bacteriological examinations, they should not ex-
ceed 75% of the required examinations. In place of each colirorm
test omitted, three chlorine residual determinations should be car-
ried out It was considered desirable that the chlorine residual
should be determined daily but it was left to the control agency
to determine the frequency of tests. CPHA recommendations differ
from the U.S. regulations in specifying a higher level of chlorine
residual but this is lower than the 0.5 mg per litre required by the
1968 Canadian Standards. Another difference is in the number ol
chlorine residual measurements recommended for each substitu-
tion The mcrease in the use of ozone as a disinfectant by water
treatment plants, especially in Quebec, has to be considered m
•, : f •• • ".' , : • , ,';':.- i; j.: _ ; _ ;;"._ ' • :L • .- j^
••'* Environmental Standards Division, Environmental Health Centre, Health and
Welfare Canada, Tunney's Pasture, Ottawa, Ontario K1A OL2.
192
-------
CANADIAN DRINKING WATER STANDARDS/T.P. SUBRAHMANYAN
evaluating the merits of substituting bacteriological tests by chlorine
residual monitoring.
5. While sampling frequency based on population served might be
realistic and adequate for large centres, it was felt that the require-
ments for smaller communities are inadequate and that the minimum
sampling requirement should be raised to 10 samples per month.
Where this is not practical and less than 10 samples are examined,
the present standards should be modified to stipulate that none of
the samples should be positive for coliform organisms. When a
sample is positive, corrective action should be .taken immediately
and tests on daily samples carried out until no coliform organisms
are found in two consecutive samples. Greater emphasis should be
placed on sanitary surveys to detect problems in the system. Sampling
from trouble spots should receive greater attention than the col-
lection of representative samples.
6. In view of the unacceptable degree of variation in the efficiency of
different membrane filters in recovering coliforms, specifications for
membrane filters are required; these should cover all the parameters
that affect the recovery of these organisms.
7, Another aspect requiring attention is quality control. The use of
simulated water specimens for comparative testing in different labora-
tories and the development of reference procedures with which
methods used in the different laboratories may be compared were
~ 7 . suggested. :
8. Research into techniques for simple methods for the detection of
streptococci of exclusively fecal origin has been recommended.
Health hazards due to the presence of Klebsiella organisms in wa-
ter and the possible use of fecal sterols as an absolute indicator of
'""" fecal pollution" merit further study. :
9. Virological examination of water was discussed at length. It was
decided not to recommend virological examination as a routine
procedure although the present bacteriological standards may not
reflect viral contamination. It was felt that since recent technological
developments are encouraging, virological studies should be car-
ried out selectively and that virological surveys of drinking water
should receive priority when developing research programmes.
10. It is necessary to determine whether standards are being met,
whether bacteriological surveillance programmes meet established
criteria and whether the water supply systems are adequate for pro-
'~ ~ "viding safe drinking wafer. Therefore, a drinking water supply
survey which encompasses" "different parts of Canada has been
recommended.
Finally, some of the differences between the Canadian position and the
revised U.S. regulations reflect the fact that the revision process in the
United States is at a more advanced stage. Further, the Canadian Standards
provide only guidelines for regulatory agencies. When the new Canadian
StanHanls" are finalized on the basis of the CPHA recommendations and other
submissions, some of these difference!? may disappear. Those based on
differences between our situations will, however, remain. We wish to ac-
knowledge the generous cooperation we received from Dr. P. Toft, Chief,
Environmental Standards Division, Health and Welfare Canada, Ottawa and
Mr. L. Vlassoff, Chief, Microbiology Section, Ontario Ministry of the En-
vironment, Toronto.
193
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
,. : QUESTION AND ANSWER SESSION
C.W. Hendricks, Office of Drinking Water, U.S. Environmental Protection
Agency, Washington, D.C.
Do you think suppliers will sample at the rate of 8 to 10 per month?
T.P. Subrahmanyan, Enteric Virus Laboratory, Ontario Department of
Health, Toronto, Canada.
The water supply utilities in Ontario, from which most of the Canadian
experience is drawn, are run either by the Ontario Ministry of the Environ-
ment or by other public agencies in cooperation with the Ministry. In the
case of the major municipalities, the collection of 8 to 10 samples per month
has not presented significant problems. I am not in a position to comment
on utilities run by private companies.
194
«|l|> llllllu'll II K : Jll I ill) Hi ;, "' , liiiiiU , l-l'Lr . ,l . ., ' . , Iliiiii - . lalllll ! Jill /,!':,; ';
-------
Kansas City, Missouri Water Treatment Plant
PANEL DISCUSSION
Alternative Means of Determining Compliance
Chairman: Dr. Sumner M. Morrison, Department of Microbiology, Colorado
State University, Fort Collins, Colorado
195
-------
K ?,••! SB iilIKi itid!
S;,;.,-!.!,:i; L 'iiiiiti '& mi11 •;., s '.-i "ii,!> . „:;::: Hi it Hit; i;•« 1 • • i, Ji: AJU, .:tv
llllllli iiiiiii ii 111 ill nil
III III
•nil
-------
ALTERNATIVE MEANS OF COMPLIANCE/S.M. MORRISON
Alternative Means of Determining Compliance
Simmer M. Morrison
Department of Microbiology
Colorado State University
Fort Collins, Colorado 80527
Microbiologists, in the area of laboratory work for the surveillance of
water supplies, are sort of in the analogous position of the common expres-
sion "painting yourself into a corner". Well developed, simple, inexpensive
tests for coliforms exist and are being productively applied in; the control
of drinking water quality for a large portion of our population. In con-
junction with the work of the engineers and health officials, the laboratory
work has contributed to the decline in the waterborne illness over the past
50-60 years. Now the mandate is to expand this supply of safe water to all
public systems, including many, many small ones. This philosophic or political
change in the need for safe water for all has created our problem. What was
considered a simple, inexpensive, expeditious and effective procedure for
metropolitan population centers now is considered a procedure, for many
reasons already discussed, too cumbersome, costly and ineffective for the
small, systems. Some inconvenience, cost, dislocation and personnel train-
ing is going to occur in meeting the charge that Congress has given EPA
and us as scientists and professionls with a stake in near zero risk drinking
water for all.
I would like to comment about a few points that may be alternate ways
of establishing compliance. Some of these ideas may be "far out" and not
worthy of consideration, but this panel is to explore all possibilities.
1. While population served seems to be a very logical way to assign re-
quired tests, is there reason to consider volume treated or volume delivered
for domestic use as a means of determining required numbers of tests?
Systems that supply a high proportion of then- water for industrial, cooling
or irrigation uses are different from systems that supply domestic water
only.
Similarly, is there reason to establish categories of requirements based
upon type of water and type of treatment system, or perhaps the availability
of engineering support? Perhaps the information we now have would not
allow such an approach, but if better statistical risk data were to be accumu-
lated we might be better able to provide the same degree of protection to
all people with a more equitable cost to each.
2. The use of monthly enumeration and averaging of samples is an
arbitrary use of time periods. A system using a 30 day, or even a 60 or
90 day running average would better indicate deteriorating water systems.
The arbitrary calendar month will cause system operators to establish test-
ing times to avoid getting caught with a high test in the last part of the month
and other similar ways of not using good statistical sampling.
3. A major fear that seems to have been expressed is that compliance pro-
cedures do not really take into account possible sampling or laboratory error.
A procedure allowing immediate retesting, perhaps for 3 consecutive days,
would allow the discard of the single high coliform test without setting in
motion reaction from the State and possibly public disclosure.
4. There seems to be an attempt to directly link the MCL of the bacteria
to day-to-day prevention of disease. From an epidemiological standpoint it
should be placed into proper context. Periodic sampling of small portions of
197
-------
MICROBIOL STANDARDS EYALUATION/C.W. HENDRICKS
water for coliforms, when coupled with proper design and operations man-
agement, has proven quite effective in preventing frequent large scale water-
borne disease outbreaks. It has stood the test of time. Other alternatives to
bacterial testing, as frequent chlorine residual sampling or engineering surveys
only, need to be tried buiin an experimental mode. Until these alternatives
have stood the test of time, as has the cpliform test, there should be both
short-term and long-term surveillance research. To avoid again not having
the answers to pertinent questions, the research must be well-designed and
directed by outside as well as in-house expertise; as free as possible from
self-serving bias. !..'.' .... f
"5. We have heard some previous speakers analyze and criticize some ot
our'long-established testing procedures on the basis of modern statistical
analysis. Among the very important research areas that needs development in
4 above is an attempt to put such items as sampling frequency, sampling de-
sign, minimal required types of tests, correlations with human waterborne
illness patterns, etc. on a solid defensible statistical basis.
6 To me, the concept of public notification when a water system appears
to be in non-compliance, is a two edged sword. While it might provide ap-
piopriate pressure on the water supplier to be in continuing compliance, it
could also serve to create complacency in the population of a community.
In some places it could lead to severe disruption of the local community,
endless litigation, suits for all types of real and imagined damage, political
ictribution, scare tactic competition and advertising and wasteful and un-
ethical methods to avoid notification. In other places, with no penalties be-
yond notification, the local people will be successfully sold on the idea
that the notices are bureaucratic garbage and they will evoke no response.
In either extreme case the end result will not be safer drinking water but
rising consumer costs for the legal and political maneuvers.
Far more effective would be a State program of compliance by cooperative
effort with a review and emergency team to attain the compliance. By
charging the costs for these emergency reaction measures back to the water
supplier, the pressure would exist for staying in compliance and the chances
for meeting the safe drinking water objective would be greater.
With these few remarks, I would like to introduce this afternoon's panel
speakers. .
The panel members for the remainder of this session are:
Dr. Warren Litsky, University of Massachusetts; Mr. Berry E. Gay, Jr., of
the Division of Laboratories in the Illinois Department of Public Health;
Mr; Ira Markwood, Director of Public Water Supplies for the State of
Illinois; Dr. G. Wolfgang Fuhs, Director of Laboratories and Research for
the New York Department of Health; and, Dr. Wesley O. Pipes, Drexel Uni-
versity. These individuals will present their ideas concerning alternative
methods of determining compliance to the microbiology standards.
UK lliMyiiliil
198
i.jkit,uiiU Jill iiJM^ i: .Livi!'iiii. 1.:.'.lia Hiii ii i. -^ • .,; l,,:;,!i...,,1, ilii. jl-: i. 't i it :, ;! liliiilf a i,iilii, tllilllliiiiiii.iliii::
-------
THE COLIFORM MCL/W. LITSKY
The Coliforin MCL: Can We Defend It?
Warren Utsky
Department of Environmental Sciences
Marshall Hall:
University of Massachusetts
Amherst, Massachusetts 02003
During the past day and a half we have heard some excellent papers con-
cerning evaluation of the coliform standard, detection and control of coliform
bacteria and compliance with the coliform standard as they apply or pertain
to .the National Interim Primary Drinking Water Regulations. We have heard
the pros and cons of many ageless scientific debates and we have also been
privileged to the arguments of a new breed of ecologically oriented micro-
biologists and. their rational for adopting alternative indicators of water con-
tamination. We have listened to the advantages and disadvantages, the
rational and the public health considerations of substituting the chlorine
residual for the traditional microbiological procedures. After listening in awe
to the parade of expert speakers I asked myself, "What will the outcome be
for all this rhetoric?"
This question was not prompted by sarcasm or smugness but rather by in-
creased activity of consumerism and consumer groups. The days are long
gone when recommendations or regulations of Government Agencies were
universally accepted. Unfortunately, the trend,now is to distrust any govern-
ment official, challenge any document put forth and above all ignore the
scientists since "they are the group that got us into this present day mess."
Many legal actions initiated by consumer groups against government in the
name of _.the ..people have been brought to my _ attention with the result that
a few of my colleagues are becoming extremely wealthy by acting as scien-
tific advisors to law firms. Because of this trend and because of Section
141.32 which mandates that persons served by a community water system
be notified when the system fails to comply with the applicable MCL, I
predict a rash of legal entanglements, the likes of which have never been
reached, and EPA better be ready for this deluge. As a mental exercise during
the past few days I have asked myself, "What advice would I give to a lawyer
who has been retained by a water supplier ajccused of exceeding the bac-
teriological MCL?" Realizing that the object of any court action is to prove
that your opponent or the regulations, etc;, that he is enforcing are incorrect,
illogical, or just plain stupid, the scenario which I present may be a trifle
biased, to. say .the least.
With these introductory remarks concluded, let me present a few items on
which I would base a case were I acting in the. capacity of a legal advisor.
I would most definitely attack the substitution of the chlorine residual for
the coliform standard. The data in Table I demonstrate that there is most
definitely fecal coliform bacteria in water showing a high chlorine residual.
While these data are based on sewage effluent and. combined chlorine
residual, as measured by the OT test, they suggest that chlorine residual
monitoring is not as infallible as its proponents wish us to believe.
Leaving this subject I would-then advise the court that while we fully sup-
port the present coliform standard, we challenge the procedure by which
this standard is measured! In as much as the present regulations give an un-
fair edge to those using the Multiple Tube Technique over the Membrane
Filter Technique for enumerating .the coliform bacteria, I contend that
199
-------
I
MICROBIOL STANDARDS EVALUAf ION/C.W. HENDRICKS
, ,, • . ..: '. . , , . ,
TABLE 1 EFFECT OF CHLORINE RESIDUAL AND CONTACT TIME
ON FECAL COLIFORM DENSITIES IN SEWAGE EFFLUENTS.
Plant
A
A
A
A
S
S
N
N
Effluent
primary
primary
primary
primary
secondary
secondary
primary
primary
CI Residual Contact F. C. MPN
rng/1 mm
1.5
1.0
1.25
0.75
• •: : :. .
1.5
0.5
1.2
1.1
10 7900
12 7700
10 9200
15 270
20 20
11 35,000
15 4,900
20. 2,400
OF* 330
OF 1,100
25 5,400
20 16,000
* outfall
the former method will be universally employed and, in the interest of time,
will limit the remainder of my discussion to the Multiple Tube Method.
While the definition of-the coliform group has been accepted for many years I
w.puld question whether the techniques are actually capable of isolating and
enumerating this particular group. We all have heard the outcries of Dr.
Victor Cabelli of EPA that members of the Aeromonas group, a common
soil inhabitant, can not be distinguished from the coliform bacteria using
the present standard methods. Thus, are we measuring coliform bacteria,
Aeromonas species, a mixture of the two, or God knows what?
The use of the MCLs intrigues me and may also amuse the court. Case in
point, Section 141.14 defines the MCL in actual numbers per 100ml when
the Membrane Filter Technique is used while when the Fermentation Tube
Method is used the MCL is defined as coliform present in a 10ml standard
portion. In other words the MCL for the latter is defined as the number of
tubes showing gas. By employing this system, the calculation of the Most
Probable Number (MPN) and its very embarrassing 95 percentile confidence
limits range are avoided. Whatever the reasons for this may be, the fact still
remains that both the MCL and the MPN are based on positive fermenta-
tion tubes and, call it what you wish, the MCL are going to be correlated
with the MPN which we have witnessed many times during the past day
and a half. , , ,
Now for the MPN or MCL of coliform bacteria. As one who has been
teaching sanitary bacteriology for the past twenty-five years, I have become
disenchanted with the accuracy and precision of the MPN technique for
coliform bacteria as outlined in Standard Methods. I have seen too much
variation on replicate samples by technicians and students employing flawless
technique. The precision of the MPN just is not there and the data in Table
2 will illustrate this. Fecal coliform determinations were made on marine
watersj chlorinated sewage effluents and laboratory chlorinated sewage efflu-
ents hi triplicate by a technician with ten years laboratory experience. As can
be seen in Table 2, the resulting triplicate MPNs show that a ten-fold range
200
-------
THE COLIFORM MCL/W. LITSKY
TABLE 2. TRIPLICATE FECAL COLIFORM MPN RANGE.
SAMPLE MPN/100 ml
Marine
Chlorinated
Effluent
Lab
Chlorinated
Effluent
49
79
94
8
700
2,200
49,000
490,000
35,000
240,000
49,000
310
350
220,
1,300
1,700
7,900
130,000
790,000
. 110,000
540,000
130,000
310
540
350
1;300
7,900
13,000
490,000
2,400,000
240iOOO
2,400,000+
490,000
is not uncommon. These data should not surprise anyone working with the
MPN technique and yet the standard is based on this type of an analysis.
Following this I would attempt to inform the court that other investigators
have-published numerous papers warning-of the fallacies concerning the
MPN technique; but.first let us consider the statistics on which the MPN
; was originally based. ...... ,
McCrady (4) in 1915 proposed the MPN based on Bayesian statistics, a
branch of statistics that modern books do not even mention. I find _ it im-
possible to explain what Mr.B ayes' theorem was and how it is involved with
the MPN. However, my friends in the statistics department informed me that
the following statement from McCrady's paper is the basic assumption of
Bayes' statistics as applied to the MPN. I hope that you are able to make
more sense out of it than I could:
"There is a certain assumption involved in this principle which
must be recognized. This assumption is that all numbers of B.
coli are equally probable; that is, that in the long run of samples,
one number of B. coli will appear about as often as any, other
number.
It.is to be noticed that this assumption is the same, in kind, that
is involved in the determination of the 'Most Probable Number.'
There, it is assumed that the 'Most Probable Number', so de-
termined, is as probable as any other number."
In 1957, Woodward (7), then Chief of Water Supply at the Robert A.
Taft Sanitarian Engineering Center, USPHS, discussed the precision of the
MPN and he concluded that:
.- "The lack of precision of MPN estimates of bacterial densities
is generally recognized—at least by those who perform the tests.
Like many other measurements used and interpreted by others
than the laboratory workers who fashion them, MPN estimates are
often credited with a precision they" do not deserve.
Precision confidence limits have been computed for three-tube
and five-tube tests using multiple decimal dilutions and for the five-
tube smgle dilution test 'These limits are broader than the approxi-
mate ones based on a log normal distribution of MPN's. For a
201
-------
i rri ifw " • i in111 'i I i inni1 ' i i •• " •• '.'• T v7^ • : •'! l;l"!i '.JT "Ji" ••! v1" • <. •. ' ;it* v' •'vM^'m 'n'j ;"1"l; "!" * «•"• ' i't:>: r iw; ira1 HI!
MICROBIOL STANDARDS EVALUATION/C.W.HENDRICK&
three-tube multiple dilution test, the 95 percent confidence limits
cover approximately a 33-fold range from approximately 14 to
458 percent of the MPN estimate. For a five-tube multiple dilu-
tion test, the" 95 percent confidence limits cover approximately a
thirteen-fold range from approximately 24 to 324 percent of the
MPN. In short, MPN estimates are not precise and many of them
are inherently improbable." .
Laubusch (2) in 1958 noted several limitations in the dilution technique
for the determination of MPN of coliform bacteria in water:
"Not the least of these is that the results depend on the number
of tubes observed in each dilution and may be greater than the
"true" coliform density (as much as 23 percent high for a series of
5-5-5 tube tests and 43 percent high for a series of 3-3-3 tests).
Other limitations include deficiencies in laboratory technique and
personal errors associated therewith, differences of sensitivity
of alternate procedure methods, influence of one variety of organisms
on another, lack of true randomness of bacterial distribution, etc.
For these reasons, the MPN by the dilution method should not be
treated with sacred reverence."
As to the standards Laubusch wrote:
"The currently accepted 'Standards,' (6) which involve planting
of a minimum of five 10-ml portions, were not altogether acci-
dental. There appears to be no rational basis in the establishment
of a minimum of five portions, rather than say ten or more, except
that it is necessary to limit the amount of labor and materials
involved to such an extent that the results are commensurate with
the requisite accuracy—an entirely reasonable approach.
" The most accurate method of evaluating fermentation tube test re-
sults hi terms of a determinate number is still open to dispute, al-
though in more recent years this is less obvious due to the general
acceptance and convenience of tables of most probable numbers
from the Hoskins (1) adaption of the Reed (5) formula based on
the Poisson random distribution."
Laubusch then summarizes by quoting Hoskins whose MPN tables we all
have used throughout the years:
Twenty years ago Hoskins asserted that ". . . it seems reasonable
to conclude that the dilution tube method cannot result in an ac-
curate enumeration of coli-aerogenes organisms and that, therefore,
the need for an acceptable and dependable method still remains."
He added further that ". . . if correspondingly as much study (as
the MPN) were devoted to the development of a suitable plating
phase of coli-aerogenes enumeration, it is quite reasonable to believe
that this effort would yield a method superior to the present one."
Finally a short quotation from a 1975 paper by. Oblinger and Koburger
(5) illustrating the mathematical bias in MPN values:
McCarthy et al. .(3) demonstrated a substantial mathematical
bias in MPN values relative to plate counts. With the agar plate
count method as the control estimator, 10 replicate MPN's indi-
cated an average bias of +29%, +10%, +6%, and —4% when
compared to the plate counts' arithmetic average, geometric mean,
median, and harmonic mean, respectively. The precision of 10.
replicate plate counts was at least three times that of replicate
MPN values. Perhaps Woodward summed it up the best: "The
202
-------
THE COLIFORM MCL/W. LITSKY
lack of precision of MPN estimates of bacterial densities is gen-
erally recognized—at least by those who perform these tests."
After disposing of the MPN technique I would focus on the subject of
check samples. Section 141.21 (D) (2) states that when the coliform bac-
teria occur in three or more 10ml portions of a single sample, two consecu-
tive daily check samples shall be collected and examined at the same.
sampling point. It can be argued that the check samples are required because
of the belief that most positive results are due to an error in. the collection
process or in the laboratory analysis but seldom due to the fact that coli-
form bacteria were actually present in the water. I have heard speaker
after speaker refer to .spurious positive analysis, laboratory errors, unsatis-
factory collection procedures .as if they were extremely common occur-
rences and thus the need for check samples. I contend that the collection
routine and the laboratory analysis should be such that a positive tube
indicates just what it should indicate—trouble! I submit that there can be no
room for error in a clinical laboratory especially during the ;typing and
matching of blood; likewise, there should be very little room for error when
analyzing finished waters. Like it or hot, the check sample does not and
cannot indicate the actual condition of the original sample. I see no reason
why these common errors in the collection of the sample or in the laboratory
analysis should be condoned as routine happenings. They need not occur and
moreover they should not occur to the degree that has been indicated at
this symposium. ^ .
"I will be the first to admit that my arguments presented here are ex-
tremely biased and somewhat facetious. Nevertheless, these are the types of
arguments that may use in future legal confrontations—and EPA had bet-
ter be ready for them.
REFERENCES
I, Hoskins, J.K., and C.T. Butterfield. 1935. Determining the bacteriological
quality of drinking water. JAWWA 27:1101-1109.
2. Laubusch, E.J. 1958. As a yardstick for pollution, how good is the MPN
coliform index? Water & Sewage Works J05:334-338.
3. McCarthy, J.A. et al. 1958. Evaluation of the reliability of coliform density
tests. Am. J. Public Health 48:1628-1635. .
4. McCrady, M.H. 1915. The numerical interpretation of fermentation tube
TesultsT J. Infect. Dis. 17:183-212.
5. Oblinger, J.L., and J.A. Koburger. 1975. Understanding and teaching the
most probable number technique. J. Milk Food Technol. 35:540-545.
6. United States Public Health Service. 1946. Drinking water standards. U.S.
Government Printing Office, Washington, D.C.
7. Woodward, R.L. 1957. How probable is the most probable number? JAWWA
49:1060-1068.
---.- -QUESTION AND ANSWER SESSION
Mr. Kimball
What is the problem of false samples occurring from improperly sterilized
bottles which, are done by commercial labs?
203
-------
Ill 11111111 1 1 III 1 ' !".''• :"~ 111 < 1 III |i Hill IIP II 111 " ; " •• ,'' ;:""; • 1* ' "'" ' 1 1 'in 1
Hill II
I'l ' 1 1 '" :''i:, 1 111 111 1 1
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
• ! •
W. Litsky, Department of Environmental Sciences, University of Massachu-
setts, Amherst, Massachusetts
It is assumed that employing standard methods for th'e analyses of water the
results obtained reflect the condition of the water. This requires that all
chemicals, solutions, media and glassware be sterile. A further requirement
is that the analyses be carried out by a qualified technician m order to pre-
vent contamination of the sample, glassware, diluent or media from^the
environment, fingers, etc. In other words, the bacteria which are counted on
the plate or by the MPN technique must originate from the sample and
from nowhere-else. An improperly sterilized sample bottle can in itself, be a
source of bacteria thereby negating the value of the water analyses. Improperly
sterilized bottles have no place in the water sampling routine. Any legitimate
commercial laboratory should be able to supply sterile sampling bottles If
they can not accomplish this simple feat they should not be tolerated. They
should, like the desert dwellers, fold their tents and silently slip away.
i , • '..
Comment by G.W. Fuhs, Division of Laboratories & Research, New York
Department of Health, Albany, New York
In addition to laboratory imprecision of MPN results, environmental
variability in natural surface waters (patchiness) can cause a standard devia-
tion (in a lognormal distribution) corresponding to a factor of 4, and some-
times corresponding to one order of magnitude and more. This was de-
termined at optimal laboratory precision (20 to 80 colonies per plate,
membrane filter technique).
...•-.,.. •...•.... , , :.... .. .„ ; .....v| , , „ ..,!'!
Participant from New Jersey
Collectors' error had been found to occur in the form of a batch of
nonsterile bottles which gave false-positive results. This meant that isolated
positive results could very well be in error.
G.W. Fuhs
This observation reflects a typical quality control problem and is not in
conflict with Dr. Litsky's and my own views. In the case mentioned, posi-
tive results were discarded upon investigation and for a good reason. Dr.
Litsky and I argue that isolated positive results should never be discarded
without cause.
204
-------
OTHER MEANS OF COMPLIANCE/B.E. GAY
Alternative Means of Determining Compliance
Berry E. Gay, Jr.
Division of Laboratories !
Illinois Department of Public Health
Springfield, Illinois 62701
I understand there is a lawyer here in Washington, D.C. who is attempting
to aid bureaucrats so they will write their agencies' regulations in plain
English. With that in mind I'll make no attempt.to write a regulation.
I realize that the federal government has not stipulated by regulation that
all community water supplies shall be chlorinated. Illinois Environmental
Protection Agency regulations do require that public water supplies shall be
chlorinated. Based upon the number of public water supply samples sub-
mitted to two of the three Illinois Environmental Protection Agency labora-
tories from July 1, 1975 to June 30, 1976, I have documented from the sum-
maries provided by these two laboratories the following—together the two
laboratories analyzed 60,665 potable water samples and 4,131 of them were
positive for total coliform, in terms of percentage that's about 7%, but after
verification procedures the actual percentage was about 2%. The two per
cent verified positive for coliforms may have been spurious results due to a
fluke in sample collection, an error by laboratory or a miscue by water treat-
ment plant personnel. - , ... .,...:
In the Statement of Basis and Purpose for the Proposed National Interim
Primary Drinking Water Standards (3), on page 9, there is the following
statement, "The presence of any coliform bacteria, fecal or non-fecal, in
treated water should not be tolerated". In. the Illinois Pollution Control Board
—Rules and Regulations'—Chapter 6 Public Water Supplies (2) there is a
definition of the word safe—"Safe" means that the water contains no sub-
stances or organisms which are or may be injurious to a person in normal
health who ingests the water". Currently the application of check samples (4),
whenever a positive coliform result is obtained whether >4 coliform colonies/
100 ml by membrane filter (MF) or three or more tubes with gas by most
probable number (MPN) an immediate resampling from the same sampling
point is required until two consecutive daily check samples show no coliform
colonies by MF or positive tubes by MPN. Thus we can assume the goal is
zero coliform.
Each Environmental Protection Agency laboratory sends out procedures
for collecting samples, but at times it appears to be a wasted effort as samples
arrive without complete sampling data, bottles are completely full, and no
one has signed the collection sheet. Before going any further, it ;appears we
have a weak link to discuss. The weak link is the sample collector. No where
in the National Interim Primary Drinking Water Regulations (4) is this per-
son's responsibility described. We know in a few months the final criteria for
Water Laboratory Certification will be published and a reference to approved
laboratories is provided in Section 141.28 of NIPDWR (4). Those of us that
have reviewed the proposed Criteria and Procedures for Certification of
Water Laboratories Involved in Public Water Supply Analyses noticed there
is a definite procedure given for collecting water samples when the laboratory
is assigned the responsibility—unfortunately not all laboratory personnel col-
lect samples. Perhaps upon rewriting of NIPDWR someone might add a
statement such as this—"At the discretion of the State Regulatory Agency—
water sample collectors, after a successful training period, may be issued a
205
-------
MICROBIOL STANDARDS EVALUATION/CAvl HENDRICKS
of mistakes, cutting coUs or "improving "JJ^^^J™
when U.S. Public Health Service wrote Drmhng Water
<£=
es
A
to tnt eaenand their
many commercial laboratories
One omy co-workers lives
and pays in the neighborhood
water supplier sh
cndar
since one S£.
the water supply. This bring
Samplmg and Analy
«The supplier
density samples
main function is to be of service. We have
approved for water analysis and we feel rtn
tt-st
indfcatine presence of coliforik whether within or exceeding the MCLs.
One olmy co-workers lives in a small community of about 650 popdation
of $20.00 per month for water He feels his
uld provide safe water (no conform bacteria
y waer suppe
and further believds one sample per month is not adequate I
th **»
.mple per month
me to Section 141.21
ical Requirements I have a definite dislike the
Con-
ina Samplmg an nay ca e
way the words monitoring and frequency are used m this section.
Sd£Lm b-«The supplier of water for a community water
of water for a community water system shall
at regular time intervals and in number pro-
sy sa
population Served. In no event shall the frequency be less
forth below". Then there is the chart showing Population and
um number of sampl
.
3 that are required to be taken each month.
jr means to watch or check on, monitoring is
JYLUilxlujL aa u.v.ui-1-" ^j <•— 'L "«"•" A~fi*»*A \\\r WphstFT
the action word meaning checking on. Frequently also ^defined by^Webster
means happening at short intervals, and frequency the fact orCondition of
" " ."• ff °.^ _T___. .4 ^iam \X7~ VIQVP cnmft laboratories capable
Monitor as defined by Webst
occurring frequently. Here's t
of examining all the required
206
problem. We have some laboratories capable
minimum number of samples in one day. For
-------
OTHER
example, in one of our metropolitan laboratories there are fourteen analysts
approved for total coliform procedures and
required for them to analyze is 480 per rnonth. If their collection; unit de-
cided to collect all samples in one day and
these fourteen analyists could easily anal1
community of 10,600. population the Registered Sanitarian could easily col-
lect all required twelve samples in one day
yst in the laboratory who has the capability to process them after delivery.
In both cases cited, both laboratories; have
the minimum number of samples required
MEANS OF COMPLIANCE/B.E. GAY
the minimum number of samples
wrought them in to the laboratory,
rze them. Likewise, in a smaller
and deliver them, to a lone anal-
the capabilities of examining all
per month in one day. Thus, as
it is now written, this would be satisfactory, but it is hardly monitoring at
short intervals.
I suggest the minimum number of samples per month be increased to four
(one per week until all four quadrants of the community are sampled.) If only
one sample is taken, the community could be in trouble the rest of month
and no one would detect anything until the next month. Besides, how could
those that go for bacterial result averaging do it with one sample? Not only
that, it is hardly what one would call monitoring when frequency means
monthly!
As of this date, our small and large community water suppliers ,are in a
state of chaos—they all have to secure tie services of an approved water
laboratory or spend the monies to esitablis! i their, own. A "most unfortunate
situation. Our state has approximately 1,650 public water systems—there are
1,348 below the 5,000 population level, 119 hi the 5,000 to 10,000 level and
183 greater than 10,000 level. Of Chose 1,348 communities below 5,000
level, we have approximately 700 small cdrimunities in the 25-1,000 popula-
tion level. Our non-community supplies nuriber approximately 11,000-12,000.
A majority of these non-community supp ies are currently being sampled
monthly. 1 -
Since March 14, 1977, my office has hid approximately 90 requests for
laboratory certification information. Many requests are from individuals
wanting to start a commerical laboratory str ctly for the total coliform param-
eters. Some requests have come from county health departments, county re-
habilitation centers, factories, schools, smell and large communities. Three
small community representatives expressed their desire to establish a central
laboratory and do samples for other comriunities within a 25 mile radius.
They also felt sure the various mayors involved would agree to having a
central laboratory with each community sharing expenses. As it appears now,.
there are numerous groups with the same hought, and this, in my,opinion,
is a feasible arrangement. As an individual community, the cost would be too
great, but together it is affordable. There is also another route to go. We can
establish reciprocity with our neighboring states. We have two of our com-
mercial laboratories in Northern Illinois aialyzing water samples for com-
munities in Wisconsin and Iowa, so for tie southern area of our state we
will enter into a reciprocity agreement witi water laboratories approved by
states of Missouri, Indiana and Kentucky. ,
We have the technology to send men to the moon, and we have the tech-
nology to have a safe water supply.
While I decline to write a regulation, I am sure the Regulatory Engineering
Staffs can. I do suggest small communities shouldn't be singled out any more
than large metropolitan areas. The presence of coliform bacteria in public
drinking water La either location is very undesirable. A Federal Chlorination
Regulation would be of great value. We'vfe had laboratory certification for
years, but no one state has really insisted water plant laboratories obtain cer-
207
-------
•I '
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
tification. The certification program to date as carried out by Geldreich and
Nash is of the highest caliber. Those states that have approval programs
only surveyed laboratories upon request—as a voluntary program Some, ol
us take pride in our programs and feel we have accomplished standardization
of procedures so as to produce reliable data for the enforcement agencies and
weV°done so with the aid of Geldreich and Nash. I take umbrage that
spurious positive results are solely a laboratory error and I assume ^the engi-
neering sfaff in the water treatment plants would take offense i^boratory
personnel responded with a statement "someone goofed in the plant . While
the sample collector may be a weak link, I am sure they too feel .resentment
n being blamed. It is a matter of educating all concerned as to their respectrve
responsibilities-to, provide a safe water supply by plant personnel, proper
procedures that must be used when collecting water samples, and the proper
methodolosv that must be adhered to by laboratory personnel.
Ke comments from the various regions of U.S. Environmental Protec-
tion Agency I was surprised to note a desire to give a n equal ranking of
MF wkh MPN results-I suggest a reading of HoW Probable is the Most
frobable Number? by Richard L. Woodward (5) Mr Edwin Geldreich.*
Handbook for Evaluating Water Bacteriological Laboratories (1) page^izu
For those rewriting 141.21-why is it necessary to notify the public based
upon the averaging for the month? It seems to me when the first check sample
ffpositive for Ae presence of coliform the State Regulatory Agency should
be notified by the laboratory performing the analyses.
As for substituting residual chlorine determinations for a bacteno ogical
sample, I suggest deletion. Bacteriological examinations detect pollution,
ch£; residuals do not. State Regulatory Agencies should be given the
option or allowed to use their own discretion regarding chlorine residuals,
but not as substitutes for bacteriological • sampling. Perhaps sample collector
will determine chlorine residual at point of sampling and include information
on collection sheet. Sewage treatment plants have been given the go ahead
s'm to use MF on chlorinated effluents and on proving results should use
MPN for fecal coliform determinations. Thus coliform bacteria can be de-
tected in the presence of chlorine.
We have no laboratories in our state using five 100 millihter standard por-
tions for MPN. Perhaps the use of the five 100 ml portions could be deleted.
Those of us with active laboratory approval programs have a shortage or
survey forms. Mr. Geldreich's supply is very limited. Will each region allow
the states to come up with the supplemental procedures survey form for SPC,
Fecaf Coli and Fecal Strep and permit attaching it to the federal forms that
will be distributed for use? At times, it is necessary to supplement total coil-
form counts with results by Standard Plate Counts (SPC). SPC's are also
necessary for quality control within the laboratory. Fecal coliform procedures
are also used as a supplement to determine quality of source waters The
new survey forms are directed only to total coliforms. I guess Id better go
now and see the lawyer that is teaching bureaucrats to write in plain English.
'" ! •'•• -;" ' '"! • '••"'•••" :i: ":iiir '• ' ; ;
.*;! : REFERENCES
! Geldreich, E.E. 1975. Handbook for evaluating water bacteriological labora-
' tones. 2nd ed. Environmental Research Laboratory, Cincinnati, Ohio.
2. Illinois Pollution Control Board, Rules and Regulations, Chapter 6, Public
' Water Supplies—November 22, 1974.
3 Office of Water Supply. 1975. Statement of basis and purpose for the pro-
' pOSed national interim primary drinking water regulations. U.S. Environ-
mental Protection Agency, Washington, D.C.
208
-------
OTHER MEANS OF COMPLIANCE/B.E. GAY
4. U.S. Environmental Protection Agency. 1975. National interim primary drink-
ing water regulations. Fed. Register 40:59566-59574.
5. Woodward, R.L. 1957. How probable is the most probable number? JAWWA
49:1060-1068.
QUESTION AND ANSWER SESSION
VJ. Cabelli, U.S. Environmental Protection Agency, Narragansett, Rhode
Island.
Is it advisable to have a certification program for sample collectors due
to the turnover in personnel?
B.E. Gay, Division of Laboratories:, Illinois Department of Public Health,
Springfield, Illinois.
Yes, it is most advisable to have a certification program for sample col-
lectors.
As each public water supply treatment plant is evaluated for requirements
by State EPA Engineers and the laboratories are surveyed for compliance
with Standard Methods by Sjtate Water Microbiology Survey Officers, some
responsible person such as the plant engineer, chief operator or laboratory
supervisor should be evaluated by the State authority for proper sample col-
lection procedures. Upon certification toy the State authority this person would
then become responsible for training, surveying and certifying his own water
supply sample collectors. A report of these surveys should be sent to the
State EPA office, the Laboratory Survey Officer, the sample collector cer-
tification authority, and the Regional USEPA office.
This will tie in the sample collector to the program," "as he may make or
break the monitoring quality control program in the field.
209
-------
• • t ' liar :t
; I
'.if
-------
OTHER MEANS OF COMPLIANCE/I. MARKWOOD
i
Alternative Means of Determining Compliance
Ira Mssrkwood
Director, Public Water Supplies !
Springfield, Illinois 62706 j
i
I suppose you expect me to talk profoundly on thje questions that were
just proposed earlier. I don't intend to do so; I am not sjire I really understand
the questions, let alone the answers. I was supposed to come here and give
some alternate methods to improve the bacteriological . standards or replace
them. Joe Harrison got here first. I therefore will say I agree with Joe's pro-
posal No. 1, but there is one other thing I would like to see done; that is
that the sampling period be a running period rather tian a calendar month,
for example.
If you go by a calendar month, a supply can take samples the last week
of the month and the first week of the following mon:h and then go almost
two months before they have to take samples again. Oq if they should happen
to miss the end of the month by one day because their bottles came late, then
they are listed as not having sampled that month everi though their samples
came to the laboratory the following day. That is one| change that I haven't
heard specifically mentioned, which I would like to se
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
Montitpring is the last line of defense, it is really only there to see that what
should have been done before has been done, and that nothing has been
overlooked. So, while it is important, I wouldn't say it is the most important
thing that has to be done. I don't think any of the parts are more important
than another. They all have to be done and done right if we are to assure
the safety of a water supply.
I suppose we are no different than other State agencies. We don t have
enough money and enough people to do the job the way we think it should
be done, but we do the best we can and, because we can't make all the engi-
neering evaluations we would like to, and send our people out to all the in-
vestigations we would like to, we have to rely quite a bit on the monitoring
more than I think we really should. But that is one of the facts of life and
we have to live with it. '
Now, getting to the monitoring itself. The ideal for a test would be some-
thing that could be run accurately by an operator, would give a definite—
let's say binary—result, either yes or no, nothing in between. He would then
know just what to do. He wouldn't have to use his judgment. We don't have,
that. Instead we have the coliform test. As we all know, this is not perfect;
it has its problems. We spent most of two days here trying to discuss how
to interpret the results, so that it is obvious that there must be judgment with
the knowledge of the system, with the knowledge of the people who are
ppe that system, and of any other conditions which may affect it. It
can't be done by the ordinary operator.
One of the things we can't do, as you all know, is shut off a supply. That
is a last resort. We must try to keep the supply in operation and yet make
it safe, which obviously is the reason for boil orders. However, this in itself
creates problems because if we know that a system can be contaminated, then
how can we keep it operating? If we issue a boil order, we have problems
with compliance. We know people don't boil water. We had a case where we
notified a supply that their water was contaminated and a boil order should
be issued. The water superintendent said, yes, he would see that it was done.
We checked back a couple of hours later and it hadn't been done. We told
hhn to issue it immediately and he said, "Yeah, okay." We checked back an
hour later and found that he still hadn't issued it; he was going to wait until
the next day. We gave him a few choice threats, and insisted that it had to be
done immediately. He went on television and said, "The Illinois EPA says
you should boil your water, but I don't think you have to and I am not going
to boil mine." That is almost an exact quote.
When we take a sample we hope that it is representative. But really it is
only representative of the water that is passing by the sampling point at
that instant. We will never be able to get a representative sample of that par-
ticular water again because as it goes through the distribution system, it
changes and it mixes with other water in the system, is dispersed, so the
sample we get is a sample of one discrete portion of water. However much
Ve might try, it is not representative of the whole system and never will be.
AVe have to tal^e that in consideration when we set a standard.
We have to be very careful when we have public notification. I can give
you an example in a slightly different area. We have mandatory fluoridation
in Illinois. I had a woman call me up to ask me if there was some way in
which a petition would get her municipality released from mandatory fluori-
dation because she didn't want it. I talked with her for a little while and
found out that she was drinking bottled water? They were buying bottled
water for drinking purposes—she, her husband and the children—and they
were only using the public water supply for bathing and so forth. We got a
212
^^^ i
-------
OTHER MEANS OF COMPLIANCE/I. MARKWOOD
sample of the brand of bottled water that she was using. The municipality
was holding pretty well between .9 and 1.2 parts of fluoride. The bottled
water had about 5 parts.
When people start distrusting their water, t iey'11 go elsewhere, and almost
invariably the other source will be less safe tian the public water supply. If
we start giving public notice with little or no real backing for the notice, in
other words, if we give it just because a number has been exceeded (a num-
ber which has been selected somewhat arbitrarily), then we are going to have
two results. We will probably have first one, land then the second. The first
result will be panic and will have people going off in all directions trying to
avoid the use of the water at all costs, including using the nearby pond, and
that stream that looks so nice and clear as it runs past the sewage plant outfall.
Then, they are going to decide that this is just| a lot of crying of "wolf", and
after a while, when we have a major emergency,, we will send out a notice and
nobody will pay attention. And that is one of the things that I am worried
about. I would hate to see that attempts to implement the Safe Drinking Water
A -.*. •.*._..1J t. it__ _ 1 , • i 1 _ _ "
Act would eventually result in poorer or less s;
this country.
Thank you.
e water for the communities in
213
-------
! I.1 :,„!!,; ,,"ll nlffl""1!1'!1, : ' "!!'i'''i||lp;''
t1'!"•f|'?:;'i '•'.13i'li'"'iiiiiiii!IFilii
If fii
1 •,
II
IK; i. liiiiiiibiii, •!, I"!),,, iiiiiiiiiiiiiiii „ it ' !t i IMIK hi.. tiiifllB^^^^^^^ illllliii i, ii' :. in
., lip,;,' ; ....... i' ;.:\ , j.' ..... "tfi ..... i ^ ;
..... jliii il'liiiiiliiiidiii' ''
ik. i i'l ..... mll
-------
OTHER MEANS OF COMPLIANCE/W.O. PIPES
Alternative Means of Determining Compliance
Wesley O. Pipe's
Department of Biological Science
Drexel University
Philadelphia, Pennsylvania 19104
Over a period of a day and a half it has become quite clear that there is a
real problem associated with establishing microbiology standards for drinking
water. What is not clear is the exact nature of the problem. It may be a
public health problem. It may be a microbiology problem. It may be a
statistics problem. It may be the engineering problem of providing adequate
treatment facilities and distribution systems
whose integrity is difficult to
broach. It may be the personnel problem of fiading capable people and train-
ing them properly as operators, water samplers, and laboratory technicians. It
may be a regulatory problem. It is probably k mixture of all of these.
Previous, speakers have concentrated mostlj? on the technical parts of the
overall problem. Thus, it is appropriate for ire to make a few comments on
the nature of the regulatory problem.
First, I would like to say that I have never seen a technical problem solved
by a regulation. Technical problems cannot ie legislated (or regulated) .out
of existence. They have to have technical solutions if they are to be solved at
all. However, regulations can create technical problems. There needs to be
some very serious thought about separating the regulatory part of the overall
problem from the technical parts and about how to word the regulation so
that it motivates efforts to solve the technical parts of the problem.
There is a basic conflict between regulatory approaches and technical ap-
proaches. Lawyers write regulations and we can be sure that whatever sub-
jects and numbers technical people put into the regulation to begin with, the
lawyers will have the last chance to determine the wording. Lawyers talk
about "good draftmanship". What they mean is that in wording a law or a
regulation it should be made clear that, if a person does A, B, and C, he is
operating perfectly legally and within the law,
he has violated the law and is subject to a fin
other penalty they can dream up. They want
from C to D so high that few people will take
In technical matters, usually A is excellent,
but if he does A, B, C, and D
; or a prison term or whatever
to make the penalty of going
the risk.
B is good, C is poor, D is bad,
E is worse, and F is awful. Somewhere within the range of C, D and E, a line
between the acceptable and the unacceptable is passed but the exact location of
-this line is difficult to determine precisely. As an example, if the monthly
average concept is used in the microbiology standards for drinking water,
a specific number needs to be selected. That number may be 1 coliform per
100 ml or 2.2 per 100, or 4 per 100. AH of these numbers are in the gray
area between acceptable and unacceptable and probably none of them is
technically defensible in an absolute, quantitative sense. If any one of those
numbers is selected, it is quite likely that someone will come up with a num-
ber of examples of water supplies which exceeijl the monthly average standard
in communities where there is no evidence ofl waterborne infectious disease.
On the other hand, an average of 16 coliforms per 100 ml is probably far
over the black and would be running too great a risk of outbreaks of water-
borne disease in too many communities.
The Primary Drinking Water Regulations are written in terms of maxi-
mum contaminant levels (MCL's). It appears that there is some attempt to
apply the MCL concept to different types of parameters on a somewhat uni-
215
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
form basis, f his may be a source of part of the problem Water treatment
Processes are not in particular designed to remove arsemc barium, cadmium
cSomum, lead, mercury, .nitrate, selenium, or silver although some Amount
of °ome of these substances ma^ get removed during treatment. Water treat-
ment processes are in particular designed to remove ^^.
bacteria from water. There is k statement in Standard Methods (1
. no coliform bacteria of any kind should be tolerated in a
snene to differentiate between the use of the coliform
coliforms as indicators rather than to establish a drmkmg
water standard and it is an exireme viewpoint. However it does emphasize
'wlr the concept that treied water should have no coliforms m it.
properly operated treatment system usually turns out finished water
coliforms present. When there are no coliforms m the finished water
cforl found in the distribution system indicate that water torn some
other source has been introduced. The discovery of coliforms m either the
SsnedTater or in the distribution system indicates that something is wrong
and corrective action should bl taken. If coliforms are found in .to _ finished
water at the treatment plant, tne treatment system should be investigated to
Sd oS wSy ^£e letting through and adjustments should be made »
*at they don't. Ifcoliform organisms are found in the .dutnbutom system .the
sbiirce M the water containing1 coliforms should be located and etaunated.
On the other hand, if the positive coliform tests are the result o : poor sam-
pling or poor laboratory technique, this needs to be discovered and corrected
Regulations written in terms bf monthly averages and percent of samples
Slowing positive results just don't get the message ac ^ . »
lowing positive
OK to have some coliforms
.
present. The regulations should be
o ave so .
that it is clear that positive resflts are not O.K.; they indicate a problem that
_ - .» Jt _ 1 __ _ ^lv*n .3 •w.-rt-rv-V'rt'f tIT
_ - .» Jt
needs to be solved and needs
Testing finished water at a
bacteria is not a very rewardi
so
:o be solved promptly.
well run water'''tr'eatm'eht plant for coliform
Lg activity because practically all the results
bactera s no a very rng, a^umj www«,~ — r~ ------- j — %
are 0 coliforms per 100 ml. That is why my research to determine the ire-
cmency distributions for colifortn organisms in water has been done using raw
water samples. The frequent result of 0 coliforms per 100 ml is a reason to
reconsider the use of monthly kverages in the regulations. That result doesn t
mean that there are no coliforkns present; it just means that the sample was
too small. If a liter, or 10 litlrs, or 100 liters, etc., of water were filtered
sooner or later a coliform org^ism would turn up. The 0 coliforms per 100
ml iust indicates less than 1 coliform per 100 ml. Averaging a string of num-
bers which include several zeijos which really aren't zeros creates problems
which the statisticians haven't solved yet.
There are statistical problem^ in estimating the number (or range of num-
bers) wnicn" best represent the results of a series of coliform determinations
I say this in full knowledge of the work of Clarence Velz during the 1940s
(7) and Harold Thomas during the 1950's (6). Their work really was ex-
cellent considering that they were sanitary engineers doing statistics. The
questions we are asking now are somewhat different, we are demanding more
precise answers, and we have more serious and more economically far-
reaching decisions to base on those answers. Also the field of statistics has
seen several new development^ in the last 15 years and some of the newer
statistical techniques offer better answers to our questions if we had the right
kind of data.
216
-------
OTHER MEANS OF COMPLIANCE/W.O. PIPES
Dr. Litsky presented a strong indictment of the proposed coliform stand-
ards based on the lack of precision of the most probable number (MPN).
What he presented is correct; however, he didn't give a fair and balanced
presentation. It is true that McCrady (5) was the first to use the term, "most
probable number", and it is true that he used Bayes' concept of inverse proba-
bility in his derivation of the MPN equation. It is also true that the Bayesian
approach has been discounted by mathematicians and no statistician would
use that approach now. McCrady was a bacteriologist not a statistician and
his entire mathematical approach was crude. However, the same MPN equa-
tion was derived more elegantly by Greenwood and Yule (4) two years later.
It was also derived by R. A. Fisher (3) in his classic paper on parameter
estimation in 1921 and there are several other derivations of the MPN equa-
tion which have been published since. I think that we have to accept the
mathematical soundness of the MPN equation if we accept the assumption
on which it is based.
The term, "most probable number," is an unfortunate one because statis-
ticians have rejected the Bayesian concept of inverse probability and because
the term implies a precision the number doesn't have. It should be called the
"maximum likelihood estimator of the mean coliform density" but, of course,
McCrady didn't know this when he wrote his paper because Fisher didn't
invent the term "maximum likelihood estimator" until six years later. It is re-
markable that a bacteriologist anticipated Fisher's general result for this
specific case by six years.
The Bother references which Dr. Litsky cited were trying to point out that
the precision of the MPN is not as high as most people who use the multiple
tube fermentation test think it is. They \vere not saying that the MPN value
had no precision at all. It would, of course, be a good thing for the people
who use MPN values to understand what its precision actually is and for
this to.be taken into consideration in .the regulations.
Standard Methods (1) does give the 95 % confidence interval range for
MPN values. This range in most cases is a factor of 5; that is, using the table
in Standard Methods we can distinguish between water with 1 coliform per
100 ml and water with 10 coliforms per 100 ml but we Can't distinguish be-
tween water with 1 coliform per 100 ml and water with 5 coliforms per 100
ml. The table in Standard Methods is a bit optimistic. The assumption behind
the MPN equation is that the numbers of bacteria in small volume aliquots
from a water sample will fit a Poisson distribution. The characteristic of the
Poisson distribution is that the variance is equal to the mean. Anyone who
has had much experience doing coliform counts on water samples knows that
the variance is much greater than the mean. It even says so in Standard
Methods (1). That was the main point Dr. Muenz was trying to make yes-
terday when he was discussing the data I had sent to him.
The statistical model which Dr. Muenz developed is really rather simple
compared with models other statisticians have suggested to me and it is very
powerful. If we had the right data on coliform counts in treated water and
water in distribution systems his model could be used to put the sample fre-
quency table of the regulations on a rational basis. I suspect that his "bowl-
ing ball" model; that is, most of the water with no coliforms and only an
occasional slug of water with some coliforms is the correct one. From that,
I suspect that if. the sampling frequency table were put on a rational basis,
it would indicate a need for a much greater sampling frequency than anyone
has suspected so far.
I have talked mostly about the regulatory problem and the statistical prob-
lem because the other technical aspects of the overall problem have been well
217
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
, . , „ , .1 . . , , . ,,
covered by other speakers. I would like to end by agreeing with a couple of
grew McCabe about the need for real time response to bacterial
conation of a public water supply. Such a contamination «j^ to be
detected as soon as possible and needs to be corrected as soon as P"»"«.The
bacteriological tests we have are not suited for a monitoring program because
SteWSmnd time. Some parameter such as residual chlorine needs to
be used ior monitoring purposes. It would be more rational to use the micro-
b o^gtaHesTng for short-term, intensive studies of the level .oi ^contamination
raw water, of the effectiveness of the treatment system, of the m-
3 ^distribution system, and of the validity of some alternate mom-
n I want to say is that I agree with what Archie Greenberg
said thi morn?ng about flexibility in the regulations. Colif orms are surrogates
for patiiSens or for public health effects. Standards and regulations are
urroPgates for judgments. The regulations attempt to prejudge every sihiafaon
which might arise. We just aren't that smart. A regulation might coyer 90%
S the^fes adequately but there always will be exceptional cases m which
following a regulation literally will lead to the wrong response to the situ-
ation The regulation must be written with enough flexibility so that the
SicaT peopfe who have the direct responsibility for water supply can find
the right response and solve the problem.
REFERENCES
1!,1- ,'.?; u ••••
3 ££;TA.^
Philos. Trans. R. Soc. London (Series A) 222:309-368.
4. Greenwood, M., Jr., and G.U. Yule. 1917. On the statistical interpretation of
some ^ bacteriological methods employed in water analysis!. Hyg./6. 36O4.
5> McCrady, M.H., 1915. The numerical interpretation of fermentation tube re-
' suits. J. Infect. Dis. 17:183-212.
6. Thomas, H.A., 1955. Statistical analysis of coliform data. Sewage and Ind.
7. Velz, C.J., 1951. Graphical approach to statistics. IV. Evaluation of bacterial
density. Water and Sewage Works 95:66-73.
l|l|| I Ill
-------
OTHER MEANS OF COMPLIANCE/G.W. FUHS
Alternate Means of Measuring Compliance with Bacteriologic
Standards of Public Water Supplies
G. Wolfgang Fufas
Environmental Health Center
Division of Laboratories and Research
New York State Department of Health
Albany, New York 12201
• The question of what constitutes an adequate measurement of compliance
with bacteriological standards in water supplies should be extended to a
consideration of the standard itself and of its utility for the protection of
public health. For the purpose of this conference, therefore, we should at-
tempt to develop a logical connection between the absolute , measurement of
performance by bacteriological testing and the meaning of that performance in
terms of public health. . .
In an earlier paper (1) I attempted to define mathematically the relation-
ship between bacteriological standards and public health risk for bathing
beaches. A distinction was made between two situations, governed by what
might be called the laws of, large and small numbers. These lines of reason-
ing are easily applied to considerations of water system safety.
The terms large numbers and small numbers refer to the size of the human
population in a watershed. For sources of potable water polluted by large
: populations (e.g., the Great Lakes and the Mississippi and lower Hudson
rivers), the probability of ingesting an infectious dose can be calculated from
the probability of ingesting a certain number of viable infectious particles and
«the probability, of infection .from. :suc.h_ a particle.. This calculation must in-
clude proportonality factors, such as . (-a) -the abundance of excretersjof Jj£__
fectious particles in the human population a"nd j5T~tBe~ratib of infectious
bacteriaTA large polluting popu-
^
lation virtually"" assures the presence of the infectious agent, but in numbers
that are. reasonably stable and predictable and generally quite low.
When water polluted by a large population serves as a source of potable
water, the public health risk associated with its use depends to a great extent
upon the abundance of excreters in the watershed population, which deter-
mines the ratio of infectious .particles to indicator bacteria. The density of
the infectious agents can be reduced, often by several or many orders of
magnitude, by precipitation, filtration andean initial disinfection step. For
infectious agents which can be effectively removed by this treatment, the
product will often be quite safe for drinking even without final disinfection —
not because the pathogens are absent, but because their numbers are so
small as to virtually preclude ingestion of an infectious dose by any one con-
sumer. Final disinfection can then be considered mainly protection against
sources of contamination affecting the treated water.
The law of small numbers applies to sources of drinking water affected
by small populations and to drinking water distribution systems. Whenever
the polluting population is small, the ratio of pathogens to indicator bacteria
is subject to wide fluctuations depending on the presence or absence of
excreters in the watershed. The safety of the supply with regard to a speci-
fied disease is assured when there are no excreters in the population, but
the water may be highly unsafe if and when only a few or even one excreter
is present. In the extreme case the probability . of one excreter being present
at any time is virtually identical with the probablity of an outbreak. Assum-
219
-------
Figure 1. AREA COVERED IN SURVEY OF WATER SUPPLIES IN NEW YORK
STATE. COUNTY LINES ARE SHOWN.
SCAIE IN MIIES
10 0 10 20 30 40
-------
OTHER MEANS OF COMPLIANCE/G.W. FUHS
ing that excreters are randomly distributed and as mobile as the general
population, this probability will obey a Poisson distribution with time, with
the ratio of excreters to general population as the mean. Therefore the
public health risk in this case also is predicated upon the prevalence of in-
fection in the general population.
This principle explains three phenomena: (a) short episodic outbreaks of
waterborne disease related to occasional pathogenic contamination of a water
supply whose treatment system fails intermittently or contiuously, (b) the
absence of waterborne disease in certain small, isolated communities, despite
almost constant contamination of their water systems with fecal bacteria,
and (c) the occasional water system which produces illness whenever a vital
disinfection step fails, indicating access to the system by a resident excreter.
In dealing with situations governed by this law of small numbers^ and for
tbejpurpose of protecting (rather than predicting) public safety, the^nuniber
o'F^^ters_ for any given ^d7s^a?en^rH^>o^lat|on^houId be abitrarily set at
the theoreticaj^gxpectea" number plus one. This additional excreter repre-
sents thF1B^ividlIal~wE6~may haveTccess to the system at some time in the
immediate future and thereby create an acute hazard. This approach, in
effect, leads to a policy of zero permissible contamination for all potable and
recreational waters affected by small populations and, in particular, for all
public water supplies, since their distribution systems are susceptible to
localized access or failure.
The strategy for a zero-contamination policy consists in setting up one or
several protective barriers: watershed protection with fences and patrols;
impermeability of sources, reservoir structures, and distribution systems; and
disinfection. Each barrier is intended ,to be absolute but is in fact not so,
due to the imperfection of all human endeavor. The effect of any small num-
ber of failures of any barrier, can be reduced, often dramatically, by a second
protective barrier, as long as the failures of .^either barrier-are randomly dis—
tributed with time and their causes are independent of one another. In this
case, the over-all probability of failure of the barrier system is the product of
the probabilities of failure of the individual barriers (together with the proba-
bility that some person having access to the system is excreting pathogens, if
fecal bacteria are the indicators of system failure).
Since these probabilities combine in a multiplicative fashion, we should not
be surprised to find a lognormal distribution of the number of incidents of
contamination with fecal bacteria per unit of time. This indeed is the case.
Bacteriologic records from 726 public water supplies in 42 mostly rural
counties of New York State_(Fig. 1).were..examined for reports of conform
bacteria. (This is the geographical area served by the sanitary bacteriology
laboratories of the New York State Department of Health.) Detection of
any coliform bacteria in one or more samples in any calendar month during
which samples were received was counted as one occurrence. This was done
in an attempt to assign equal weight to failures of the source, of the treat-
ment, and of. the distribution system, .although the use of the calendar month
is rather arbitrary. Most records covered 3 years (July 1972-August 1975),
with a minimum of 12 months. Sampling frequency was based on popula-
tion according to the 1962 Public Health Service table, but with a minimum
frequency of 12 rather than 24 per year.
The cumulative frequency distribution of the average number of occur-
rences per year, plotted on log-probability paper, shows a straight line over
most of the range (Fig. 2). This reflects the multiplicative behavior of com-
ponent probabilities, as explained above. TLie different slope ofjhe_gurve in
the range of supplies which fail most fri3quentiy~may indicate a different
221
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
re^singie jLuu-im aam.k'i^j LWI.WV.. ..
wTthatT7% of the supplies wefe free of coliforms for 3 years;
an value was 1 event every 13 months, or 0.9 -ents per year How-
ever 10% of the systems .showed cohforms. every 3 months or more ire
quStly"and 1% were more or less continuously contaminated Typical ex-
ampks of poorly performing systems are <«) a supply with a defectBranch
inlne distribution system which continuously delivers contaminated water to
an outlyng section of the community and (b) defective small supplies m
economically very depressed regions where improvements cannot reasonably
be expected without financial assistance from the outside.
. This method of determining- the performance of a water ^ »«"£
more stringent than the existing regulatory procedure, which allows the
99.9
•0 99.8
H 99.5
E1 99
% 98
m- 95
U_
O ,90
loJ
O
tr
UJ
O
80
70
60
50
40
30
20
10
5
O.i,
MONTHS BETWEEN/OCCURRENCES
100 ' 60
30 20
T
15 12
~1—T~
6
T
4
~T
2 1.5 I
_L__J—L
0.3 '0.5 I 2 3 4 5 6 8 1012
- OCCURRENCES PER YEAR
.. , , : , . ,.. „,
Figure 2. CUMULATIVE AVERAGE FREQUENCY blStRIBUTION OF COLIFORM
OCCURRENCES IN NEW YORK STATE WATER SUPPLIES, FOR
DEFINITION OF OCCURRENCE AND PERIOD COVERED, SEE TEXT.
FOR AREA doVEREb, SEE Figure 1.
II1II 111 111 || tiliiililPliiLH! llki; ,,mki. „,, IKJjL.ili.i ,b iL
-------
OraER MEANS OF COMPLIANCE/G.W. FUHS
0
I 2 3 4
OCCURRENCES PER YEAR
Figure 3. CUMULATIVE FREQUENCY DISTRIBUTION OF COLIFORM OCCURRENCES
IN FIVE LOCAL JURISDICTIONS IN NEW YORK STATE. THE SUMMARY
CURVE (HEAVY LINE) APPROACHES THE CURVE OF THE LOGNORMAL
DISTRIBUTION IN Figure 2.
averaging of coliform counts over all samples taken in any given month. It
also is suitable for characterizing the status of an entire region or jurisdiction.
Differences among five local jurisdictions in northeastern New York State are
illustrated in Figure 3. While the summary curve approaches the lognormal
distribution, the individual curves show markedly different degrees of per-
formance, which are suggestive of both regional socioeconomic differences
and the varying intensities with which corrective actions are sought on the
local level.
223
-------
I i ii 11 (nil; I « Ij|| !l
; i
to
I'll 1 1 1 1
o
•g
w
s
f
co
H
I
O
j FMAMJ JA SOND
Figure 4 SEASONAL DISTRIBUT10N.OF COLIFORM OCCURRENCES IN THE FIVE
JURISDICTIONS. HEAVY LINE: ACTUAL OCCURRENCES. LIGHT LINES:
CONSTANT FUNDAMENTAL, AND FIRST FIVE HARMONICS OBTAINED
BY FOURIER ANALYSIS. DOTTED LINE: LONG-TERM MEAN MONTHLY
TEMPERATURES FOR ALBANY, NEW YORK. THE DISCREPANCY
BETWEEN ACTUAL EVENTS AND PREDICTIONS IN DECEMBER AND
JANUARY WAS AN ARTIFACT OF THE SAMPLING SCHEDULE.
-------
GIBER MEANS OF COMPLIANCE/G.W. FUHS
The 726 supplies in our sample produced an average of 650 occurrences
per year, or almost 2,000 occurrences over the 3-year period covered by this
study. Yet there were only two outbreaks of disease. One occurred in a
larger supply which had about 4 occurrences a year but was mostly in com-
pliance by regulatory standards because these are based on average counts,
which in this supply were generally low. The outbreak was ascribed to water-
borne Giardia lamblia. The other outbreaw of disease was of a particular
freaky nature and can hardly be called waterborne in a strict sense.
It was ascribed to a rare pathogen, Yersinia enterocolitica, which was
traced to a supply with an almost flawless bacteriologic record. The supply
itself did not carry an infectious dose for any consumer, but the bacteria
multiplied in a commercial dairy product, causing the outbreak. This low
ratio of outbreaks to system failures shows that the pattern of outbreaks of
waterborne disease is governed by the: laws of probability.
The seasonal distribution of coliform occurrences in the five local juris-
dictions covered in the previous sample is shown in Figure 4. Fourier analysis
(the first six harmonics are shown) indicates a seasonal pattern which closely
follows the long-term mean monthly temperature curve for the area. The
abnormality in December-January is an artifact caused by the sample col-
lection schedule.
Another important consideration in formulating regulations on bacterio-
logic sampling and in determining compliance is the statistical power of the
bacteriologic test. The most powerful statement that can be derived from the
standard tests is the proof that coliforms (or any other selected group of
bacteria or agents) were present in the sample and therefore were present in
the system at the time of sampling. Whenever we deal with small numbers
of bacteria in a small number of samples, the quantitation of a positive result
is characterized by poor precision. But the power of the negative test is also
quite poor. A negative result for a single 100-ml sample, if interpreted at the
95% confidence level (which is standard in scientific work), should read
"fewer than three bacteria per 100 ml" and not "fewer than one," as has
.been customary in water supply monitoring. The probability that the state-
ment "fewer than one" is correct is only about 70%. (The remaining per-
centage points are attributable to the possibilities that there are one or two
bacteria per 100 ml.) Even this statement, which is based on the Poisson dis-
tribution, .holds only for small parcels of water, which are likely to show a
random and homogeneous distribution of bacteria, and does not hold for
entire distribution systems.
Whenever the statistical power of a. single test is small, we are told by
statisticians to continue testing until a satisfactory conclusion can be reached.
This is exactly the purpose of the long-term evaluation shown in Figure 2. As
long as a water system does not undergo a significant change the record
reflects "present Conditions" except for seasonal variations and random
events. If the sampling period spans at least 1 year and if at least 24
samples were taken, a rating of the system by this method is possible. I
tend to prefer a series of monthly samples taken during 3 years over the
same number of samples taken more frequently during 1 year because the
3-year series may include events such as floods, droughts, or other unusual
stresses on the water supply. —
This type of long-term evaluation has the advantage that a single occur-
rence of bacterial contamination is placed into proper perspective. Individual
occurrences of bacterial contamination may be ascribed to unusual circum-
stanes and may be excusable, but a high rate of recurrence of such incidents
is an indication of an unreliable source, ailing equipment, sloppy operation,
225
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS!
or any combination of these anjd other factors. The single incident also rarely
produces an outbreak of disease, but the aggregate of all incidents is a direct
measurement of the tightness of the system and therefore of the public health
risk associated with it, ! - .,
In this context I feel Strongly that no positive bacteriologic results should
be discarded without cause. In our experience, isolated positive results tend
to become confirmed in the long run, although not necessarily in the check
samples immediately following* which may not represent the same parcel of
water. Also, in our experienced false-positive results by laboratories are less
frequent than false-negative results, but this may in part be due, not to
laboratory error, but to the low statistical power of the negative result, as
explained above. j ."" . .
Long-term evaluation of the type indicated can be used to set priorities
for a detailed investigation of'systems that are performing poorly. Such an
investigation would include sanitary surveys with intensive sampling, using a
more complete spectrum of bacteriologic and chemical techniques. These
should include nonstandard but nevertheless firmly established microbiologic
procedures for the identification of bacterial types indicative of specific
sources of contamination. (The summary methods of coliform determination
and standard plate counts are| appropriate for determining the presence or
absence of any number of prjoblems, but they do relatively little to .reveal
the exact nature of the problem.)
1 The detailed investigation of poor or marginal supplies should be directed
at finding the most effective ways of upgrading them. The investigation of a
supply with average performance should aim at determining whether a sig-
nificant problem exists and whether performance could be improved at a
reasonable cost. j
Long-term evaluation and investigation are directed at finding ways for
long-term improvement of water supply operations and long-term improve-
ment of the prevalence of infections in the general population, which, as we
have seen, is essential in assuring that the fewest possible system failures re-
sult in disease. Nothing should i discourage regulatory peronnel from following
u£ positive bacteriologic results immediately and taking all action necessary
tp protect the public from acute health hazards. But an equally important
long-terni goal is a conclusivej rating of each water supply. This cannot be
obtained by a highly variable, month-to-month rating, which only breeds
confusion, indecisiveness, and, [if combined with a public notification scheme,
public apathy. j
As the problems of many water supply systems are chronic or of a long-
term nature, the solutions are often not of the quick-and-easy type. Disinfec-
tioncreates only-one barrier where several are needed. Chlorination cannot
compensatecprnpletely for the!shortcomings of deteriorating distribution sys-
tems, and theformation of chlorinated organic materials has become a cause
for concern. There is need forj improved watershed protection, for treatment
by advanced techniques, and particularly for the rehabilitation of old dis-
tribution systems. The need for costly, long-term solutions involving con-
structionprojects is particularly'obvious in areas in the northeast which are
already economically depressed.
Public notification of a poor long-term and overall performance of a supply
—nas opposed to proclamation {of an acute public' health hazard—can be an
important instrument to obtain support from an informed public for long-
term improvement of a water Supply. The prohibited-use classification of the
U.S. Interstate Carrier Sanitation Program was a valuable precedent in this
regard. Jn the absence of a construction grant program for water supplies,
226
i mi nil in 11 n ill n in i in n i i i n i n • n 11 i n i H i i in
-------
OTOER MEANS OF COMPLIANCE/G.W. FUHS
correct handling of the public notification issue assumes particular importance
if activities under the Safe Drinking Water Act are to become more than a
rather fruitless, frustrating exercise.
REFERENCES
1. Fuhs, G.W. 1975. A probabilistic model of bathing beach safety. Sci.
Total Environ. 4:165-175.
QUESTION AND ANSWER SESSION
V.J. Cabelli, U.S. Environmental Protection Agency, Narragansett, Rhode
Island.
What is the appropriate action for a one-barrier system to provide ab-
solute safety from an outbreak "of disease?
G.W. Fuhs, Division of Laboratories & Research, New York Department
of Health, Albany, New York.
Make sure that the barrier works at all times. If the barrier is a chlorina-
tor, it must be watched 24 hours a day, 7 days a week.
227
-------
JIB it
• iiK ijii
:i:
'JIB!',*
Hull III" Hi":'!11'1
'"(;"Hi!;"'n,, lib'illlllS^^^'Ii'iisiiiiiiit'i'i'liii1,' t' :'i' "''Jill'.!'!!;::'!' .:.>'• 'iillili;i!"'i'ii
.v 'tSS**,.&.i. ..... iS ....... 6, "Li •:.; .; ..... ., 1 ^'M, 'Slat ...... li ........... Uv I ! .:*: , I li ..... si A£, ...... nMO . » at ..... : "i '
........... !!lliiili:ii ......
-------
JLIST OF ATTENDEES
EVALUATION OF THE MICROBIOLOGY STANDARDS
FOR DRINKING WATER
OFFICE OF DRINKING WATER
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C.
APRIL 13-14, 1977
Martin J. Allen
Water Supply Research Laboratory
U.S. Environmental Protection Agency
26 W. St. Clair Street
Cincinnati, Ohio 45268
Ray E. Anderson, Jr.
Head, Public Water Supply. Section
201 W. Preston Street
Baltimore, Maryland 21203
Robert Becker
Indianapolis Water Company
1220 Water Way Blvd.
Indianapolis, Indiana 46202
James W. Berry
Fairfax County Water Authority
P.O. Box 151
Occoquan, Virginia 22125
Robert H. Bordner
Environmental Support Laboratory
U.S. Environmental Protection Agency
26 W. St. Clair Street
Cincinnati, Ohio 45268
Fran Brezenski
Region II
U.S. Environmental Protection Agency
Edison, New Jersey 08817
Victor J. Cabelli
U.S. Environmental Protection Agency
West Kingston, Rhode Island 02892
Craig D. Cameron
Fairfax County Water Authority
P.O. Box 151
Occoquan, Virginia 22125
Dean Cliver
Food Research Institute
University of Wisconsin
1925 Willow Drive
Madison, Wisconsin 53706
Rita R. Colwell
Department of Microbiology
University of Maryland '
College Park, Maryland 20742
Steve Cordle
Water Supply Research Staff
Office of Research and Development
U.S. Environmental Protection Agency
Washington, D.C. 20460
Joseph A. Cotruvo
Criteria & Standards Division
Office of Drinking Water (WH-550)
U.S. Environmental Protection Agency
Washington, D.C. 20460
Gunther F. Craun
Health Effects Research Laboratory
U,S. Environmental Protection Agency
26 W. St. Clair Street
Cincinnati, Ohio 45268
Patricia Creamer
U.S. Environmental Protection Agency
Region — II
26 Federal Plaza
New York, New York 10007
Rodney S. DeHar
2562 Executive Center East
Tallahassee, Florida 32303
229
-------
in ill ",:.,: wi" i n vf, \wn\vr
i
MICROBIOL STANDARDS EVALUATION/C.W. HENDRlCKS
John E. Delaney
Lawrence Expert nent Station
37 Shattuck Street
Lawrence, Massachusetts 01843
Robert J. Drye, Jr.
Environmental Sciences Branch
North Carolina Division of Health
,il Services
P.O. Box 28047 ,
Raleigh, North Carolina 27611
G. Wolfgang
Division of
New York
Funs
Laboratories
Department
Scotland Avenue
& Research
of Health
Albany, New Ydrk 12201
Barry E. Gay, Jr.
Division of Labc ratories . .
niinpis Departmmlof Public Health
134 N: 9th Street
Springfield, lilinbis 62701
Edwin E. Geldreich, Jr.
Water Supply Research Laboratory.
U.S. Environmental Protection Agency
26 W. St. Clair
Cincinnati, Oio
Street
45268
Ralph E. Gentrf
Environmental Protection Agency
College Station Road
Athens, Georgia
30605
Arnold E. Greenberg
Chief, Bibenvironmental Laboratories
California Department of Health
2151 Berkeley Way
Berkeley, Calif drnia 94704
David R.Goff
Waterwork
180? Park Road N.W.
Washington, D.b. 20010
;" ;;.';.; •• :,;: •• • "•:.-• •; ••••
Joseph Harrison
IIS. Environmental Protection Agency
230 Sputn Dearborn Street
Chicago, Illinois 60604
Paul Haney
Black and Veatch
P.O. Box 8405
Kansas City,
230
Missouri 64114
Department of Natural Resources
270 Washington Street
Atlanta,' Georgia 30334
Charles W. Hendricks
Criteria & Standards Division
Office of Drinking Water (WH-550)
U.S. Environmental Protection Agency
• Washington, D.C. 20460
Morris L. Hennessey
Division of General Sanitation
201 W. Preston Street
Baltimore, Maryland 21203
John C. Hoff
Water Supply. Research Laboratory/
„ 'MERL
U.S. Environmental Protection Agency
26 W. St. Clair Street
Cincinnati, Ohio 45268
Riley D. Housewright
Safe Drinking Water Committee
National Academy of Sciences
(JH-308)
2101 Constitution Avenue, N.W.
Washington, D.C. 20418
William J. Jamieson
North Carolina Department of
Human Resources
Raleighj North Carolina 27611
W. Blake Jeffcoat
Division of Public Water Supplies
State Office Building
Montgomery, Alabama 36130
Bill Kesler
Department of Natural Resources
P.O. Box 1368
Jefferson City, Missouri
Arnold M. Kuzmack
Office of Drinking Water (WH-550)
U.S. Environmental Protection Agency
Washington, D.C. 20460
Frank Lambert
Division Consolidated Lab Services
Box 18"77 1":' !"" ; : :
Ricmond, Virginia 23215
Roger Lee
(ME) PNW Regional Office
901 Fourth & Pike Building
Seattle,' Washington 98101
-------
Judith Lewis
Maryland State Health Department
201 West Preston Street
Baltimore, Maryland 21203
Warren Litsky
Department of Environmental
Sciences
Marshall Hall
University of Massachusetts
Amherst, Massachusetts 02003
Ira Markwood
Director Public Water Supplies
2000 Churchill Road
Springfield, Illinois 62706
Leland J. McCabe, Jr.
Health Effects Research Laboratory
U.S. Environmental Protection Agency
26 W. St. Clair Street
Cincinnati, Ohio 45268
Gordon A. McFeters
Department of Microbiology
Montana State University
Bozemari, Montana 59715 "
A. Thomas Merski
EPA —Region III
6th & Walnut Street
Philadelphia, Pennsylvania 19106
Cameron Moffat
Washington Aqueduct
5900 MacArthur Blvd., N.W..
Washington, D.C. 20016
S. M. Morrison
Department of Microbiology
Colorado State University
Fort Collins, Colorado 80527
Larry Muenz
Field Studies & Statistics Program
National Cancer Institute
7910 Woodmont Avenue
NIH, Landow Bldg. C-403 ... ,..._._
Bethesda, Maryland 20014
Harry D. Nash, Jr.
Water Supply Research Laboratory/
MERL
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
R. Warren Norris, Jr.
Water Supply .Branch, Region VI
U.S. Environmental Protection Agency
1201 Elm Street
Dallas, Texas 752J70
Otmar O. Olson
U.S. Environments 1 Protection Agency
Region VII
1735 Baltimore
Kansas City, Missouri 64108
Henry J. Ongerth,
Bureau of Sanitary
3uef
Engineering
California Department of Public
Health
2151 Berkeley Waj
Berkeley, Califormfi 94704
Wesley O. Pipes
Department of Bio
Drexel University
Philadelphia, Penn yl
Louis A. Resi
Technical Support
U.S. Environrrienta
5555 Ridge Avenue
Cincinnati, Ohio 4 268
Division, (OWS)
Protection Agency
Pamela Rhodes
Fairfax County Wa
Lorton, Virginia
Charles E. Rundgrea
Water Supply Brancjh
Route 4, Box 160
Hillsborough, Nortl
Augustus Ruser III
Florida Division of Health &
Rehabilitative Sejjvice
P.O. Box 210
Jacksonville, Florida 32201
Michael E. Sawson
ERGO
185 Alewipe Brook Parkway
Cambridge, Massachusetts 02138
Stephen Schmidt
Pennsylvania Department of
Environmental Resources
16tfi Floor, Fulton Building
Harrisburg, Pennsylvania 17120
ogical Science
Ivania 19104
er Authority
Carolina 27278
231
-------
MICROBIOL STANDARDS EVALUATION/C.W. HENDRICKS
Peter T. B. Shaffer
Carborundum Company
P.O. Box 1054
Niagara Falls, New York 14302
M. C. Snead
Johns Hopkins School of Hygiene
615 N.Wolfe Street
Baltimore, Maryland 21205
Eloise W. Stewart
Water Supply Laboratory
47 Trinity Avenue, S.W.
Atlanta, Georgia 30334
Stanley B. Stolz
Division of Natural Resources
P.O. Box 776
St. Croix, U.S. Virgin Islands 00820
T. P. Subrahmanyan
Enteric Virus Laboratory
Laboratory Services Branch
P.O. Box 9000
Ontario Department of Health
Toronto, Ontario, CANADA M5WIR5
Clifford Summers
Nebraska Department of Health
301 Centenial Mall South
Lincoln, Nebraska 68509
Floyd Taylor
Water Supply Branch
U.S. Enyironmenial Protection Agency
Boston, Massachusetts 02203
;I Peter" Toft
Environmental Standards Division
Department ofNational Health &
>•••' Welfare
Ottawa, Ontario, CANADA KlA OL2
L i i „ , *i" si «i ' .,•„,' ,"„ • ., 'I,,,:; , ' „„., , :,' HI
Richard S. Tobin
Health & Welfare Canada
Environmental Health Center
Ottawa, Ontario CANADA KIA OL2
Joseph E. Uridil
Nebraska Department of Health
301 Centenial Mall South
Lincoln Nebraska 68509
Kenneth W. Whaley
Tennessee Department of
. Public Health
Cordell Hall Building
Nashville, Tennessee 37219
Roger D. Wible
Carborundum Company
Research & Development Division
Niagara Falls, New York 14302
John' Wilford
Division Water Resources
P.O. Box 2809
Trenton, New Jersey 08625
John Winter
EMSL — Cincinnati
Environmental Protection Agency
Cincinnati, Ohio 45211
Michael Wojton
Maryland State Health Department
201 West Preston Street
Baltimore, Maryland 21203
Harold M. Wolf
Civil Engineering Building, R.m. 202
Texas A&M University
College Station, Texas 77843
Richard Woodhull
Conn-State Health Department
79 Elm1 Street
Hartford, Connecticut 06115
232
-------
INDEX
Achromobacter 14, 66
Acid-fast bacteria 39
Actinomyces 15
Adsorption:
Aeromonas 92
algae 43, 79
microbial interference 14
nutrients 44
Aeromonas 38, 39, 192, 200
American Water Works Association 105
Analytical considerations:
false positive results 149, 204
low-level contamination 151
negative results 151
MF vs. MPN spurious results 80,
168, 173, 197
survival 105
temperature 68
Analytical methods:
American Public Health Association
91
EPA Manual 85 _•- ....
Standard Methods 85, 91 -"-':,-..
U. S. Geological Survey 86
virus 96
Analytical models:
"bowling ball" 51, 53
negative binomial 51
Poisson 51, 217, 221, 225
statistical significance 198, 201, 215
Arthrobacter 14, 66
Ascaris lumbricoides 120
B
Bacillus 14, 66
Bacteria, specific
Achromobacter 14, 66
Actinomyces 15
Aeromonas 38, 39, 192, 200
Arthrobacter 14, 66
Bacillus 14, 66 '
Bacter-aides 39
Bifidobacterium 39
Caryophanon 39
Clostridium 14, 38, 62, 66
Corynebacterium 14
Enterobacter aerogenes 91
Escherichia coli 15, 25, 37, 42, 46,
49, 91, 94, 97, 104,112-114, 153. ...
Flavobacterium 14, 20, 45, 66, 67, 73
Gallionelfa 14, 66
Klebsiella 14, 3S, 43, 45, 66, 193
Leptothrix 14, 66
Micrococcus 15, 66
Mycobacterium 4. 66
Proteus 14, 66
Pseudomonas 14, 15, 20, 39, 45, 66,
67/68, 73, 91, 121, 122
Salmonella 39, 43, 73, 91
Salmonella typhi 43, 131
Sarcina 15
Serratia 14, 66
Shigella 43, 91
Shigella paradysenteriae 136
Spirillum 14
Staphylococcus 91, 119
Streptococcus 38, 43, 193
Vibrio cholerae 43, 73
Yersinia enterocolitica 30, 225
Bacterial distribution 53
Bacterial nutrients 40
Bacterial regrowth 18, 20, 45
Bacteroides 39
Bifidobacterium 39
Bottled water 69, 71
Candida parasilosis 4
Caryophanon 39
Check samples 34, 180, 206
Chelating agents 81
Chlorine demand 16
Chlorine residual:
DPD test 61
continuous monitoring 62
Chlorine-residual substitution 101, 147,
199, 208
Chlorine-resistant organisms 4, 96, 101
Cholera iii
Clostridium perfringens 14, 38, 62
Coliform bacteria 3, 43, 112
Coliform detection 77
Coliform-pathogen ratio 130, 219
Coliform standard:
implementation 159, 164, 169
limit 191
.Coliphage 39, 97
Community Water Supply Survey 17,
57, 59 '
Community water systems 21
Compliance with regulations 195
233
-------
Contaminated samples 129
Corynehacterium 14
Crustacea 105
."; ": D
Data calculation 164, 207, 215
Dead-end mains 15, 16, 132, 178, 179
Diatoms 179
Disease, specific:
amoebic dysentery 184
giardiasis 8, 23
infectious hepatitis 23, 31, 95, 105
salmpnellpsis 23, 25
shigellosis 23, 24, 33
typhoid Hi; 23, 24, 131, 181
Disinfectants:
chlorine, demand 110
chlorine dioxide 111, 121, 139
hypochlorous acid 106-114, 120, 121
.bzbrie, 121, 139
Disinfection:
chlorinatibn 32, 128, 132
interference 103
effectiveness 104, 119
residuals 17 1
Distribution systems 15, 67, 70, 81, 132,
178, 179 .
Drinking fountains 69
Drinking Water Standards:
see Legislation
Drug-resistance factor 120
Engineering control practices 125
Entamoeba histolytica 24, 115
Enteric bacteria
Enterobacter aerogenes 91
Enterotoxic Escherichia coli 23, 25
Esdierichia coli 15, 23, 25, 37, 42, 46,
49, 91, 94, 97, 104, 112-114, 153
Fecal coliform bacteria 8,,37, 43
Fecal coliform-fecal streptococci ratio
43
Fecal sterols 193
Fecal streptococci 38, 69, 193
Filters: ' ' " ' .' ""_
carbon 45
dual media 138
Flavobacterium meningosepticum 67
Gallionella 14, 66
Gastrointestinal illness iii, 23,' 32, 34, 96
Giardia 9, 24, 31, 39, 115
Giardiasis 8, 21, 23, 26
Ground water 31
H
Hepatitis A 24
234
Heferotrophic bacteria:
growth 40, 41
injury 40
..... , lh ........ ,, „ , I , ,
Indicator organisms
characteristics 37
concept 3
specific organisms 37
rapid methods 66
Interstate Carrier Program 153, 160
................. '. . K
Klebsiella 14, 38, 43, 45, 66, 193
Laboratory certification 88, 207, 208,
209
Legislation:
California 185-187
Canada 189-193
EPA Interim Primary Drinking Water
Regulations (1975) v, xi, 3, 9, 18,
29, 30, 59, 103, 140, 145, 146-149,
154, 159, 164, 166, 175, 179, 183,
205
Public Health Service Drinking Wa-
ter Standards (1942) 3, 154, .184
Public Health Service Drinking Wa-
ter Standards (1946) 3, 154
Public Health Service Drinking Wa-
ter Standards (1962) iii, 145, 154,
206
Safe prinking Water Act (1974) iii,
v, ix,2i, 145, 160, 167, 168, 180,
213 • .
Treasury Standards (1914) 3, 125,
153
Treasury Standards (1925) 3, .181,
184
Leptpthrix 14, 66
"'• Limuliis lysate 39, 46, 47, 48
' M ;
Maximum Contaminant Level (MCL)
46, 47, 97, 103, 146, 160-162, 164,
168, 180, 197, 199, 206, 215
Membrane diffusion chamber 41, 42
Membrane filter procedure 3, 15, 66,
80, 146, 167, 191, 199, 200
Microbial contamination 50
Microbial ecology 72
Micrococcus 15, 66
Monitoring:
chlorine residual 29, 57, 192, 218
frequency 55 • ' ' •
quality control 54, 193
statistical considerations 49
Monthly averages 148, 163, 168, 173,
176, 197, 208
-------
Most probable number procedure 3, 15,
66, 146, 167, 173, 191, 199, 200-
204, 208, 217
Mycobacterium 4, 66
O
Organic halogens 139
P
Pipe "C" values' 133
Poliovirus 97
Potable water 116
Proteus 14, 66
Protozoa, specific:
Entamoeba histolytica 24, 104, 115
Giardia lamblia 3, 9, 21, 24, 26, 31,
39, 115, 225
Pseudomonas 14, 15, 20, 39, 45, 66-68,
73, 91, 121, 122
Public notification 148, 159, 160, 180,
186, 198, 206, 212, 213, 226
Public water supplies 226
R
Rapid indicator systems 39
Ratios:
coliform-pathogen 130, 219
fecal coliform-fecal streptococci 43
virus-coliform 5
Safe Drinking Water Act iii, 21, 45
Salmonella 39, 43, 73, 91
Salmonella typhi 43, 131
Salmonellosis 23, 25
Sample collection 79, 205
Sampling:
frequency 79, 149, 193, 207
techniques 79
Sanitary survey 32
Sarcina 15
Semi-public systems 27
Serratia 14, 66
Sewage contamination 46
Sewage poisoning 23
Shigella 43, 91
Shigellosis 24, 33
Shigella paradysenteriae 136
Spirillum 14
Standard Methods 85-87, 91-93, 97,
-147, 171, 191, 200, 216, 217
Standard plate count 15, 20, 42, 175,
208
Standards evaluation 58, 145
Staphylococcus 91, 119
Streptococcus 38, 43* 193
Tap water 69
Temperature dependence 82
Transit time 80
Treatment deficiencies:
cross connections 25, 96
slime formation 133
Total coliform bacteria 3, 42, 43, 131,
193
Turbidity:
control practices 139
interferences with measurement 13,
18
research 105, 111
Typhoid iii, 23, 24, 131
V
Vibrio cholerae 43, 73
Virus 4, 91, 101, 104, 113, 121, 138
Virus, specific
adenovirus 4
bacteriophage 4
coliphage 39, 97
coxsackievirus 4
echovirus 4
hepatitus A 4, 24, 31, 95
poliovirus 4, 97, 104, 111
reovirus 4
Virus-coliform ratios 5
Virus concentrator 1
Virus detection 95
W
Water distribution system 11, 32
Water quality:
characteristics 65
data 6, 7
goals 130
Water softening 33
Water treatment 44, 98, 104, 137
Waterborne disease:
epidemics 21, 22, 134
etiology 21, 61, 95
indicator organisms 21
surveillance 27
Wells 32
•j}{- U.S. GOVERNMENT PRINTING OFFICE: 1979 O 275-621
235
-------
f
aiiiBSSt"!!11!!!!!!!1111 :i II|;!!!I!!I!!I!!;;;!!|I!I|I,IB! 'UKI'1!!!!!;;,:"!1!;!1!:;,1 ! !,/
i >''"% ''''• ' i ", " ',' '.''
=:~i~^^^^^^^^ RI5!!1R
. • 1 Wmi K 1 WF if! "I'llMB! rifflSBIfimffl MttlllllMnSff ''
i;i" ' ,'""'"!";I:': '" "'irl,':1",' ; =;,; ":• is :::'i: ,' ''if;' :i-; iiiii"1 ill""1: °""1, '"'
111"'] 'I! I!1:,!!
- ItHi!'! -ililirii', , !l
iliii'ji ....... ' .'I,:1!1 ...... ......... IF:i' :«' : IT
, T!l ,? 1, i,T I :!j|liil!!!i! W'
!"" i "111* 111,
XT
!'i (I
1
------- |