SYMPOSIUM
on the RECOVERY of
       INDICATOR
       ORGANISMS
employing membrane filters
33

\,
         LU
         C3
Environmental Monitoring and Support Laboratory
        Office of Research and Development
        U.S. Environmental Protection Agency
                Cincinnati, Ohio 45268

-------
                RESEARCH REPORTING SERIES

Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology.  Elimination of traditional grouping  was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
      1.  Environmental  Health Effects Research
      2.  Environmental  Protection Technology
      3.  Ecological Research
      4.  Environmental  Monitoring
      5.  Socioeconomic Environmental Studies
      6.  Scientific and Technical Assessment Reports (STAR)
      7.  Interagency Energy-Environment Research and Development
      8.  "Special" Reports
      9.  Miscellaneous Reports
 This document is available to the public through the National Technical Informa-
 tion Service, Springfield, Virginia 22161.

-------
EPA-600/9-77-024

PROCEEDINGS OF THE  SYMPOSIUM ON THE
                        RECOVERY OF
                        INDICATOR ORGANISMS
                        EMPLOYING MEMDRANE FILTERS

                        EDITED BY
                       ROBERT H. BORDNER
                       CLIFFORD F. FRITH
                       JOHN A. WINTER
                        September 1977
                    Co-sponsored by:
            • The American Society for Testing and Material
              • The Environmental Protection Agency

    UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
          OFFICE OF RESEARCH AND DEVELOPMENT
   ENVIRONMENTAL MONITORING AND SUPPORT LABORATORY
                CINCINNATI, OHIO 45268

-------
                                        DISCLAIMER

    This report has been reviewed by the Environmental  Monitoring and Support Laboratory-Cincinnati,
U.S.  Environmental Protection Agency, and approved for publication. Mention of trade names or com-
mercial products does not constitute endorsement or recommendation for use.

-------
                                           FOREWORD
     Environmental measurements are required to
determine the quality of ambient waters and the
character of  waste effluents. The  Environmental
Monitoring  and  Support  Laboratory-Cincinnati
conducts research to:

     *    Develop  and  evaluate  techniques  to
         measure the presence and  concentration
         of  physical,  chemical, and  radiological
         pollutants in water, wastewater, bottom
         sediments, and solid wastes.

     *    Investigate methods for the  concentra-
         tion,   recovery,  and  identification  of
         viruses, bacteria and  other microorgan-
         isms in water. Conduct studies to deter-
         mine the responses of aquatic organisms
         to water quality.

     *    Conduct an Agency-wide  quality assur-
         ance program to assure standardization
         and  quality  control  of  systems for
         monitoring water and wastewater.
     This publication of the Environmental Moni-
toring  and Support Laboratory, Cincinnati, en-
titled:

     Symposium  on the  Recovery  of  Indicator
     Organisms Employing Membrane Filters, re-
     ports  the  proceedings  of  meetings  co-spon-
     sored  by ASTM  and EPA  for the specific
     purpose of solving current analytical problems
     in microbiology. Such meetings  involving the
     combined expertise of government, academic
     and industrial  laboratories working with the
     manufacturers  have the greatest chance for
     solutions  that will be acceptable to everyone
     involved in monitoring and  controlling  pol-
     lution in the environment.
                   Dwight G. Ballinger
                   Director, EMSL - Cincinnati

-------
                                           ABSTRACT
     The  Symposium on the Recovery of Indicator Organisms Employing Membrane  Filters sponsored
jointly by the United States Environmental Protection Agency and ^he American Society for Testing and
Materials (Committee D-19 on Water)  brought together users, manufacturers, research scientists and repre-
sentatives  of government agencies to exchange technical information and review the performance of mem-
brane filters. Problems had been reported with the recovery of bacterial indicators from water and waste-
waters by the membrane filter procedures. They were most pronounced in the fecal coliform test. A key
question was  whether  the cause was  differences in sample types,  membrane filters or the test method
employed.
     Professionals experienced in water analysis presented  relevant field experiences, laboratory data and
research findings and discussed problems concerning recovery of organisms stressed  or injured by environ-
mental factors. Media,  transport phenomena, physical and chemical characteristics of membranes, mem-
brane sterilization methods, incubation temperatures, techniques for comparison of methods, data analysis,
and the status of the proposed ASTM methods for evaluating membrane filters were discussed.
     Solutions suggested at the Symposium included use of  two-step incubation, overlay and/or enrichment
techniques and modification of membrane  filter structures. Recommendations were made to  manufac-
turers and to users to develop and improve intralaboratory quality control  programs, to standardize inter-
laboratory testing procedures, to participate in these collaborative studies and to generally improve com-
munications among users, manufacturers and standard-setting organizations.
                                                IV

-------
                                          CONTENTS
Foreword	  iii

Abstract  . .	  iv

Acknowledgement  	   vii

List o.f Attendees	viii

     Color Plates	  1

     Welcome	  5

     Summary	  6

     Recommendations	  7

     Session I Uniform Procedures and Quality Control
         Session Chairman — Warren Litsky

         The Membrane Filter Dilemma	    8
              Robert H. Bordner

         Performance Variability of Membrane Filter Procedures	   12
              Edwin E. Geldreich

         Quality Control of Membrane Filter Media	   20
              David Power

         Statistical Interpretation of Membrane Filter Bacteria Counts	   26
              Karl J. Sladek* Clifford F. Frith and Richard A. Cotton

         Effects of Injury on the Recovery of Indicator Organisms on Membrane Filters	   34
              Alfred W. Hoadley

         Effect-of Temperature on the Recovery of Fecal Coliforms	   42
              James B. Hufham

         Optimum Membrane Structures for Growth of Fecal Coliform Organisms	   46
              Karl J. Sladek,* Robert V. Suslavich, Bernard I. Sohn and Fred W. Dawson

     Session II Comparision Studies of Membrane Filters
         Session Chairman — Phillip E. Greeson

         A Comparison of Membrane Filters and Media Used to Recover Coliforms from Water	   58
              Michael H. Brodsky* and Donald A. Schiemann

-------
    Comparison of Membrane Filters in Recovery of Naturally Injured Coliforms	   64
         John E. Schillinger, Gordon A. McFeters and David G. Stuart*

    Efficiency of Coliform Recovery Using Two Brands of Membrane Filters	.•	   67
         Frederick A. Harris* and Carl A. Bailey

    Comparison of Membrane Filter Brands for the  Recovery of the Coliform Group	   73
         Alfred P. Dufour* and Victor J. Cabelli

    A Comparison of Membrane Filters, Culture Media, Incubation Temperatures,
    Polluted Water and Escherichia coli Strains in the Fecal Coliform Test	   82
         Paul J. Glantz**

Session III Modifications to Improve Recovery
    Session Chairman — Phillip E. Greeson and Robert H. Bordner

    Recovery Characteristics of Bacteria Injured in the Natural Aquatic Environment	   98
         Gary H. Bissonnette, James J. Jezeski, Gordon A. McFeters* and David G. Stuart

    A Layered Membrane Filter Medium for Improved Recovery of Stressed Fecal Coliforms ...  101
         Robert E. Rose, Edwin E. Geldreich* and Warren Litsky

    Measurement of  Fecal Coliform in Estaurine Water	   109
         Alanson P. Stevens, Rosario J. Grasso* and John E. Delaney

    An Evaluation of Methods for Detecting Coliforms and Fecal Streptococci
    in Chlorinated Sewage Effluents	   113
         Shundar D. Lin

    The ASTM Proposed Membrane Filter Test Procedure for the
    Recovery of Fecal Coliforms	   133
         Don W. Davis* Margareta Jackson and George R.  Kinser

    Critique on ASTM Test for Recovery of Fecal Coliforms and
    Proposal for Modified Method	   153
         Norman H.  Goddard

Summary of Symposium	   157
    Francis Brezenski and John Winter

Final Discussion	   162

Final Remarks	   175
    Clifford F.  Frith

APPENDIX

    Comparison of Membrane Filter Counts and Plate Counts on Heterotrophic and
    Oil Agar Used to Estimate Populations of Yeast, Fungi and Bacteria	   178
         J.  D. Walker, B.F. Conrad, P.A. Sessman, and R.R. Colwell
 *
**
Speaker
Paper summarized by Bernard J. Dutka, CCIW, Hamilton, Ontario
                                              VI

-------
                                   ACKNOWLEDGEMENT
    We wish to thank the  attendees and speakers for  their  lively participation and interest. We also
acknowledge the wholehearted support given to the Symposium by the Subcommittee  D-19  on  Water,
ASTM.
                                            VII

-------
                                      LIST OF ATTENDEES
Dr. Robert K. Alico
Research Scientist
NASA — Ames Research Center
Biological Adaptation Branch
Moffett  Field, CA 94035

Mr. Thomas S. Arline, Jr.
Arline's  Pollution Abatement
1627 N. 29th Ct.
Hollywood, FL 33020

Dr. Frank T. Baker
Water Microbiologist
Ecotest  Laboratory
P.  0. Box 247
Elderton, PA 15736

Mr. D.G. Ballinger
Di  rector-MDQARL
EPA
Cincinnati, OH 45268

Mr. Ken Barnhill
Director of Publications
Camp Dresser & McKee
One Center Plaza
Boston, MA 02108

Mr. Frank Bazogh
Biologist
Lee Co.  Env. Protection
P.  0. Box 398
Ft. Myers, FL 33902

Mr. J.B. Bell
Supervisor, Microbiology
Laboratory
Dept. of Environment
5320 122 St.
Edmonton, Alberta,
Canada

Dr. Gerald Berg
Chief, Biological Methods Branch
MDQARL
NERC, EPA
Cincinnati, Ohio 45268
Dr. Herbert G. Berger
Regional Engineer
NCASI
P.O. Box 14483
Gainesville, FL 32604

Dr. Gary K. Bissonnette
Asst. Professor
West Virginia University
Rm. 401 Brooks Hall
Morgantown, WV 26506

Dr. Michael Bloomstein
Microbiologist
Pall Corp.
30 Sea Cliff Ave.
Glen Cove, NY 11542


Mr. Robert L. Booth
Technical Coordinator
MDQARL
NERC, USEPA
Cincinnati, Ohio 45268
Mr. Robert H. Bordner
Chief, Microbiological Methods
MDQARL, U.S. EPA
1014 Broadway
Cincinnati, OH 45202
Mr. James A. Brewtenstein
Environmental Specialist
City of Ft. Lauderdale
Utilities Department
4241 N.W. 11th Ave.
Ft. Lauderdale, FL 33309
Mr. Francis J. Brezenski
Edison Laboratory
U.S. EPA
Raritan Arsenal
Edison, NJ08817
Mr. Michael Brodsky
Scientist, Environmental
Bacteriology
Ont. Min. Health
Box 9000, Term "A"
Toronto, Ont., MSW 1R5, Canada

Dr. John D. Buck
Assoc. Prof, of Microbiology
Marine Research Lab.
University of Connecticut
Noank, CT 06340

Mr. Robert M. Carter
Product Administrator
Beckman Inst.
Box 6100
Anaheim, CA 92806

Mr. Richard Chen
Sr. Filtration  Engr.
Amerace Corp.
Ace Road
Butler, NJ 07405

Mr. Harold P. Clark
Bio. Lab. Tech.
EPA
4676 Columbia Parkway
Cincinnati, OH 45268

Mr. Jas. L. Collins
The Standard Oil Co.
2109Glendale
Toledo, OH 43614

Mr. Richard A. Cotton
Technical  Liaison
Millipore Corp.
Ashby Road
Bedford, MA  01730
Mr. D.W. Davis
Research Associate
Johns-Manville
Box 5018
Denver, CO 80217
                                              VIII

-------
Mr. F.W. Dawson
Director of Research
Millipore Corp.
17 Cherokee St.
Bedford, MA 01730

Mr. Frank  Dazzo
Microbiologist
Dept. of Soil Microbiology
University of Florida
2169McCarty Hall
Gainesville, FL 32611


Mr. Rodney S. Dehan
Microbiologist
Dept. of Pollution Control
Tallashassee, FL

Mr. R.M.Dille
Supervisor
Texaco
P. 0. Box 26747
Richmond, VA 23234


Dr. Alfred P. Dufour
Microbiologist
EPA
N.M.W.Q.L. Liberty Lane
W. Kingston, Rl 02892

Mr. Bernard J. Dutka
Head Microbiology
Canada Center of Inland Water
3450 Spruce Ave.
Burlington, Ont.,
Canada
Mr. Theodore A. Ehlke
Microbiologist
U.S. Geological Survey
P. 0. Box 1350
Albany, NY 12201

Mr. Karl M. Fox
Consultant
Research Service
300 Yale Ave.
Swarthmore, PA 19081
Mr. Thomas J. Furdek
Chief Chemist
USA Corps of Engineers
210 N. 12th St.
St.  Louis, MO 64501
Mr. S.L. Furman
7035 Commerce
Pleasanton, CA 94566

Mr. Edwin E. Geldreich
Research Director
Microbiological Control in
Water Supplies
US EPA
Cincinnati, OH

Mr. Ralph E. Gentry
Microbiologist
EPA, Region IV, S & A Division
College Station Road
Athens, GA 30601


Mr. Walter Ginsberg
Chief, Water Bacteriology
Chicago Bureau of Water
1000 E. Ohio St.
Chicago,  IL 60011

Mr. Norman H. Goddard
Export Area Manager
Sartorius Membranfilter
34 Gottingen
Weender, Landstr 96-102
W. Germany
Mr. Rosario J. Grasso
Sanitary Biologist
Comm. of Mass. D.P.H.
Lawrence Exp. Sta.
Lawrence, MA 01842

Ms. Barbara Green
Research Asst.
Univ. of Massachusetts
Dept. Envir. Sci.
Marshall Hall
Amherst, MA 01002

Mr. Phillip E. Greeson
Hydrologist
U.S. Geological Survey
National Center, MS 412
Reston, VA 22092
Dr. Edward F. Gritsavage
Microbiologist
U.S. EPA
240 Highland Ave.
Needham, MA02194
Dr. Leonard J. Guarraia
Microbiologist
U.S. EPA
401 M St., N.W.
Washington, DC 20460

Mr. Frederick L. Harris
Microbiologist
EPA
25 Funston Rd.
Kansas City, KS66115

Dr. Walter Harris
President
Med-Ox Chemicals Ltd.
145 Bentley Avenue
Ottawa, Ont., Canada

Ms. Patricia Haynes
Microbiologist
Carborundum Co.
R& D
Buffalo Ave.
Niagara Falls, NY 19302

Dr. Charles W. Hendricks
Microbiologist
EPA, WSD,WSME 1011
2005  Kenley Ct.
Alexandria, VA 20308

Dr. Alfred W. Hoadley
Assoc. Prof.
GA. Institute of Technology
School of Civil Engineering
Atlanta, GA 30332

Mr. Lyman Howe
Research Chemist
TVA Water Quality Branch
150-401 Bldg.
Chattanooga, TN 37402
Dr. James B. Hufham
Asst.  Prof, of Life Science
University of Missouri
Rolla, MO 65401

Mr. Lothar Jeschke
Schleicher & Schuell
28 Greenbriar Road
Keene, NH 03431

Mr. Wayne Johnson
Chemist
Lee County
813 Dellena  Ln.
Ft. Myers, FL 33905
                                               IX

-------
Mr. Ernest Karvelis
Supervisory Aquatic Biologist
U.S. EPA-Cincinnati
5555 Ridge Ave.
Cincinnati, OH 45268

Ms. Harriet Kennedy
Laboratory Director
Evanston N.S. Health Dept.
1806 Maple Ave.
Evanston, IL 60204

Mr. Virgil Kessinger
Chief, Oper. Water Plant
Palm Springs Util.
226 Cypress Lane
Palm Springs, FL 33460

Mr. Robert T. Kirkland, Jr.
Chief, QW Service Unit
U.S. Geological Survey, WRD
244 Federal Bldg.
Ocala, FL 32670


Mr. William L. Klein
Manager, Surveillance Operations
ORSANCO
414 Walnut St.
Cincinnati, OH 45202

Mr. Edward C. Kosakoski
Microbiologist
Miami Reg. Lab.
1350 N.W.  14th St.
Miami, FL 33134


Mr. A. L. Lane
Director of Laboratories
Difco Laboratories
920 Henry St.
Detroit, Ml 48201

Mr. Edwin W. Lard
NSRDC Annapolis Lab.
U.S. Gov't Code 286
12703 Beaverdale La.
Annapolis, MD 21402
Dr. Morris Levin
Microbiologist
EPA
National Marine Water Qual. Lab.
Liberty Lane
W.  Kingston, Rl 02892
Mr. Stan Liebaert
Technical Representative
Gelman Inst. Co.
600 S. Wagner
Ann Arbor, Ml 48106

Dr. Shundar Lin
Professional Scientist
III. State Water Survey
P.O. Box 717
Peoria, IL 61601

Dr. Warren Litsky
Professor
Institute of Agricultural and
Industrial Microbiology
Marshall Hall
Univ. of Mass.
Amherst, MA 01002

Mr. J.P. Lively
Chief,  Laboratory Operations
Division
Environment Canada
Ottawa, Ont., Canada

Mr. John Loft
Products Manager
Amerace
10 Linda Lane
Summit, NJ 07901

Mr. Robert M. Lollar
Director of Environmental Affairs
Tanners Council
216 Eastern Ave.
Clarendon Hill, IL 60514

Dr. W.IM.Mack
Professor
Inst. Water Research
Michigan State  University
Dearborn, Ml 48104

Mr. Jim Marshall
Manager, Membrane R & D
G,elman Inst. Co.
600 S. Wagner
Ann Arbor, Ml 48106

Dr. Maria T. Martins
Senior Biologist
CETESB
Av, Prof. Frederico Hermann Je 465
S.,Paulo, S.P.,
Brazil
Dr. Gordon McFeters
Assoc. Prof, of Microbiol.
Microbiology Dept.
Montana State  University
Bozeman, MT59715

Mr. Gene Medley
Biologist
State of Florida, Dept. of Poll.
Control
1773 Pine Ave.
Winter Park, FL 32789


Mr. Stephen Megregian
Director, Water Quality Programs
Wapora Inc.
6900 Wisconsin Ave., N.W.
Washington, DC 20015


Mr. Amar S. Menan
Bacteriologist
Dept. of Environment
P.O. Box 2406
Halifax, Nova Scotia,
Canada

Mr. Gerald  Moore
Med-Ox Chemicals Ltd.
145 Bentley Ave.
Ottawa, Ont., Canada


Dr. Bernard Newman
Grad. Dept. of Marine Science
C.W.  Post College
 Long Island University
Greenvale, NY 11548

Mr. William A. O'Connor
Laboratory Director
Serco Laboratories
2982 Cleveland Ave.
Roseville, MN  55113


Mr. Al Pendleton
Hydrologist
U.S. Geological Survey
USGS National Center
Reston, VA 22092


Ms. Priscilla Positano
Chemist
3588 N.W. 27th St.
 Lauderdale Lakes, FL 33311

-------
Mr. Vincent Positano
Chief Chemist
Fort Lauderdale Utilities
3588 N.W. 27th St.
Lauderdale Lakes, FL33311

Dr. David A. Power
Manager, Technical Service
BioQuest
214 Burning Tree Rd.
Timonium, MD 21093


Mr. Robert Pratt
Director
Palm Springs Utilities
226 Cypress Lane
Palm Springs, FL 33460


Mr. Maynard W. Presnell
Research Microbiologist
USPHS-FDA
P.O.Box 158
Dauphin Island, AL 36528

Mr. William G. Presswood
Microbiologist
Tenn. Valley Authority
150-401 Chestnut St.
Chattanooga, TN 37402


Mr. Philip B. Reed
Manager, Ionics — Lyo Products
Ionics, Inc.
65 Grove St.
Watertown, MA 02172


Mr. Louis A. Resi
Microbiologist
EPA
5555 Ridge Ave.
Cincinnati, OH 45268

Ms. Cynthia Root
Microbioiogist
Gelman Instrument Co.
600 S. Wagner
Ann Arbor, Ml 48106
Dr. Donald F. Rothwell
Professor — Soil Services
University of Florida
2169McCarty Hall
Gainesville, FL 32601
Mr. Augustus Ruser
Microbiologist
Fla. State Div. of Health
P.O. Box 210
Jacksonville, FL 33033

Mr. Dave Rusnell
Environmental Specialist
Fla. Dept. Pollution Control
3201 Golf Course Blvd.
PuntaGorda, FL 33950

Dr. John E. Schillinger
Research Assoc.
Mont. State Univ.
816 N. 17th
Bozeman, MT59715

Mr. John W. Scottie
Assistant Director
Margate Utility Authority Inc.
6700 N.W. 6th Court
Margate, FL 33063

Mr. Irving Seidenberg
Chief, Microbiology, Region II
U.S. EPA
6 Sterling Ct.
E. Brunswick, NJ 08816

Mr. Karl V. Shallenberger
Biologist
City of Ft. Lauderdale
290 N.E. 40th St.
Ft. Lauderdale, FL33315

Mr. K. L.Shull
PEAC
320 N.Cleveland
Chagrin Falls, OH 44022

Mr. Joseph E. Sims
Dir. Membrane Development
Helena Laboratories
4180 Kenneth
Beaumont, TX 77705

Mr. William J. Stang
Senior Microbiologist
EPA, NF 1C-Denver
10626 W. 7th Ave., Apt. 102
Lakewood,C080215

Dr. David G. Stuart
Associate Professor
Montana State University
Dept. of Microbiology
Bozeman, MT 59715
Mr. Harry C. Torno
Staff Engineer
U.S. EPA
Office of R & D (RD678)
Washington, DC 20460

Mr. Ben Trasen, AW
Marketing Mgr.
Johns—Manville
Greenwood Plaza
Denver, CO 80120

Mr. Albert D. Venosa
Research Microbiologist
U.S. EPA
National Environment Res. Center
Cincinnati, OH 45268

Mr. L.T. Vlassoff
Manager
Min. of Env. Canada
Box 213
Rexdale, Ont., Canada

Mr. Gene T. Waggy
Union Carbide
P.O. Box 8361
So. Charleston, WV 25303

Mr. Russell West
Operator
Margate Utility Auth. Inc.
924 Pine Ridge Dr.
Plantation,  FL 33317

Dr. Ted Williams
Chief, San. Bact. & Chemistry Sect.
Michigan Dept. Public Health
3500 N. Logan
Lansing, Ml 48914

Mr. John Winter
Chief, Qale Br., MDQARL
U.S. EPA
NERC-CINTI
Cincinnati, OH 45202

Mr. Gilman Wommach
Director of Lab. Water Quality Program
Missouri Dept. of Natural Resources
1407  Rehagen
Jefferson City, MO 65101

Mr. Charles C. Wright
Senior Technical Advisor
ARAMCO
1539 W. 16th
Long Beach, CA90813
                                              XI

-------

-------
 A.   Total coliforms
                                                 B.   Fecal coliforms
C.   Fecal streptococci
           Plate 1.    Recovery of Indicator Microorganisms
                      by Membrane Filter Methods.

-------
A.   Fecal coliform and non-coliform colonies
                                                         B.    Fecal coliform colony showing
                                                              crystalline structure
     C.   Total coliform sheen colony
                                                          D.   Fecal streptococci colonies

              Plate 2.    Close-ups of Indicator Microorganisms on Membrane Filters.

                                               2

-------
           A.   Poor distribution
  B.   Leaking filter assembly
         C.   Non-wetting area
D.   Excessive turbidity and confluency
      E.   High background count
       F.   Wrinkled membrane
Plate 3.    Problems in Recovery of Indicator Microorganisms as Shown in the Total Coliform Test.

-------
           A.   Problems in recognition
              C.   Turbidity effects
    B.   Poor colony definition
D.   Channelling, poor distribution
     and overcrowding
                         E & F.     Normal recovery and effects of stress
Plate 4.   Problems in Recovery of Indicator Microorganisms as Shown in the Fecal Coliform MF Test.

-------
                                           WELCOME

                                        Robert H. Bordner
     On behalf of the American Society for Test-
ing and  Materials  and the  U.S.  Environmental
Protection  Agency, we  warmly  welcome you to
ASTM Committee  D-19  on Water, and  to the
Symposium on the Recovery of Indicator Organ-
isms Employing Membrane  Filters. We have esti-
mated  the  number of  attendees to be 105. This
large attendance reflects the high  interest  in the
use of membrane filters to recover coliforms and
other bacteria from water.

     ASTM Committee D-19 has been sojourning
to Florida  every  January  for ten  years and last
year the D-19  meeting attracted about 200 enthu-
siastic  members to Fort  Lauderdale.  Committee
D-19 has grown rapidly from the original subcom-
mittees for organic substances,  metals and  inor-
ganic constituents  in  water  and   now includes
many other test areas, such as oil identification,
sediment  chemistry,  automated  analyses,  and
biological methods. This year the total attendance
of the D-19  meeting  has doubled, reflecting the
increasing activity  in biological and microbiologi-
cal monitoring.

     The subject of this symposium falls under the
Subcommittee D-19.08 on Membrane and Ion Ex-
change Materials. A task group of this subcommit-
tee was formed to develop test procedures for eval-
uation of membrane filter materials. It is appro-
priate  that  this symposium  be held  under the
sponsorship  of ASTM  as well as EPA, because
of the Society's almost-unique structure that brings
together  representatives  of  the  manufacturers,
users, regulatory  agencies and research  workers
to attack problems of common concern.

-------
                                            SUMMARY
     The  results of  the studies  of  membranes
described  in this symposium cannot be compared
directly because  of  non-uniform test  conditions.
However,  we can reach the general conclusion that
variable recoveries  of coliforms,  fecal coliforms
and  fecal  streptococci occur with testing of water
and  wastewater  samples  by membrane filtration
(MF), further, that  low counts result from  injury
caused by natural stream conditions, chlorination,
the elevated  incubation  temperature of the fecal
coliform test and the structure of 0.45 iim mem-
brane filters.
     The effects of stress on recovery are confused
by the variable and unpredictable low recoveries of
indicator organisms from different lots and brands
of  membrane  filters.  Because most laboratories
are not conducting routine quality control checks
on  materials, media, equipment and methodology
as part of  a  within-laboratory  QC  program, are
not  analyzing  enough split or replicate samples,
and  are not verifying sufficient test results to in-
sure the validity of their data, further discrepancies
occur in comparative data.
     It was also obvious from discussion that the
poor communications which  exist between manu-
facturers and users contribute  to the problem.

     Proposed solutions to the  problem of  low
recovery of  indicator  bacteria  on  MFs were:

     1.   Incubation of filtered samples on a  non-
         selective medium prior to transfer  and
         incubation  on  the selective medium.

     2.   Short  term incubation  of filtered  sam-
         ples on a  selective medium at 35 C prior
         to incubation at 44.5 C.
     3.
     4.
                                                        5.
Short term incubation of filtered sample
on  a  non-selective  medium  at  35  C
prior  to transfer  and  incubation  on a
selective medium at 44.5 C.

Use of an overlay of a non-selective agar
on  a  base  layer selective agar for  short
term  incubation at 35C prior to incuba-
tion at44.5C.
         Use of
         filter
         standard filter.
        a larger surface-pore membrane
      in  place  of the  0.45 Aim  pore

-------
                                      RECOMMENDATIONS
To the ASTM Subcommittees

     Develop MF collaborative testing procedures
for physical,  chemical  and microbiological char-
acteristics that include tightly written protocols,
which  control test variables and include randomi-
zation   and  standardized  statistical  evaluation.
     Develop a  collaborative testing  program for
comparison of methods for indicator organisms.
     Develop a uniform test protocol for compar-
ing membranes, media and other test conditions.
Until this is done it will be impossible to evaluate
and  select improved  methodology  for indicator
bacteria.
To Membrane Filter Users

     Establish a quality assurance program within
the laboratory for supplies, equipment, and analy-
ses. Such  control on membranes, media, and test
conditions will assure valid data.

     Follow standard test protocols developed by
EPA/ASTM  committees  for future within-labora-
tory or interlaboratory studies.

     Support ASTM, EPA and other testing groups
in evaluations  of improved MF procedures using
the standard test protocols.
To Manufacturers

     Expand and improve the quality control  of
membrane filters, media, reagents, and equipment
used in MF tests.

     Encourage the  certification of specific pro-
ducts for water analysis.

     Improve communications with users. Provide
them with  necessary  information  on pore  size,
configuration, additives (extractables), membrane
materials, manufacturing dates of media, changes
of formulations etc.

     Establish a voluntary program to notify users
about unacceptable  lots and recall  such  products
if necessary.

To EPA

     Develop a test  protocol  for the comparison
of test methods.

To Researchers

     Investigate  the  physiological  basis  for en-
vironmentally-stressed  cells  and apply the results
to improved MF methods.

     Solve the  problem  of  testing  chlorinated
effluents by MFs.

     Define MF specifications for a) a completely
inert, or b) highest count membranes.

-------
                               THE MEMBRANE FILTER DILEMMA

                                        Robert H. Bordner

                               Chief, Microbiological Methods Section
                          Environmental Monitoring and Support Laboratory
                                   Environmental Research Center
                       U.S. Environmental  Protection Agency, Cincinnati, Ohio
                 ABSTRACT

     Reported variances  in recovery of indicator
organisms, symposium objectives and requirements
for improvements  in  membrane filter  (MF) pro-
cedures are  described.  The  applications  of  MF
methods to  water  quality  control and  recent
enforcement  legislation are reviewed. The  import-
ance of quality assurance procedures and the role
of The American Society for Testing and Materials
in  the  development  of  standardization test
methods for water are discussed.

               INTRODUCTION

     The dilemma  that we face here today is that
for the past year and a half  the membrane filter
has been the target of serious charges concerning
its inability to adequately recover indicator organ-
isms from  water  and wastewater.  Conflicting  re-
ports in the literature have documented the differ-
ences in recovery among various membrane filter
brands. Some laboratories have reported variations
in membranes and  media from lot to lot. Other in-
vestigators  have encountered  low recoveries when
using  the  membranes  to  enumerate  indicator
groups  from marine  waters  and chlorinated  ef-
fluents, or those containing toxic materials.

     The  requirements  to enforce and  monitor
water  and wastewater  standards recently estab-
lished  by  the  U.  S. Environmental  Protection
Agency have  greatly accentuated  the  need for
precise  standard  procedures and reliable, uniform
test materials.

           SYMPOSIUM OBJECTIVES

     This  symposium was organized to focus on
the  real ability  of  membrane filters  to  recover
indicator  organisms.  Much consideration  will  be
given  to  fecal  coliforms, because  they  are  the
indicator group of primary concern. Also, recovery
problems  are intensified at the elevated tempera-
ture of the fecal coliform test.

    The  ultimate  objectives of this  symposium
are to:

     1.    Identify   problem   areas  and  future
          needs for the use of membrane filters.

    2.    Review  differences in the recovery of
          microorganisms and investigate the cause
          of these differences.

    3.    Determine  the  factors  that affect  re-
          covery.

    4.    Define the type(s)  of filters required for
          water analysis.

    5.    Develop test  procedures for the evalua-
          tion  of  membrane  filters  that   will
          assure quality control.
     THE MEMBRANE FILTER PROBLEM

     The membrane filter, first introduced in this
country about 25 years ago, has developed over the
years into  an estimated 5 million  dollar a year
industry, and is  widely accepted by water micro-
biologists  for many different  tests. The micro-
biologist has  learned to appreciate the advantages
of membrane filter procedures: the rapidity, ex-
pediency of direct counts, ease of testing, minimal
space and labor  requirements, ability to  examine
large  sample volumes  and  portability  for  field
testing. He  must now consider which  brand of
filter to use or,  indeed, whether to use the mem-
brane filter at all  to obtain good  recovery.

-------
     The manufacturers know the needs and con-
cerns of the users, and have taken a close look at
their product, conducted their own investigations,
and  in  some cases  modified membrane formula-
tions in an  effort to improve the filters for water
analysis. They  are becoming more aware  of the
importance  of good  quality control of  membrane
filter materials.
     There  are  several questions  related to  the
membrane filter dilemma that we should consider
during this symposium:

     1.   What degree  of  recovery is required? Is
         the  ultimate goal the recovery of  all
         organisms that  will grow under one set
         of test conditions for  a given  parameter
         (e.g.,  specific membrane, medium, and
         incubation temperature), or  is  it  the
         recovery  of  all  viable  organisms,  in-
         cluding attenuated  organisms,  to pro-
         duce  a result as close  as possible to the
         true count?

     2.   What type of membrane filter is neces-
         sary?  Do  we need a filter that is com-
         pletely inert and  does not in itself  af-
         fect growth  and recovery,  or one that
         does  not contain nutrients, materials or
         inhibitory  materials? Should  a filter  be
         considered that enhances recovery and
         growth by  releasing  soluble materials
         which stabilize pH or  have a beneficial
         buffering effect on the medium? Manu-
         facturers point out  that such materials
         are available.

     3.   Will membranes provide good recovery
         under prescribed test conditions for one
         indicator but not  provide equally good
         recovery  for other   indicators  under
         different conditions?  Is it  desirable  to
         design  a  membrane  material for  the
         optimal recovery of each  specific  indi-
         cator group?

     4.   What  is the relationship of  the  mem-
         brane  to  the  underlying   media? For
         example,  is  recovery  affected by agar
         or broth-saturated pad  substrates?

     5.   Does  the  method  of sterilization affect
         the ability of the  membrane  to recover
         organisms, as recently reported?

     6.   How  can  better recovery   be provided
         for problem samples such as marine and
         estuarine waters, chlorinated  effluents,
         and  wastewaters  containing   phenols,
         metals   or   other  toxic  compounds?

    7.   What  product  specifications  or limits
         are  critical  for membrane  filter mater-
         ials?

    8.   What is the real shelf  life of the mem-
         brane filter?

    9.   What quality assurance does the manu-
         facturer perform?

    10.  What quality  control  procedures  must
         the  laboratory  carry out on membrane
         filters as well as other materials, such as
         the  media,  reagents,  distilled water and
         other supplies?

    11.  What are the microbiological and chemi-
         cal  characteristics of  membrane  filter
         materials for which  ASTM  subcommit-
         tees  should develop practical  test  pro-
         cedures,  for  example,  recovery, inhibi-
         tory effects, retention and extractables?

    12.  What statistical measurements should be
         used  uniformly  so  that  membrane
         filter tests can be compared fairly among
         laboratories?


   USES OF MICROBIOLOGICAL ANALYSES

     Microbiological  analyses of water and waste-
water are conducted in order to:

     1.   Assure  the  quality of potable water  at
         the water  treatment plant and in the
         distribution system,  raw water  sources,
         ground water, and bottled water.

     2.   Determine  the quality of water for
         recreational,  argicultural, irrigation, in-
         dustrial, shellfish-raising, and other uses.

     3.   Investigate   the  quality  of  municipal
         and  industrial  wastewaters,  and  the
         effectiveness of treatment.

     4.   Plan and develop water resources.

     5.   Perform in-plant studies.

    6.   Identify the source or trace the disposal
         of bacterial pollutants.

-------
     7.   Carry out research investigations.

     8.   Monitor and enforce established stand-
         ards for  wastewater effluents and  re-
         ceiving streams.

         WATER AND WASTEWATER
         STANDARDS AND CRITERIA

     The requirement for monitoring  and enforc-
ing water quality standards is one of the most com-
pelling  reasons for developing standard  methods
and uniform materials.

     The standards for potable water quality were
provided for  in  the  Safe  Drinking  Water  Act
(Public  Law  93-523),  dated December 16,  1974
(1). Within 90 days after enactment, the maximum
allowable levels  of constituents  should  be pub-
lished in the  Federal  Register as EPA standards.
The MPN or  MF procedure may be used to moni-
tor these limits.  For  the  membrane filter  tech-
nique, the quality limit is one total coliform per
100 ml  and the  action limit is more than 4 total
coliforms per 100 ml. The minimum action  re-
quired is immediate  repeat sampling. The volume
sampled  by the MF technique must be 100 ml. The
minimum number of samples collected each month
is  based  on the population served by the supply.
A  standard plate  count incubated at 35 C for 48
hours is recommended.

     The suggested  criteria  for  natural waters
were  spelled  out  in  the  Water Quality  Criteria,
1972 (2). This  updated  volume of the original
Water Quality Criteria  published in  1968 (3), was
developed for EPA by the National Science Foun-
dation.  A supplemental document will be forth-
coming  to  support the  enforcement of these
criteria.  The  recommended criterion  for recrea-
tional water is 200 fecal coliforms per 100 ml. The
criteria  for  shellfish-raising waters are 70  total
coliforms and  14  fecal  coliforms per  100 ml.

     The Federal  Water   Pollution Control  Act
Amendments  of  1972 (Public  Law 92-500) (4)
established guidelines for  the levels of constituents
in  municipal and industrial effluents. The monitor-
ing of these standards,  under the new national per-
mit  system,  is  entitled  National Pollutant  Dis-
charge Elimination  System, (NPDES). The  mini-
mum level of effluent quality attainable by secon-
dary treatment for fecal coliforms was described in
the Federal Register, August 17, 1973 (5). The Act
states that  the geometric mean  of the fecal coli-
form  value for effluent  samples  collected over a
period of 30 consecutive days shall not exceed 200
per  100 ml  for  fecal coliforms. The geometric
mean for 7 consecutive days shall not exceed 400
per 100ml.

     The microbiological guideline  for industrial
effluents from  the food  processing, textile,  feed-
lot, meat  products, tanning and sugar processing
industries  is a maximum  fecal  coliform value not
to exceed 400 counts per  100 ml at any time.

            QUALITY ASSURANCE

     One  of  the primary  responsibilities of the
microbiology laboratory  is  the adoption  of a for-
mal  quality  control  program  to assure  the reli-
ability and validity of laboratory and field  data.
The quality assurance program includes the syste-
matic practice of accepted sampling  and analytical
procedures described  in  Standard Methods  or in
the forthcoming  EPA manual  on microbiological
methods. Quality assurance by trained laboratory
and field personnel and application of good quality
control over materials,   equipment,  instrumenta-
tion analyses and the resultant data are required.

     Assured  validity  of  results  is  particularly
important  if the data are  used  in  court, in the
exchange  or  compilation  of data  from  other
laboratories,  or entered in a data storage  bank for
other users.

    MEMBRANE FILTER SPECIFICATIONS

     In 1965 the Department of Defense, Office
of Medical Materiel, developed a detailed set of
interim specifications to control the  quality of
membrane  filter  materials.  Test  procedures  were
described  for  characteristics  such  as  recovery,
toxicity,  retention,  flow  rate,  porosity,  pore
size,  and  extractables. These  specifications  have
been  superceded  by  Military Specifications dated
September, 1973  (6).

DEVELOPMENT OF ASTM TEST PROCEDURES

     The American Society for Testing and Mater-
ials is directly concerned with the quality control
of materials such as membrane filters because of its
fundamental  interest in the development  of stand-
ards  of characteristics and performance of ma-
terials and the promotion  of this knowledge.

     ASTM   provides  a  unique  system  whereby
experts from several fields can  get together and
develop, evaluate and approve test procedures by
group concensus. ASTM  publishes the procedures
                                               10

-------
using an established protocol and a standard for-
mat. If an evaluation process is required to charac-
terize  materials available from various manufac-
turers, to determine that materials are uniform, or
to know that purchase from random sources will
give  acceptable results, the mechanism is available
through ASTM. This  is the real  reason  for this
symposium.  The test  procedures for membrane
filters  are  being developed through this  system.

     To develop a test  procedure,  a  task group
chairman solicits proposed  methods. These draft
methods are  circulated for review until the chair-
man is satisfied that he  has a consensus of agree-
ment on a proposed test procedure. This procedure
is then tested by volunteer laboratories in  a round
robin  procedure if  it is amenable to this  type  of
collaborative  testing.

     Under ASTM  rules, it is  necessary  to have
supporting  data to show that  the method is ac-
ceptable.  Statistically  significant   results  must
demonstrate the desired precision, and, if possible,
accuracy. The  proposed  test  procedure  is then
submitted to  the main  D-19 committee for appro-
val  by  ballot.  If the procedure  is approved and
negative votes,  if any,  have  been  resolved, the
method  is  published in  the ASTM  manual as  an
official procedure with  tentative status.

ASTM TASK GROUP ON MEMBRANE FILTERS

     The  present  subcommittee  task  groups  on
microbiological  properties and  physical-chemical
characteristics of membranes originated  in June,
1971   with   subcommittee  D-19.08, which  was
organized to  investigate test procedures for mem-
branes used  for separating processes such as ion
exchange,  electrodialysis,  reverse  osmosis,  and
ultrafiltration. At the initial meeting of this group
much  confusion  arose over the basic question,
"What  is a  membrane filter?"  It soon  became
apparent that the need  for controlled pore size
filters  varied  with the  use. Because  membrane
characteristics are related to their application, sub-
sections  were  developed by  use under  specific
categories.

     The present session of this particular sub-
committee  that is developing test procedures for
membrane filter materials used  in water analyses is
the third since the organizational  meeting  in June,
1973. The development of test  methods has pro-
ceeded  through the early discussion and  review
stages, but not without growing pains. Test pro-
cedures for recovery  and inhibitory effects were
submitted  to   the  subcommittee.  Preliminary
round  robin tests were carried out on 6 brands of
filters  by  10  volunteer water  laboratories.  The
statistical  results indicate  a need for a better test
procedure. A detailed  report on this procedure
and the round  robin test  results will be presented
later at this symposium.

     The present status of ASTM test procedures
for the microbiological  properties of membrane
filter materials  is:  1. A preliminary procedure for
recovery has been developed but  requires modifica-
tion. 2. A  modified test procedure for inhibitory
effects is  ready for resubmission to the subcom-
mittee. 3.  A proposed procedure for retention  has
been prepared for task group discussion.

    The conflicting and confusing reports on  the
recovery of indicator  organisms  and the lack of a
test procedure for recovery precipitated the organi-
zation  of  this symposium to  provide  a forum for
the in-depth examination of these problems.
                 REFERENCES

 1.   Safe Drinking Water Act, Public  Law 93-523,
     December 16, 1974, 88 Stat. 1660. 42 United
     States Code (USC) 300f.
 2.   Water Quality Criteria, 1972, EPA-R3-73-033,
     December,  1973. Office of Research and De-
     velopment, EPA, Washington, D.C.
 3.   Water Quality Criteria. Report of the National
     Technical Advisory Committee to  the Secre-
     tary  of the  Interior, April 1,  1968. Federal
     Water   Pollution  Control  Administration.
     Washington, D.C.
 4.   Federal Water Pollution Control  Act Amend-
     ments  of 1972, Public Law  92-500, October
     19, 1972, 86  Stat.  816, 33  United States
     Code (USC) Sec. 1151.
 5.   Water Programs, Secondary Treatment Infor-
     mation, EPA Federal Register  CFR 40,  Part
     133, August  17, 1973.
6.   Military Specification, Disk,  Filtering, Micro-
     porous, MIL-D-37005   (DSA-DM),  Defense
     Personnel Support Center, DPSC-ATT, Phila-
     delphia, Pa., September 5, 1973.
                                               11

-------
              PERFORMANCE VARIABILITY OF MEMBRANE FILTER PROCEDURES

                        Edwin E. Geldreich, Microbiological Research Director
                                 Water Supply Research Laboratory
                              National Environmental Research Center
                               U.S. Environmental Protection Agency
                                          Cincinnati, Ohio
                  ABSTRACT

     Performance  variability in membrane filter
procedures can be traced to a variety of  factors
including  variations  in  membrane  filter  (MF)
manufacture, absorbent pad impurities, MF steri-
lization procedures, commercial  media  inconsis-
tencies, and technician knowledge, skill and judge-
ment in applying MF procedures to water analyses.
The acceptability of special MF devices in terms of
data reliability is also reviewed.

               INTRODUCTION

     Many of the problems to be identified in this
presentation on the membrane filter dilemma can
be traced  to technological developments that have
failed to recognize the critical  requirements asso-
ciated with  microbiological  applications.  As  a
consequence, a variety of factors contribute to the
overall  problem  whose magnitude threatens the
future of  this valuable microbiological tool. Prior
to cataloguing these  factors,  a brief historical
resume should place these issues in better perspec-
tive and hopefully point to directions that both the
manufacturers and laboratories must take to  re-
store  confidence  in membrane filter procedures.

Historical  Development of the Membrane Filter

     The  initial  attempt to develop  an  artificial
membrane as a substitute for those found  in nature
has been credited to Fick in 1855 (1-3). However,
Fick  experienced  difficulties in using these fragile
collodion  membranes  and eventually abandoned
the idea.  Once the conceptual design of a collo-
dion  sac  for dialysis  became apparent, numerous
applications of this membrane form were reported
during the period  of  1893 and 1905 (1). Soon it
was discovered that the mixture and concentration
of  alcohol-ether  and  glacial  acetic  acid  solvent
systems would change the  porosity  in the nitro-
cellulose during evaporation and that the resulting
permeability  could be  preserved  by placing  the
forming membranes in water before solvent evapor-
ation was completed (4-8). Glycerol, soluble in the
alcohol-ether  mixture  but  insoluble  in  nitro-
cellulose, was found  to increase membrane per-
meability (9). Improved flexibility of the finished
membrane was achieved by the addition of  a small
concentration of castor oil. From these pioneering
investigations  it  became evident as early as 1915
that: a) the control of porosity was the key to the
successful development of nitro-cellulose  mem-
branes;  b) reproducibility of pore size was easier
in membrane sheets than in the production of a sac
configuration and; c)  there was a need for  careful
control  of production methods and exact concen-
tration  of ingredients used in the preparation of
membranes (10).

    By  1916 the usefulness  of a  nitro-cellulose
membrane as a tool in bacteriology had been recog-
nized.  Removal  of bacteria from any reasonable
quantity of  fluid and  cultivation  of the  micro-
organisms in place on a membrane surface had long
been a goal  for  many investigators. However, the
usefulness  of such   membranes was extremely
restricted  because their preparation was difficult
and the product was of uncertain quality and por-
osity.  Zsigmondy  and Beckman were the first to
develop a method for the  preparation of a mem-
brane that could readily be adapted to commercial
production and  they  were issued a United States
patent in 1922  based on an application submitted
in 1919  (11-14). This procedure consisted  of dis-
solving nitro-cellulose or nitro-cellulose  acetate in
a mixture of acetone and glacial acetic acid, pour-
ing the solution  on a glass plate in a thin layer;
allowing the volatile solvents to evaporate at 18 C
in a 60 percent relative humidity; and washing the
finished membrane in water.  Pores of  a specific
size range were  produced by controlling the con-
centration of nitro-celhjlose in the basic solution,
                                                12

-------
composition   of  solvent  mixture  and   relative
humidity  during  solvent evaporation.  Increasing
the percent of water in the solvent mixture or ele-
vating the percent  relative  humidity during the
drying process resulted in larger pore sizes.

     The  commercial  manufacturing  technology
for membrane filters remained basically unaltered
from the  Zsigmondy process until the post World
War II period. From a review of German military
intelligence gained during a scientific  reconnais-
sance after the war,  Dr. Alexander Goetz prepared
a complete report on  the manufacturing process,
characteristic  properties and  bacteriological appli-
cations of the Zsigmondy membrane filter  (15).
Research   into  production  techniques  for  the
U.S. Army Chemical Corps led Dr. Goetz to de-
velop a membrane filter that did not require stor-
age in water, to  incorporate a grid imprint, and to
design an  associated apparatus  for  use  in filtering
water. This  improved  processing was  then  con-
tracted to the Lovell Chemical  Corporation which
later organized the Millipore Filter Corporation for
American  production of the  improved filter made
from domestic  materials.  Numerous refinements
relative to automation of  the  process  have  since
been  made to improve the uniformity and quality
of the product  in a competitive market that  in-
cludes several other  American  membrane  filter
manufacturers:  Gelman, Schleicher and  Schuell,
Nuclepore, Helena, and Johns-Manville. Among the
foreign manufacturers, Oxoid  (British) and Sar-
torius (German)  are  the most available  mem-
branes in this country.

     Application of the nitro-cellulose  membrane
developed  simultanously  with  the  technological
advances  in  membrane filter manufacture. How-
ever, application to bacteriological procedures was
hindered  in early years by the uncertain  quality
and  inadequate supply of the material.  In 1919, a
nitro-cellulose filter was first reported  to recover
organisms  of  tuberculosis  from  urine  (16). Ap-
parently,  the first  attempts  to culture  micro-
organisms  on  membrane  filters were done  in
Russia and Germany during the 1930's  (17-21).
Dr. Mueller (Hygienic Institute, Hamburg) adapted
these techniques to the urgent needs for an ade-
quate monitoring of public water supplies and for
emergency  situations   resulting  from   the  war
devastations of Germany in the period of  1943 to
1945.  Routine analysis of potable waters for coli-
forms was accomplished by placing the membrane,
after filtration, on a substrate of seven filter papers
saturated with Endo broth (20).  Mueller was also
successful  with this type of procedure in her inves-
tigation of an epidemic of typhoid fever occurring
in Hamburg (21). Samples were filtered through a
nitro-cellulose  membrane filter which was then
placed on a modified  bismuth sulfite medium for
cultivation of typhoid organisms. The  interest in
membrane filter techniques  for the bacteriological
examination of water became widespread through-
out the  English speaking world following studies
in the nineteen-fifties  by Clark et al. (22,23) and
Goetz et al   (24, 25)   in the United States and
Taylor  and  Burman  (26, 27)  in  Great Britain.

     Analysis  of  data  available to  the U.S. EPA
laboratory evaluation program indicates that state
health,  state  environmental, city-county  health,
municipal water treatment and private laboratories
are examining approximately 3.5 million samples
annually  from this nation's public and  private
supplies  and  in gathering and monitoring data on
natural  waters relative to state and  federal  stand-
ards for  a variety of water  quality uses (28). An
estimated  one  million  additional   samples are
analyzed by  local laboratories in quality control
monitoring of industrial and municipal waste dis-
charges as required  in  the National  Pollution Dis-
charge Elimination  System. Of course, not all of
these analyses involve  the membrane filter  proce-
dure  but national statistics  suggest 52.1%  of all
laboratories currently  involved in the nationwide
laboratory  evaluation  program are  using  the
membrane filter  procedure  on a variety of water
samples  estimated to  be  1.8  million analyses per
year.  Thus, there is a substantial amount of moni-
toring data being developed from membrane filter
procedures which should be  reliably  measuring
water quality.

Membrane Filter Quality for Microbiology

     Commercial  brands of  membrane filters may
vary in performance as a result of manufacturing
technology,  materials,  and degree of quality con-
trol  exercised. For  microbiological  applications,
there  must be a complete retention of organisms
on or near the surface  of  a non toxic, inert matrix
which permits a continuous contact with nutrients
from  a medium held in a substrate below the mem-
brane. These  basic  conditions place demanding
requirements on  the quality  of every  membrane
used in the  laboratory. Basic difficulties encount-
ered with membrane filters generally relate to pore
distribution, hydrophobic filter areas, grid line ink
restrictions,   membrane  materials,   sterilization
practices, and poor storage  characteristics that
cause   increased  filter  brittleness  and  surface
warping (29-39).
                                                13

-------
     Membrane  filter pores should  be uniformly
distributed and  have a diameter of 0.45 micron
(± 0.02  micron) for routine bacteriological  tech-
niques.  Pores of some  commercial  lots of mem-
brane filters  have been  found to be so small  in
some areas of the filter that serious local reduction
in the flow rate  occurs. The filter should be free of
visible non-porous areas which prevent the diffu-
sion   of  nutrients  to the  upper surface of the
membrane. Any  bacterial cells entrapped on such
surfaces  will  not develop into visible colonies for
lack  of nutrients. When M-Endo is used in a test of
diffusibility,  non-wetting areas on the filter will
remain white and dry. Such  observations should
not  be  confused with air bubbles,  which can be
removed  by   reseating  the membrane  over the
medium-saturated pad or agar base. At the other
extreme, pores larger than 0.7 micron will not re-
tain  organisms  associated with  indicator  groups.
For  complete bacterial  separation  from  liquids,
membrane filter porosity of 0.22  micron is re-
quired to insure retention of the smallest bacteria
through  physical impingement  or  electro-static
entrapment.

     The ink  used to imprint the grid system on
the membrane filter should be non-toxic to all bac-
teria  cultivated  on  the filter' surface (30). Some
inks   have been found  to be  bacteriostatic or
bactericidal.   Such   effects  can  be  recognized
through  restrictive  colony development adjacent
to the imprinted lines. These growth restrictions
may not only be caused  by inhibition  from  toxic
inks  but also from  thick ink  imprints  that "wall-
in" grid squares and by  hydrophobic inks which
prevent  nutrient diffusion to  sites  in  the ink
imprint.   As  an  additional  characteristic,  inks
selected  for  grid imprinting should not "bleed"
across the membrane surface after a 24 hour con-
tact  with any medium  normally  used at 44.5 C
incubation.  Heavy  imprinting of  the grid system
can also  result in a  network of "canal-like" inden-
tations  that  frequently  become filled with  con-
fluent growth.

     The physical   structure  of  the  membrane
filter material should be such  as to provide an
optimum retention  of bacteria on the surface with
little migration  to  areas  within  the pore matrix.
When surface penetration  occurs, growth  may be
limited  in development during the colony counting
procedure.
     Chemical composition of  membrane filters
has  largely been limited  to polymerized cellulose
esters since  membrane filter  technology  initially
developed in this  direction. Conventional media
designed for selective recovery of bacterial indica-
tor groups  or  pathogens using agar  pour plates,
streak plates or broth cultures had to be redesigned
to compensate  for the physical-chemical properties
characteristic of nitro-cellulose materials  (29, 40,
41). For example,  the selective adsorption of dyes
excluded the use of acid to neutral  dyes as indica-
tor systems and necessitated the use of increased
amounts of brilliant  green as a suppressive agent in
Kaufmann's Brillian  Green agar to  obtain the de-
sired suppression  of some of  the unwanted bac-
terial population. Similarly, various nutrients such
as tryptone, polypeptone and  proteose  peptone
No.  3  were found  to  be superior  in membrane
filter media than in the same media used originally
with peptone in their formulation.  The result has
been the  creation of a family of media designed
specifically for use with nitro-cellulose membrane
filter products. With these experiences  in  mind,
manufacturers should be careful about revising the
Goetz  membrane  filter  process. Changes involve
the risk  that recommended  media  may suddenly
become less sensitive or less selective. Some com-
pounds introduced  to  the membrane filter may
improve flexibility, flow rate or stabilize porosity.
However,  these substances  should  not become a
source  of  fermentable  carbohydrates that  cause
false colony differentiation, create pH shifts in the
indicator systems,  are selectively toxic for specific
organisms,   or  adversely  depress  the  selective
action  of  differential  media  by  providing  the
bacteria with a  highly nutritive organic compound.
In essence,  membrane filters should  remain inert
to  bacterial  reaction,  and  unchanged in  those
physical-chemical  characteristics that  effect  media
selectivity and sensitivity.
     Sterilization  of the membrane filter  is essen-
tial to  all applications involving filtration of liquids
for bacterial  removal or for use in bacterial cultiva-
tion. Prior to the development of the Goetz mem-
brane  filter process, membrane filters were steri-
lized in the laboratory by gentle boiling in distilled
water for 20 minutes and repeating the procedure
a  second time  with fresh  distilled  water  (15-29).
This  procedure  served the  double  purpose  of
sterilizing  the  membrane  and  of  extracting any
residual  toxic  substances.  In retrospect,  the con-
tinued use  of  this leaching and sterilization pro-
cedure would have avoided many of the variations
in  membrane  filter performance  now  evident.
However, the  procedure  does take  more time to
execute and  is  a recognized inconvenience in busy
laboratories examining 50 to 200 samples per day.
     Goetz proposed  the use of ethylene oxide
(0.5 ml per liter volume)  for 3 to 4 hours at room
temperature in  a dessicator followed by air flushing
                                                14

-------
for several hours to remove the sterilizing gas resi-
dual.  This sterilization procedure is very effective
but requires a thorough flushing, preferably in a
vacuum  system at an  elevated  temperature for
several hours, with the procedure  being repeated
on  two  or three  succeeding days for complete
removal of gas entrapped in  pockets  in the mem-
brane matrix. As  a  safety precaution relative to
the explosive  nature of pure  ethylene oxide in  air
at certain  concentrations,  it is more desirable to
use  a mixture of  ethylene oxide  and  carbon
dioxide  to decreased explosive and flammable
properties.

     With   improvements  in   membrane  filter
technology, subsequent research demonstrated that
membrane filters (packed with absorbent pads at
top and  bottom of  a stack  of filters wrapped in
Kraft paper)  could be  sterilized prior  to use  by
autoclaving at 121 C  for 10 minutes.  Immediately
following  the  time-temperature   exposure,  the
autoclave should be  rapidly  exhausted to atmos-
pheric pressure and  the membranes promptly
removed to minimize total heat exposure. Exces-
sive  exposure to  sterilization  temperatures  can
cause membranes to become  brittle and distorted.
This  problem  is also aggravated  by sterilization
of  membrane filter  stocks  held  in storage for
periods beyond 18 months.

    The  introduction of prepackaged  and pre-
sterilized membrane  filters in resealable envelopes
by  several  manufacturers was considered a desir-
able convenience by  the laboratory and  immedi-
ately  accepted.  Now we  have  evidence from  a
recent comparative  study of these  presterilized
membrane filters  that  there are  significant  in-
creases in bacterial recovery  rates for steam steri-
lized  membrane  filters  compared  to  ethylene
oxide presterilized membranes (37).  As a result,
one  manufacturer that previously  used ethylene
oxide sterilization is now reported  to be using
gamma  radiation for  sterilizing  membranes pack-
aged  in  single  service  envelopes  while  another
manufacturer  has  switched  to  steam  sterilizing
their packs of 10 menbrane filters. For laboratories
that currently have  supplies of ethylene  oxide
sterilized membranes  it may be desirable to submit
them  to steam sterilization (121  C for 10 minutes
with  rapid  steam  exhaust)  to further flush out
latent toxicities. These membranes should then  be
compared with other membranes from the same lot
of ethylene oxide treated membranes  in  a  pure
culture recovery experiment. Possibly some resi-
dual toxic  effect may still persist either from en-
trapped ethylene oxide or its  reaction products.
     Despite manufacturing claims to the contrary,
nitro-cellulose membrane  filters do undergo some
change in their physical characteristics upon stor-
age  in  the  laboratory  for  periods   beyond  18
months.  Upon aging,  membrane  filters may  lose
their flexibility and break apart at pressure points
created during manipulation. During filtration, sur-
face warping often occurs making  a complete con-
tact with the medium substrate  impossible.  The
solution to this problem is not to  stock pile mem-
brane filter supplies beyond  what is estimated to
be needed for a 12 month period.

Other Variabilities in the Membrane Filter Test

    The quality of the membrane filter is not the
sole source  of  unreliable performance.  Bacteria
retained on the  MF  surface may receive nutrients
from a broth saturated absorbent pad or from an
agar based  medium. When a liquid culture medium
is preferred,  the absorbent pad substrate material
must  be  of  high quality  paper fibers, uniformly
absorbent and free of sulfites, acids, or other  sub-
stances that could inhibit bacterial growth. Recent
quality control testing of absorbent pads supplied
with membrane  filters of  various manufacturers,
has demonstrated a significant reduction in colony
counts and colony size associated with use of the
absorbent pad substrate in comparison  to the same
medium prepared  in a 1.5 percent agar base (35).
Until the absorbent paper quality improves,  it will
be necessary  for the  laboratory to remove residual
toxic  materials, such as bleaching agents, by  pre-
soaking  pads in  distilled water held  at 121  C for
15  minutes in the autoclave, decanting the rinse
water, and repackaging pads in large petri dishes
for  sterilization at 121 C for 15 minutes, using a
rapid exhaust to quick dry the pads (22).

    The alternate approach  is to prepare all  MF
broths with  the  addition  of 1.5 percent agar.
However, it should be noted that these agar prepar-
ations must be carefully added  to culture dishes so
as to create a smooth, moist surface, free of pock
marks caused by  foam  and rapid mixing of air
bubbles in the liquid agar preparation.

     Media  manufacturers have  also contributed
significantly  to  the  membrane  filter dilemma
through variations in media  quality  (42). Formu-
lations of  media  currently  available and  recom-
mended  in  various  reference  sources contain  a
variety of peptones, bile salts and dyes which are
not chemically pure compounds and  thus subject
to variations in composition and performance. As
a result, medium sensitivity and selectivity  will
vary unless manufacturers maintain an adequate
                                                15

-------
quality control program to insure that these prod-
ucts meet the requirements for their intended use.
Although  a quality check is made of these  com-
mercial products (43).  it appears to be inadequate.
Poor quality  total coliform sheen development and
significant  reductions   in  coliform  recovery  on
M-Endo medium have been observed by  several
laboratories in recent years. Apparently the use of
poor grades of basic fuchsin and inadequate dye-
sulfite balance in the medium are responsible.  Basic
fuchsin may differ in dye content, both from lot to
lot   and   from  manufacturer  to  manufacturer,
making  it  essential  to  standardize the fuchsin-
sulfite proportion used each time a new lot of dye
is employed  (22). Variations in  intensity  of the
blue color of  fecal coliform  colonies on M-FC
medium  may  be  caused by residual acidity  in
absorbent pads or membrane filters and also  from
unsatisfactory lots of aniline blue used in the com-
mercial preparation of  this medium. The intensity
and structure of bile salt crystals  that precipitate
on  fecal  coliform colonies relates to the type of
bile  salts  complex  incorporated  in  the  medium.
Formulations of commercial  media  containing
sodium azide (M-Enterococcus, KF and PSE agars)
have an approximate shelf life of two years after
production,  because   of  the  deleterious  effects
created  by the slow decomposition of the  azide
compound. For these reasons it is desirable for the
laboratory  to establish a quality control analysis
on each new  lot of  medium purchased, comparing
it with  a  lot of  the same medium  known to  be
satisfactory in terms  of differentia.1 qualities and
sensitivity.

     No  discussion  on   the  variability  of any
laboratory  technique  can be complete without a
recognition of the human element. Reliable labora-
tory performance by every technician in the deter-
mination of bacterial quality of water requires the
continued  application  of  knowledge, skill and
judgment. Only through a uniform application of
careful  technic and  rigid  adherences to  details
will  the  procedures yield the  maximum benefits
of  reliability and  accuracy  (32).  Deviations in
laboratory  procedures  occur as a result of many
factors  including attempted shortcuts, ignorance
of  technical   procedures,  inexperience  in  new
methods,  equipment  failures, inadequate facili-
ties, technical carelessness,  shifts of competent
personnel  to other laboratory assignments, and
lack of interest in the phase of public health bac-
teriology (28). These sources of  variability in the
membrane filter procedure  can be held to  a  mini-
mum through a vigorously  pursued certification
program  at  both the Federal and  State  levels.
This program can best be achieved through  train-
ing in proper methodology, supported by periodic
laboratory evaluations  and a  bacteriological refer-
ence  sample  protocol to  test  laboratory  pro-
ficiency and to reaffirm the continued production
of reliable data (44).
Special Membrane Filter Application Considerations

     The  unique  properties  of  the  membrane
filter and  the  compactness  of  basic  apparatus
stimulated the development of several devices that
appear to have  special potential for field use ap-
plication. One of these devices, the field monitor,
serves initially as the filtration chamber and then
as the  culture  package upon injection  of con-
centrated modified media into a pad below the fil-
ter.The  unit is then ready for incubation and sub-
sequent  colony  counting. Several  independent
evaluations of  the  field monitoring concept for
total coliform recovery from polluted water  indi-
cate  that only  70 percent  recovery of the known
bacterial density is being obtained. The remaining
organism loss occurs from  : a) some bacterial by-
pass  around the filtration area to the pad below the
membrane or direct to discharge through the bot-
tom  port; and  b) failure of some  debilitated cells
to grow on the membrane and medium. In an at-
tempt to seal off the by-pass loss, the manufacturer
has added  a  hydrophobic  substance to the outer
periphery of  the filter.  Unfortunately,  one lot
tested  in our  laboratory  had  a  much  reduced
effective filtering area due  to the  non-wetting
agent. Inclusion of a consistent amount of normal
strength medium is dependent upon displacement
of the water entrapped in  the pad with  1.3 times
normal   strength   ampouled  medium  filtered
through the field monitor  following water sample
filtration.  Vapor  blockage  and  uneven  flow
through will  result in uncertain medium concen-
trations in  the  pad substrate, ultimately affecting
bacterial growth. Ampouled media has a limited
shelf life that must be recognized by  the labora-
tory; 6  months for   M-FC  and  18  months for
M-Endo  when  stored  in  the dark, preferably at
refrigerated temperatures.


     The  bacteriological  "dip stick"  appears to
offer the ultimate yet achieved in test simplicity
at some sacrifice in flexibility. This device consists
of a sterile  rectangular shaped membrane filter
positioned above a medium impregnated pad, both
being secured to a plastic  frame which is inserted
into  a mating plastic  case. The basic principle of
operation is the controlled absorption  of one ml
                                                16

-------
of sample  through the membrane to the medium
impregnated  pad  of critical  thickness when  the
"dip stick" is held in  a water sample for approxi-
mately  30 seconds. The  small volume of sample
makes the "dip stick" self  limiting  for total coli-
form  analysis  in  potable water  because  the test
base-line is established as "less than one  coliform
per 100 ml."  Preliminary evaluation of the "dip
stick"  for  use  as a standard plate count measure-
ment  in potable water compared to the Standard
Methods procedure shows the method to result in
significantly  lower bacterial counts, possibly  be-
cause of the toxicity  inherent  in  the  gray-black
membrane filter and to the inadequately  enriched
medium. The fecal coliform "dip stick" appears to
offer  the field investigator a convenient prelimi-
nary  screening tool for water pollution surveys.
However, before this  procedure can be accepted
as producing definitive data for  stream  standards
and effluent qualities, it  must be evaluated on a
variety  of waters  including acid mine drainage,
highly nutritive paper  mill wastes and chlorinated
effluents. The critical unknown factors involve the
effect  of  source  water  chemistry  on  indicator
bacterial survival  and suppression of false positive
reactions  from non-indicator organisms.  In  this
respect,  a  critical  need exists  to determine  the
adequacy of the  "dip stick" fecal  coliform  pro-
cedure  for monitoring chlorinated  sewage. Con-
sistent reliability of the bacteriological "dip stick"
has yet to be determined and will  relate to basic
qualities of the membrane filter, absorbent pad and
media stability  during storage.

    Special qualities of the membrane filter must
be recognized by  the researcher involved  in radio-
active detection of bacterial indicators. An adverse
effect  of membrane filters on bacterial release of
carbon  labeled CC>2 from  radioactive  tagged  so-
dium  formate was observed  by  Levin et  al. (45).
Filtered  cultures  of bacteria invariably  released
much  less  radioactive  C02 per  cell than did cul-
tures  that  were not filtered. Either  toxic material
from  the membrane filter inhibits metabolic activ-
ity of  the  bacteria and thus, the release of radio-
active C02 or  mechanical  rupture of some of  the
bacterial  cells  by vacuum  filtration  effectively
reduces  ^C02 release. Studies on the  problem
suggest that although there may be some mechani-
cal damage to  cells by impaction,  the mere pre-
sence  of the membrane filter in the growth medi-
um along with unfiltered cells was  enough to re-
duce the 14C02 re'ease- The problem needs more
research investigation so as to control this factor in
radioactive carbon measurement in a rapid test for
bacterial indicators in water.
     Membrane filters can be  used  as one of the
adsorbents  for  the concentration of virus from
water  (46, 47). The clogging  effects of turbidity
can be partially circumvented by use of membrane
filters of 293 mm diameter. However, sterilization
of this size membrane by autoclaving may  result
in increased  brittleness,  making  UV  sterilization
more desirable. The choice  of membrane filter
material is critical, with nitro-cellulose membrane
filters  having a high adsorption affinity for virus
particles at a low pH (in the absence of interferring
substances)  even to pore  sizes 285 times the virus
diameter (48, 49). By contrast, cellulose triacetate
filters  absorb  few  virus  particles,  even  at  pore
sizes no larger than  three  times the virus diameter.

     Some laboratories filter sterilize tissue culture
media  for cell line maintenance. This practice may
introduce a toxic contaminant to the medium from
detergents incorporated in  the filter  to promote
filtration efficiency and sterilization by autoclaving
(50).  Therefore, it  may  be  advisable to flush all
membrane filters in  hot distilled water followed by
an ice cold saline rinse before use.

     The recent development of a dialysis chamber
with membrane filter side walls offers an excellent
opportunity to study bacterial survival in a variety
of natural and artificial water environments (51).
Membrane filters  used  for  this  purpose must be
sturdy enough to withstand the buffering effects
of water currents and for this reason tear resistant
micro-web membranes  with  a  nylon  backing are
recommended. Substitute membranes with a rein-
forced backing must also  be free of biodegradable
materials that will  encourage  the development of
microbial  films  over their surface, restricting  the
in-flow of water and solutes to  interact with  the
bacterial suspension. Finally,  it is of critical  im-
portance that these membrane filters be  non-selec-
tive in their passage of solutes which might alter
the chemistry of the  water  under  investigation
and thus effect  the bacterial survival patterns pro-
duced.


                  SUMMARY

     There can  be no  doubt that the membrane
filter dilemma is real and needs urgent resolution.
Membrane filter and  media  manufacturers must
heed the outcries from anguished microbiologists
and reevaluate their product and quality  control
programs. If not, these corporations risk the loss
of their multimillion dollar market through a grow-
ing wave of no-confidence in any membrane filter
procedures with subsequent abandonment of this
                                                17

-------
heretofore  approved  laboratory  tool  in water
microbiology. This indeed would be a tragic turn
of events that need not happen.

     The laboratory staff should not be influenced
by advertising claims of product excellence which
we pay for but may not be getting. Possibly 15
percent of the laboratory activity should be appor-
tioned  to a quality control program on membrane
filter and  media lots  and a variety  of routine
quality control  procedures. The researcher is also
cautioned to include adequate controls in all exper-
iments involving membrane filters so that variables
introduced  from this  source will  not adversely
affect the data interpretation.
                REFERENCES

  1.  Bigelow, S., and  A.  Gamberling,  Collodion
     Membranes. J. Amer.  Chem.  Soc.  29:1576,
     1907.
  2.  Pierce,  H.F.  Nitro-cellulose  Membranes  of
     Graded  Porosities. J.  Bio. Chem.  75:795,
     1927.
  3.  Ferry, J.D. Ultrafilter  Membranes  and Ultra-
     filtration. Chem. Rev. 18:373,  1936.
  4.  Bechhold,  H. Kolloid Studien mit der Filtera-
     tiones Methode.  Z. Physik.  Chem.  60:257,
     1907.
  5.  Bechhold,  H. Durchlassigkeit von UltrafiItem.
     Z. Physik. Chem. 64:328, 1908.
  6.  Bechhold,  D. Ultrafiltration. Biochem Ztschr.
     6:379, 1907.
  7.  Brown, W. II. On the Preparation of Collo-
     dion Membranes of Differential Permeability.
     Biochem. Jour. 9:591,  1915.
  8.  Brown, W.  VI. Further Contributions  to the
     Technique  of  Preparing  Membranes  for
     Dialysis. Biochem.  Jour. 11:40, 1917.
  9.  Schoep,   A.  Uber ein  Neues   Ultrafilter.
     Kolloid Zeit. 8:80, 1911.
 10.  Walpole, G. XXVI. Notes  on  the  Collodion
     Membrane  for  Ultrafiltration  and  Pressure
     Dialysis. Biochem.  Jour. 9:284, 1915.
 11.  Zsigmondy, R. and W. Beckman, Uber Neue
     Filter. Ztschr. F.  Anorg.  Chemie 103:119,
     1918.
 12.  Zsigmondy, R. and G.  Jander, Die chemische
     analyse  mit Membranfilter. Z. Anal.  Chem.
     58:241,  1919.
 13.  Zisgmondy, R. U.S. Patent 1,421,341; June
     27,  1922.
 14.  Zsigmondy, R. and W. Beckman, Uber Fein-
     porige Filter und  Neue Ultrafilter.  Biochem.
     Ztschr. 171:198, 1926.
15. Goetz,  A.  Fiat  Final  Report  1313.  Joint
    Intelligence Objectives Agency,  U.  S.  Dept.
    of Commerce, Wash., D.C. 1947.
16. Rasumov, A. S.  Priamoi  metod uchets bak-
    terii  v  vode. Srav  nenie ego  s  metodom
    kokha,   Mikrobiologiya   (USSR),   1,   131,
    1932.
17. Rasumov, A. S.  Novye  metody i put!  Kach-
    estvennogo   i  Kolichestvennogo  izucheniia
    mikroflory vody, Mikrobiologiya (USSR),  2,
    346,  1933.
18. Eschenbrenner, vbn  H., Keimfreie Filtration
    in  Apothekenbetrieb. Pharm.  Montash, 14:
    169,  1933.
19. Thomann,   von   J.   Uber  die  Herstellung
    Kiemgreier  Injektianss-Losungen in Apothe-
    kenbetrieb,  Pharmaceutical  Acta  Helvetias,
    9:2, 1934.
20. Mueller, G.  Lactose-Fuchsin Plate for Detec-
    tion of Coli  in Drinking Water by Means  of
    Membrane Filters. Ztschr. Hygiene  u. Infek-
    tionskrankh 127,3/4, 187, 1947.
21. Mueller, G.  Eine  trinkwassergebundene Ruh-
    repidemie. Zentralbl.  Bait. Ab. I. Orig.  152:
    133,  1947.
22. Clark, H.F.,  E.E. Geldreich, H. L. Jeter, and
    P.  W. Kabler, The Membrane  Filter in Sani-
    tary Bacteriology. Pub.  Health Rep. 66:951,
    1951.
23. Clark, H.F., P. W. Kabler, and E. E. Geldreich,
    Advantages and Limitations of the Membrane
    Filter  Procedure, Water &  Sewage  Wks.
    104:385, 1957.
24. Goetz,  A.  and N. Tsuneishi, Application  of
    Molecular  Filter  Membranes  to  Bacterio-
    logical  Analysis  of  Water,  J. Amer.  Water
    Works Assn. 43:943,  1951.
25. Goetz,  A.  R.H.  Gilman, and A.M.  Rawn,
    Application of Molecular Filter Membranes to
    Specific Problems in  Water Analysis. J. Amer.
    Water Works Assn. 44:471, 1952.
26. Windle  Taylor,  E., N.P.  Burman, and  C.W.
    Oliver,  Use of the Membrane Filter  in the
    Bacteriological  Examination  of  Water.  J.
    Applied Chem. (London) 3:233, 1953.
27. Windle  Taylor, E.,  and  N.  P. Burman. The
    Application of Membrane Filtration Techni-
    ques  to the Bacteriological  Examination  of
    Water, J. Appl. Bact.  27:294, 1964.
28. Geldreich,  E.E.  Application  of  Bacterio-
    logical  Data in   Potable Water  Surveillance.
    J. Amer. Water Works Assn. 63:225, 1971.
29. Clark,  H.F., H.L. Jeter, E.E.  Geldreich, and
    P.W.  Kabler. Domestic  and European Mole-
    cular  Filter  Membranes. J.  Amer.   Water
    Works Assn. 44:1052, 1952.
                                               18

-------
30.  Caspar,   A.J.,   and J.M.  Leise.  Inhibitory
     Effect of Grid  Imprints on Growth of Pasteu-
     rella tularensis on Membrane  Filter. J. Bact.
     71:728, 1956.
31.  Loesche, W.J.  and S.S. Socransky. Defect
     in  Small Millipore Filters  Disclosed by New
     Technique  for Isolating Oral  Treponemes.
     Science  138:139, 1962.
32.  Geldreich, E.E., H.L. Jeter, and J.A. Winter.
     Technical  Considerations  in  Applying  the
     Membrane  Filter  Procedure,   Health  Lab.
     Sci.4:113, 1967.
33.  Levin,  G.V.,  V.L. Strauss, and W.C. Hess.
     Rapid   Coliform   Organism  Determination
     with C14.  J.  Water Poll.  Contr.  Fed.  33:
     1021, 1961.
34.  Brown,  O.R.   Inhibition of Escherichia  coll
     on  Cellulose  Acetate  Membrane   Filters.
     Microbios 7:235, 1973.
35.  Presswood,  W.C.,  and  L.R.  Brown.  Com-
     parison  of Gelman and  Millipore Membrane
     Filters   for  Enumerating  Fecal   Coliform
     Bacteria. Appl. Microbiol. 26:336, 1973.
36.  Hufham, J.B.  Evaluating  the  Membrane
     Fecal Coliform Test by Using Escherichia coll
     as the Indicator Organisms. Appl. Microbiol.
     27:771,  1974.
37.  Dutka,   B.J.,  M.J.  Jackson, and  J.B.  Bell.
     Comparison  of  Autoclave  and   Ethylene
     Oxide-Sterilized Membrane Filters  Used in
     Water  Quality Studies.  Appl. Microbiol.
     28:474,  1974.
38.  Schaeffer, D.J., M.C. Long, and K.G. Janar-
     dan.  Statistical  Analysis of the  Recovery of
     Coliform Organisms on Gelman and Millipore
     Membrane  Filters.  Appl.  Microbiol.   28:
     605, 1974.
39.  Alico, R.K. and C.A.  Palenchar.  Staphylo-
     coccus aureus  Recoveries on Various Brands
     of  Membrane  Filters. Health Lab. Sci.  12:
     341-346, 1975.
40.  Kabler,  P.W., and  H.F. Clark.  The  Use of
     Differential Media  with the Membrane  Fil-
     ter. Amer. J. Pub. Health 42:390, 1952.
41.  Burman,  N.P.  Developments in  Membrane
     Filtration  Techniques.  1.  Coliform  Counts
     on MacConkey Broth. Proc. Soc. Water Treat.
     Exam. 9:60, 1960.
42.  Wolochow,  H. The Membrane  Filter Tech-
     nique  for Estimating  Numbers. Appl. Micro-
     biol. 6:201, 1958.
43.  Vera,  H.D. Quality  Control of  Diagnostic
     Microbiology. Health Lab. Sci. 8:176, 1971.
44.  Geldreich,   E.E. A  Total System  Approach
     to  Monitoring the Bacteriological  Quality of
     Water. 2nd Annual Water Tech. Conf. AWWA.
     Dec., 1974.
45.  Levin, G.V.,  V.L.  Strauss, and W.C. Hess.
     Rapid   Coliform  Organism   Determination
     with  C14.  J. Water  Poll.  Contr.  Fed.  33:
     1021,  1961.
46.  Oliver,  D.O.  Factors in  Membrane   Filtra-
     tion   of   Enteroviruses.   Appl.   Microbiol.
     13:417, 1965.
47.  Wallis, C.M., M. Henderson, and J.L. Melnick.
     Enterovirus Concentration on Cellulose Mem-
     branes. Appl. Microbiol. 23:476, 1972.
48.  Oliver,  D.O. Factors in the Membrane  Fil-
     tration  of  Enteroviruses.  Appl.  Microbiol.
     13:417, 1965.
49.  Oliver,  D.O. Virus Interactions with  Mem-
     brane   Filters.  Biotechnol.   and  Bioengr.
     10:877, 1968.
50.  Cahn,  R.D. Detergents  in Membrane  Filters.
     Science 152:195, 1967.
51.  McFeters,  G.A.,  and  D.G. Stuart. Survival
     of  Coliform  Bacteria  in  Natural  Waters:
     Field and Laboratory Studies with  Membrane
     Filter  Chambers.  Appl. Microbiol. 24:805,
     1972.
                                              19

-------
                      QUALITY CONTROL OF MEMBRANE FILTER MEDIA

                                      David A. Power, Ph. D.
                                   Manager of Technical Services
                        BioQuest, Division of Becton, Dickinson and Company
                                      Cockeysville, Md. 21030
                 ABSTRACT

     The manufacturers of dehydrated membrane
filter media provide quality control, lot-to-lot uni-
formity and relative stability of their products. The
quality control procedures test the media compon-
ents and  the dehydrated  products for  suitable
physical  and chemical  characteristics, microbial
contamination and  growth support. In addition,
the  dehydrated  products are subject  to  perfor-
mance tests based on the use of the specific media.
However, the users must maintain quality control
after receiving the media.  Quality control proce-
dures in the laboratory  include age of product,
storage  conditions, accurate weighing of  dehy-
drated media,  good  quality of  distilled water,
clean utensils,  complete  mixing  and  solution,
controlled  heating, accurate pH  determinations,
approved  supplements  only,  and  appropriate
checking with stabilized  test cultures. Tests that
the user  should perform  on the completed media
include observation of appearance, pH, sterility,
and  membrane filter  performance  with  selected
test organisms.
               INTRODUCTION

     The development of membrane filter techni-
ques for the isolation, enumeration and  differen-
tiation of microorganisms in water, sewage, milk,
foods,  air, solutions, specimens, etc., created a
need for culture media  especially suited for  use
with these techniques.

     Kabler and Clark (5) reported that  formulas
of most media for conventional use must be modi-
fied before the best results  are obtained  with  the
membrane. This modification may be in  the form
of  changed  quantities of ingredients  or  the sub-
stitution of nutrients. These authors noted that in
many  instances there is no correlation  between
results obtained with an agar medium using con-
ventional methods and the same formula without
agar when used on a membrane.

    This  need  for specialized media has resulted
in  the  production  by  commercial  dehydrated
media manufacturers of a broad line of media for
membrane filter techniques, which are denoted by
a "M" prefix  to  the product name.  Dehydrated
membrane  filter  media offer the laboratory the
same advantages as standard dehydrated media —
relative  stability,   lot-to-lot  uniformity  and  the
quality control "built  into" dehydrated products
by commercial media manufacturers.

         QUALITY CONTROL BY THE
              MANUFACTURERS

    As with other dehydrated media, membrane
filter media are subjected to  quality control pro-
cedures at least twice before reaching  the user.
Initially,   the   raw  materials incorporated into
dehydrated media are  tested before inclusion in
various  media  formulas.  Peptones  are examined
for physical appearance of the powder, pH, clarity,
and color, and for microbial  contamination. Pep-
tones are also subjected to growth support tests in
which a single peptone serves basically as the only
nutrient. Agar is tested for gel strength, clarity by
nephelometry, color by colorimetry, gelation and
melting temperature, pH, solution time, and visual
appearance.  Only   satisfactory   ingredients  are
employed in the manufacture  of the various media.

     The second check is performed on the dehy-
drated product. Dehydrated  media are  subjected
to the same types of tests described for ingredients,
such as appearance, pH, clarity, and color. Perfor-
mance tests with microbiological cultures are based
on  the end use of the medium. For selective media,
it is necessary to demonstrate satisfactory growth
of  desired  species  and   inhibition  of undesired
species, recognizing that a selective medium repre-
sents a compromise in  that the selective agent may
                                                20

-------
somewhat inhibit strains of desired species as well
as undesired species. The performance of differen-
tial media  is assessed  by the use  of appropriate
species,  including  those  producing positive and
negative reactions.

     For example,  the BBL  laboratory  evaluates
M-Coliform  Broth,  the BBL equivalent of M-Endo
Broth, by determining that the broth colorimetric
reading is satisfactory, and that the pH after heat-
ing and cooling to room temperature is 7.2 ± 0.2.
Growth support tests include the use of strains of
Enterobacter aerogenes, Escherichia coli, Pseudo-
monas  aeruginosa.  Salmonella  typhimurium, and
Shigella  sonnei.  The  medium is tested  for  the
ability to support growth following straight inocu-
lation  of 10"' dilutions of test organisms, and for
the recovery and differentiation of E. coli in the
standard procedure by adding  one ml of a dilution
containing 30 to 300 organisms to 50 ml of water,
followed by  filtration through a  membrane filter
subsequently incubated on a medium-soaked pad
for 24 hours at 35  C. Growth  must be satisfactory
and reactions correct. Streptococcus fecalis is also
included  in  the   performance  evaluation,  and
growth must be inhibited.

     The use  of stabilized, freeze-dried  cultures
permit results to be predictable and reproducible
on standard or reference lots of media.

      QUALITY CONTROL BY THE USER

     The  need  for an  internal  quality  control
program in the microbiology laboratory  has been
documented  in  numerous papers  and has been
legislated for laboratories involved in the interstate
practice  of  laboratory  medicine  through   the
Clinical  Laboratories Improvement  Act  of 1967
(3). Abuses resulting in poor performance have not
been   uncommon,  despite  the  quality  control
procedures performed  by commercial manufac-
turers  on dehydrated  media  and  media supple-
ments  before their release and  directions on labels
and on product or methodology  manuals for re-
constitution and  preparation of  final  media and
their handling (1, 2,6, 8).

    Some of the factors that may contribute to
the preparation and use of unsatisfactory  finished
media  from dehydrated materials  when little or no
attention is  paid to the quality of the completed
media include the following (9):

     1.  Incorrect  weighing   of  dry  material,
         through  human error or use of a faulty
         balance.
 2.   Use of dry material taken from previous-
     ly opened  bottles,  which  may have
     deteriorated  from  exposure  to  heat,
     moisture,  oxidation,  or  other  environ-
     mental factors. The quantities in which
     media are purchased should be regulated
     by  the  rate  of usage.  Ideally,  bottles
     of  dehydrated  media,  once  opened,
     should be used within a few weeks. The
     purchase of media in 1/4 Ib bottles or
     in sealed  preweighed envelopes  is en-
     couraged.

 3.   Incorrect measurement of water and use
     of tap water or water from a malfunc-
     tioning still or deionizing resin column.
     Water  should  generally  meet  the  re-
     quirements of the United States Pharma-
     copeia (USP)  XVIII  (10)  for  purified
     water or be  of proven  microbiological
     quality.

 4.   Use of unclean containers or glassware,
     especially those  contaminated with de-
     tergent or other chemicals.

 5.   Incomplete mixing  or incomplete solu-
     tion  resulting  in  failure to  prepare  a
     homogeneous medium. With  agar media,
     this may even product stiff medium in
     some plates and soft medium in others.

 6.   Overheating occurring during  prepara-
     tion  and sterilization, or resulting from
     holding  too long in the molten state be-
     fore dispensing into plates, tubes or bot-
     tles, can result in the loss of productivity
     through hydrolysis of agar, carmelization
     of carbohydrates,  lowering  of pH, in-
     crease in  inhibitory action,  loss of dye
     content  in   selective  or  differential
     media,  and formation  of precipitates.

 7.   Improper determination of pH, resulting
     in the addition of too  much acid or
     alkali. The pH of a medium should be
     determined electrometrically;  the elec-
     trodes should  be  in  contact with the
     solidified  agar  medium, which  may be
     removed from a plate  or tube and placed
     in a beaker.

8.   Improper addition or incorporation of
     unsatisfactory  supplements  or enrich-
     ments,  or addition of supplements at
     the wrong temperature, possibly causing
     alteration  of  the  supplements  if  the
                                               21

-------
         temperature is  too high, or gelation of
         media  before proper mixing if too cold.

     9.   Failure of  the laboratory to  subject
         samples  of  finished  media to  quality
         control procedures with stabilized test
         cultures  before the  media  are  used.
         Regardless of the amount of quality con-
         trol built into a product, users should be
         aware  of strain variations  which  influ-
         ence recovery  and growth  and of the
         purpose  for which  each medium was
         designed.

            STORAGE OF MEDIA

     Once prepared, special attention  must be paid
to proper storage. The selective  and differential
membrane  filter media are more susceptible to
deterioration  than  similar routine culture media.
For this reason,  we recommend that broth media,
prepared in sealed tubes or bottles for later use, be
stored in a  refrigerator maintained at 2 to 8 C and
used within a couple  of  weeks. It is generally re-
commended  that   M-Coliform   Broth   (M-Endo
Broth)  and M-FC Broth be used within four to five
days (4, 7). Plates of agar  media should be pre-
pared and used promptly, or be stored wrapped in
foil or plastic to limit water loss  and used within
two  weeks. Containers  of prepared  membrane
filter media must be protected from light.
     RECOMMENDED QUALITY CONTROL
            CHECKS BY THE USER

     Checks that the  user can  perform  on the
finished media include:
     1.
     2.
     3.
     Appearance — if the medium is off-color,
     or there is a change in appearance during
     storage or signs of  drying, contamina-
     tion   or   deterioration,  the  medium
     should be discarded.

     pH  —  measured electrometrically  at
     room  temperature.  The pH should be
     within ± 0.2 of  that stated on the label.
     Sterility testing — by incubating repre-
     sentative samples,  for two or more days.
     Such sample tubes or plates should be
     discarded and not used for later culture
     work.
4.    Performance — as a  minimal  test, we
     recommend the straight inoculation of
     10~1  dilutions of test  organisms. As a
                                                       test of the complete system, we recom-
                                                       mend  the filtration through a membrane
                                                       filter of a water sample inoculated with
                                                       a  diluted suspension  of an appropriate
                                                       organism,  followed by  handling  and
                                                       incubation according to the procedure
                                                       established for that medium. This test is
                                                       a  quality control  check for the entire
                                                       system.

                                                  If  a problem  is encountered, users  are en-
                                             couraged  to contact  the  manufacturer.  The  lot
                                             number  of  the  product and dates  on which the
                                             medium  was  received  and   first opened  should
                                             accompany the observations.
                REFERENCES

1.    Blair, E.B., Media, Test Procedures and Chem-
     ical  Reagents. In H.L. Bodily,  E.L. Updyke,
     and  J.O. Mason (eds.), Diagnostic Procedures
     for  Bacterial,  Mycotic and  Parasitic  Infec-
     tions,  pp.  791-857, American  Public Health
     Association, New York, 1970.
2.    Difco  Manual,  9th ed.  1953. Supplementary
     Literature. Difco Laboratories, Detroit, 1968.
3.    Federal  Register. Clinical Laboratories  Im-
     provement Act of 1967 — Notice of Effective
     Date. 33. No. 253. U.S. Government Printing
     Office, Washington, D.C., 1968.
4.    Standard  Methods for the Examination  of
     Dairy  Products,  13th ed., American Public
     Health Association, Washington, D.C., 1972.
5.    Kabler, P.W., and H.F. Clark. The Use of Dif-
     ferential Media with  the  Membrane Filter.
     Am. J. Public Health 42:390-392, 1952.
6.    Rohde,  P.A.  (ed.).  1968.  BBL Manual  of
     Products and Laboratory Procedures, 5th ed.,
     BioQuest, Division of Becton, Dickinson and
     Co., Cockeysville, Mr., 1968.
7.    Standard  Methods for the Examination  of
     Water  and  Wastewater,  13th  ed., American
     Public Health Association, Washington,  D.C.
     1971.
8.    Vera,  H.D., and  M. Dumoff. Culture Media.
     In E.H.  Lennette,  E.H. Spaulding, and J.P.
     Truant (eds.),  Manual  of Clinical Microbio-
     logy 2nd ed., pp. 881-929. American Society
     of Microbiology, Washington, D.C., 1974.
9.    Vera,  H.D. Quality  Control   in  Diagnostic
     Microbiology.  Health Lab. Sci.  8:176-189,
     1971.
10.  United States Pharmacopoeia.  18th ed. Mack
     Publishing Co., Easton, Pa., 1970.
                                               22

-------
                 DISCUSSION

Geldreich:   Did you say that in the development
            of  M-Endo media BBL  uses a color-
            metric  method   to  determine  the
            proper  amount   of  basic  fuchsin?

Power:      There is some  color determination.  I
            think it can vary with the lot of dye
            involved, and there would be a range.

Geldreich:   The  problem  that  I  am  concerned
            about is that many or all of us prob-
            ably  know that  the actual dye con-
            tents of these  products vary consid-
            erably.  Dyes such as basic fuchsin,
            aniline  blue, brillant green, etc. are
            not  chemically   defined   materials.
            Sometimes the  microbiologist is work-
            ing with  a trade brand  of material
            which  was  originally  intended  for
            dyeing  clothes  and not for preparing
            media. Therefore, basic fuchsin, as an
            example,  may  have  a  dye content
            anywhere from  88% to 99%.

Power:      That  would  be taken  into considera-
            tion in the formulation of the actual
            dye.

Geldreich:   This  is our concern.  Years ago  we
            suggested that  the manufacturers per-
            form a biological dye titration when
            they  made  up M-Endo media. This
            titration  is done  by  holding  the
            amount of  sodium  sulfite  constant
            and  running each new lot of  basic
            fuchsin  against  it.  Each  batch  be-
            comes  an  individual lot of media and
            is checked for  recovery of the organ-
            isms  with good sheen and no evidence
            of  toxicity.  I  don't think the manu-
            facturers  do this. We often find  that
            these products have  poor sheen and
            sometimes poor  recovery.  I think it
            is  because the  sulfite-basic  fuchsin is
            not in the proper proportion.

Lane:       I am  quite sure that  BioQuest titrates
            the dye content  in  their MF media
            just as Difco does. You are absolutely
            right  that there   is  variation in  the
            dyes, and one  of the worst problems
            has been  with  the aniline blue. Basic
            fuchsin has  not  been too  bad. The
            only way  that the dye content can be
            determined satisfactorily  is by per-
            formance test. If the test organisms
            (coliforms in this  case)  don't  yield
            typical colonial morphology or color,
            the  dye  content  is  off.  The manu-
            facturers  of dyes seem to have lost the
            control that  they  had years ago.  It
            is very difficult to get a  satisfactory
            batch  of basic fushion, aniline blue or
            brilliant green. It is up to the manu-
            facturer  to   standardize  the media
            so  that the  morphology and  total
            counts are satisfactory. We do titrate.

Power:      We do have  the standards. If you ex-
            perience  a problem or if you see a
            variation,  I  wish  you would let us
            know.

Seidenberg:  You made a comment that has me in
            a dilemma; you said the M-FC and the
            M-Endo agar plates can be  kept for
            4 to 5 days.
Power:      Yes.

Seidenberg:  We in EPA are very careful to discard
            M-Endo plates after 48 hours, because
            we have found color changes.

Power:      I  notice  that our manual and APHA
            manuals allow 4 to 5 days.

Seidenberg:  I  believe that Standard Methods for
            the  Examination of Water and Waste-
            water states 48 hours.
Power:      I believe that the holding  times in the
            Standard   Methods  for  Water  and
            Dairy  Products   may  differ by  one
            day or so.
Geldreich:   Standard  Methods  allows the worker
            to hold M-Endo  and  M-FC  plates for
            use during that work week. However,
            I  think  some  of  the  laboratories
            have  found that these media are  light
            sensitive and if they don't store them
            in the dark  there  is  a real problem.
            One could cover  half a plate, leave it
            on a  laboratory  bench  about  two
            hours, take that cover away and  find
            two  shades  of media, have  resulted.

Power:      I think it is  best  to use the media on
            the day of preparation.

Geldreich:   If possible, that is great. Many work-
            ers ask how  long they can keep the
            media. The current Standard Methods
                                               23

-------
            recommends a  work  week. This re-
            commendation may be changed in the
            future.

Brodsky:    I question the use of pure cultures as
            the  only quality control that you use
            on your media. We are not accustom-
            ed   to  receiving pure  cultures  in  a
            water sample.  The relationship  be-
            tween the various  organisms in their
            natural environment is negated by the
            use  of pure cultures only. Perhaps
            pure cultures  are  used  just because
            they are expedient.

Power:      The use of the pure culture gives us a
            base for the comparison  from lot to
            lot.  Perhaps  out of this meeting you
            will come  up  with some improved
            system as  you  indicated.  Out of all
            the  possibilities  one might select,  I
            frankly don't  know which is  most
            practical. We feel that we have organ-
            isms on which we  have a history and
            back  over  the years  we have  used
            these  cultures  to test many lots. We
            are  trying  to give the user a product
            now that relates to previous lots, to
            keep  them  as  constant  as possible.
            One of the purposes of the meeting is
            to  talk about the  membranes,  pads
            and media and to come up with  final
            specifications.

Lane:       The  manufacturers of  media  must
            have a base line and the only base line
            we can presently use is pure cultures
            of Aerogenes,  E. coli, or Salmonella.
            However,  the  final assay,  the  final
            approval, and  this  again  is true for
            BioQuest as  well as Difco, is the use
            of a natural specimen. One can titrate
            a  selective  medium  using a  stock
            culture, and can get  certain results,
            such as selectivity. We recognize that
            when  we use natural specimens  such
            as   natural  water,  polluted  water,
            sewage and  stool specimens, we get
            entirely different results. The  state-
            ment  that the manufacturer does not
            take into consideration the different
            results that  could be obtained  with
            natural specimens may not be correct.

            There is  one  point I would like to
            raise;  this  is a  problem that we en-
            countered  recently. We are taking part
            in  the revision of  Standard Methods
           for the  Microbiological  Examination
           of  Dairy Products.  We  were using  a
           buffered  diluent as  in  water  micro-
           biology. We found that  different lots
           of  phosphate  gave different total
           count results. Not only  the  different
           lots of phosphates but the time that
           organisms survived in these  different
           lots varied. The organism will  survive
           for an hour in  one  lot  of phosphate
           but  in  two  hours   the count goes
           down.  The count  for another  lot
           might be satisfactory at  zero time but
           the count may go down after being
           held  for  15 minutes.  We  also  en-
           countered the problem  of the varia-
           tion in distilled  water when preparing
           a buffer from a good lot  of phosphate.
           This  is  a problem  that nobody  has
           mentioned, and I think that you must
           consider this.

Bordner:    In  water analyses we are aware of the
           potential toxicity of phosphate buf-
           fers  and  the  importance  of good
           quality distilled water. Standard Meth-
           ods allows an alternate  dilution/rinse
           water, 0.1% peptone. The next  edition
           will recommend the addition of mag-
           nesium sulfate to the phosphate buf-
           fer  as an  added protective  against
           toxicity.  There is also a test for  the
           suitability  of  distilled  water  pro-
           vided in Standard Methods.

            I have a two-fold question. You have
           stated, Aaron,  that  you periodically
           use natural samples.  Could you  give us
           any  estimate on  your  sampling  fre-
           quency  or upon what  percentage of
           the   media  sampling  per   lot   is
           based.  Secondly,  do  you gentlemen
           have any suggestions on how we as
           water  microbiologists  could   work
           more closely with the  media manu-
           facturers as I know you  do with other
           groups.  Could it be done with ASTM
           or  other  organizations and could it be
           a continuing relationship?

Power:     That  is why we are here, partly to tell
           you our story and mostly to listen to
           what  you  have to  say, because ob-
           viously you have problems. If we can
           be  of help, we certainly intend to
           cooperate. Until this week,  I haven't
                                                24

-------
           been totally aware of all the problems
           involved and I certainly am willing to
           take back  to my company anything
           that I   hear. We are very willing to
           work with you  and  try to come to
           some solutions.

Lane:      We check  every lot of the M-Endo
           broth  MF; for example,  every lot is
           checked with river water. The other
           membrane media are not checked as
           frequently.   We  spot  check  them;
           perhaps 1 out of 4 or 5 lots. The basic
           media   are  checked  every  time  we
           make a  new batch. Concerning other
           media,  for example,  media for  isola-
           tion of Neisseria gonorrhea, all media
           are assayed  with clinical specimens.
           All media for isolation  of Salmonella
           and Shi gel la are checked with  stool
           specimens  which  are  seeded  with
           Salmonella  and  Shigella. Unfortun-
           ately you can't obtain  natural speci-
           mens containing these  pathogens so
           we seed them with different concen-
           trations. I am  sure  that we  do the
           same with the medium  in which you
           are primarily interested, the M-Endo
           broth MF or the M coliform medium.
           We do  check them with river water
           every time a batch is prepared.

Brezenski:  I want to get back to the statement
           concerning the variation  in dye con-
           tent, because I  feel  this  is very im-
           portant. Every time a laboratory gets
           poor  results there  is a  tendency to
           say  there  was  a problem  with the
           medium. The microbiologist talks to
           the manufacturer and he says, "well
           we have a problem with  the  raw in-
           gredients which  we  don't have that
           much control over."  I am wondering
           whether the  manufacturer,  for ex-
           ample   Difco  or BioQuest, specifies
           the amount of the active dye ingred-
           ient that is  supposed to be  in aniline
           blue and basic fuchsin and  what the
           percentage  of  inactive  ingredients
           should  be.  These specifications could
           then be established  to  conform  to
           the quality  control. For example, in
           the FITC dye used  in  fluorescence
           antibody CDC sets up a specification;
           they must have  80 to 90 percent ac-
           tive dye and  only about 10 percent
           inactive ingredients. If the dye doesn't
           meet their specifications, they do not
           buy it. I  am wondering whether Difco
           or  BBL  have  specifications like this;
           because if  we don't, this is where we
           should start.

Power:     We do and on that particular product
           our specification  exceeds the  CDC
           specification.

Brezenski:  The statement was made that varia-
           tion  is a  problem; where  does the
           variation occur?

Lane:      There is  variation  in  natural dye
           content.  When you are  comparing a
           fluorescent dye with an aniline blue
           and a basic fuchsin you are comparing
           almost a  pure chemical  with a mix-
           ture. The only way that you can test
           and approve or reject a batch of basic
           fuchsin or a batch of aniline blue  is
           by   performance.   You  cannot set
           specific criteria for actual  dye con-
           tent. You  will have a variation, you
           will have a range  of actual dye con-
           tent, but the remaining components
           of  a batch of basic fuchsin are really
           not specific materials. They are not
           even identified, so specifications can't
           be  set. The  only  specification that
           can be set for any culture medium  is
           its  performance. Does the medium do
           what it was designed to do? Does this
           coliform   medium,  the MF  medium
           that you are using to isolate the maxi-
           mum number of coliforms present in
           that sample,  yield colonies which are
           characteristic?  This  is what  you are
           after.
           As  you pointed out, there are many
           variables in the dyes. Let me use pep-
           tones as another  example. The pep-
           tone is  not  a specific substance. A
           peptone  is either an enzymatic digest
           or  acid   hydrolysate  of meat and
           casein.  The  method  of  production
           may be the same  and  the control of
           production for dyes as well as pep-
           tones may be the  same; but the final
           criterion of a good peptone or a good
           dye is how  does   it  perform in the
           medium.
                                               25

-------
          STATISTICAL INTERPRETATION OF MEMBRANE FILTER BACTERIA COUNTS

                              K. J. Sladek, C. F. Frith, and R. A. Cotton
                            Millipore Corporation, Bedford, Massachusetts
                  ABSTRACT

     Statistical design and interpretation of experi-
ments  for  enumerating  bacteria  by  membrane
filters are discussed.  To increase precision of mean
counts, large numbers of replicates can be used. In
fecal coliform  tests,  large numbers of replicates
can be achieved using a water sample stabilized by
dilution  in phosphate  buffered peptone. Another
way of increasing precision is to reduce the scatter
among replicates. However, the random fluctuations
of bacterial  density,  even in a well mixed sample,
place a lower limit on scatter. This limit is predict-
able as a theoretical  minimum standard deviation,
which  is useful as a yardstick in comparing with
experimentally  determined  standard  deviations.
Experimental designs should provide for randomi-
zation  of procedural variables so that unexpected
variables  do  not bias results.
               INTRODUCTION

     Comparing different methods for enumerating
bacteria  is a  familiar endeavor  in  microbiology
laboratories. These evaluations include comparing
MF techniques with other standards such as pour
and  streak plates, comparing variations  in proce-
dure, comparing  membranes  of different manu-
facture, and evaluating new media  in comparison
to available ones. In  addition, in production and
quality control operations, membrane manufactur-
ers are  involved  continuously  in  evaluating  im-
proved  membrane  manufacturing and processing
methods  in comparison to existing ones  and  in
qualifying production lots of  membranes  against
standards.

     The  present paper discusses experiments  in
which the same water source or culture is used to
test different membranes, media or procedures. An
example  is to  evaluate effects of sterilization on
membrane filters. You may have had the sad exper-
ience, as we have on some occasions, of reviewing
pages of  data  only to  find  that  each  reviewer
reaches  different conclusions from  the study! We
will address here the formidable problem of carry-
ing out  an experiment  leading to conclusions upon
which all concerned can agree.

     The  factors contributing to a successful ex-
periment  can be categorized as statistical factors,
experimental design principles, and  bacteriological
considerations.  Statistical factors include  decisions
on the number of replicates to be used, the analy-
sis of the  scatter in  counts  expressed by the stand-
ard deviation of replicate counts and calculations
of  confidence  limits.  Experimental design must
provide for  isolation  of  the differences under
investigation from  extraneous effects;  random
selection  of samples  to  be tested is important
here. The bacteriological  considerations include
choice of the  water source or the culture to be
used, establishment of a detailed  procedure,  and
selection of a useful control method. This paper is
concerned mainly with statistical and experimental
design  factors;  some of the bacteriological consid-
erations in fecal coliform  tests will be  discussed
briefly.

Statistical Factors

     The  precision  of  a count determined by aver-
aging n replicate plates can be expressed  by confi-
dence  limits, which give a  range in  which the true
mean will lie.  For  example, the "95% confidence
limits"  are given by ± 1.96 a/^n, where ^/ is the
standard  deviation  of  the population of individual
measurements. The experimental mean has a prob-
ability of 95%  of being within ± 1.96 a/Vn of the
true mean. To increase precision it is necessary to
reduce a  or to increase n.  Increasing n requires an
operating  procedure  which is identical  for  each
replicate  and a water sample which  is stable for the
life of the experiment.  An  example of  sample
stabilization is given next. The scatter in replicate
counts, which  is  measured by a, is considered
afterwards.
                                               26

-------
     Fecal  coliform counts  for sewage  samples
diluted three different ways are given in Figure 1.
Each  plate was  prepared  by spreading a 0.1 ml
aliquot of  diluted sample  onto M-FC agar.  The
lower graph, representing sterile water and  phos-
phate buffer diluents, shows that the sample is un-
stable even for periods as short as 15 minutes. Use
of 0.1% buffered peptone (prepared by adding 1  g
peptone to  1 liter of phosphate buffer) stabilizes
the sample for a much longer period. "Example 1"
of the buffered  peptone  data  exhibits a stable
count throughout the entire two hour test period.
In "Example 2", however, the counts begin to fall
outside the 95% confidence limits at a dilution age
of 60 minutes. Even with a 60 minute limit, how-
ever,  peptone stabilized  samples can  be  used  in
running far  more replicates  than  could be  used
with water or phosphate buffer dilutents.
     As illustrated  earlier, the confidence limits
around an experimental  mean depend not only on
the number of replicates, but also on the standard
deviation  of  individual  measurements. An experi-
mental value, s,  of  the  standard deviation can be
calculated from  each set of replicates. However,
there is a  particular source of scatter in membrane
filter counts which  is predictable, and it is possible
to generate a theoretical value of a, against which
experimental  values can  be  compared. The  pre-
dictable source of scatter is the fluctuation in  den-
sity of bacteria throughout a water sample.

     At best, bacteria in a carefully mixed suspen-
sion will be distributed  randomly through the sur-
rounding  medium. At worst, they may be associ-
ated with particles, adhering  in  clumps, or  con-
centrated  at the  walls of the container. If they are
O
o
 o
 o
 o
60 1
1 "
50 9
40 1
30 p

60
40 1
30 1
O
20 1

A
30 1
20 A
:!
0
0.1% BUFFERED PEPTONE - EXAMPLE 1 I
0 0 c
e
n n R u *~
U 0 8 0 ° MEAN 1
-J
1

9.1% BUFFERED PEPTONE" EXAMPLE 2 °
C
„ o I
o ° o o I
9 o c
o © o c
/•>,
° • 1

A PHOSPHATE BUFFER
• STERILE WATER
A A
A A ^
15 30 45 60 75 90 105 12
                                      195%
                                      CONFI-
                                      DENCE
                                      LIMITS
                                      FOR
                                      INDIVI-
                                      DUAL
                                      COUNTS
                                 AGE OF DILUTION, minutes
                   Figure 1. Effect of Diluent on Stability of Water Sample.

                                               27

-------
randomly distributed, samples of identical volume
will  not  contain  exactly  the  same number  of
bacteria.

     The  distribution of sample counts of parti-
cles  taken from a  random  distribution  in  a sus-
pending  medium is  called  a Poisson distribution
(2,  3). This distribution has a  standard deviation
equal to the square root of its mean. For the range
of counts of usual interest in bacteriology,  the
Poisson distribution can be closely approximated
by a normal distribution, which simplifies calcula-
tions considerably.  Using  this simplification, if
the  mean count per  aliquot  of sample  is 100
bacteria, the standard deviation ap due to random
distribution of bacteria through the liquid is>/100
or 10. The 95% confidence limits on  individual
counts are ±  1.96 OR or ± 19.6. In other words, if
the average count is 100, we can expect 95% of the
measured counts to fall between 80 and 120, if the
random distribution in the water sample is the only
source of scatter in the data.
                      however,  other sources of error arising from varia-
                      tions in experimental technique. These can appear
                      either as bias in the observed mean, X, or as scatter
                      in  the data,  an  increase  in  s. In  the  following
                      section, we will discuss how an experiment can be
                      designed to prevent an unwanted  bias from enter-
                      ing into the results.

                      Experimental Design

                          An example design  is a study of the effect of
                      ethylene  oxide sterilization  on  mixed cellulose
                      ester membranes.  The purpose of the experiment
                      is to isolate the effect of sterilization and hence it
                      is necessary  to eliminate effects of all  other vari-
                      ables, known and unknown. The  idea is to take a
                      group of  identical membranes, to sterilize half, and
                      to compare  these with  the other half, which are
                      not sterilized. The isolated variable is sterilization,
                      but what  about  possible unknown variables: are
                      the membranes  really  identical,  is the  experi-
                      menter's technique identical for each plate, is each
Case
  Number of
Replicates, n
Experimental
  Mean, X
Experimental
 Standard
Deviation, s
Predicted
Standard
Deviation, OR
Streak Plates
      18
   76.2
  14.6
    8.7
Membranes,
   Unsterilized

Membranes, Ethylene
   Oxide Sterilized
      18
      18
   119.4
   123.9
   17.3
  21.8
   10.9
   11.2
     In summary, with the best case, random dis-
 tribution of bacteria in the sample, we can expect
 a  fairly  large but predictable standard deviation.
 One example of using ap is given in the upper two
 graphs of Figure I.JThe value of OR is the square
 root of  the  mean  X, found by averaging all the
 counts in the experiment. Then  the predicted 95%
 confidence limits for individual  counts are ± 1.96
 \/X. These limits are  represented on the figure as
 dotted  reference lines, showing when the experi-
 mental points are within acceptable limits.

     Another  example is to compare an  experi-
mental standard deviation, s, with the theoretical
minimum standard deviation, OR. This will be done
in  a later section.

     So far only one contribution to the scatter in
 an experiment  has  been  considered.  There  are,
                      water aliquot identical,  is  each  plate  incubated
                      identically, is each counted the  same way . .  .?
                      Although  the  bacteriologist  exercises  the strictest
                      control  over  all of these  factors, it is  still possible
                      for one of these or an entirely unknown variable
                      to  intrude into  the experiment.  To  prevent un-
                      wanted  bias,  it is safest  to assume that unknown
                      variables are   present,  and to use randomization
                      techniques to eliminate their effects.

                           For  the  sterilization study 36 membranes
                      were  chosen   and  coded  for  identification.  Half
                      were  selected  by  random number techniques and
                      were gas sterilized. Sterilization details are given in
                      our  paper, "Optimum  Membrane Structures for
                      Growth of Fecal Coliform Organisms"  (1). Then
                      each of the 36 membranes was selected in random
                      order for testing. After  incubation, plates were
                      again randomized before counting.
                                                28

-------
     The reason for all these randomizations  was
to "mix up" the effect of any undesired variables.
Suppose, for example, that  the  technician who
counted the plates became fatigued and biased the
plates counted last towards  lower values. If all the
unsterilized  membranes had  been  counted last,
their  lower  counts would  have been  attributed
erroneously  to  the isolated variable under study.
Since, however, the plates were selected at random
for counting, the (hypothetical) counting bias ap-
peared scattered randomly throughout the results.

     In summary, one can appreciate the value of
random sampling by  assuming the worst: namely,
that in spite of your best efforts, undesired vari-
                    ables are present. By randomizing at each stage of
                    the experiment, the undesired effects are scattered
                    throughout the run so that they do not bias the
                    effect of the  isolated variable. To put this another
                    way,  the signal should be due to the variable under
                    study and all  other factors should appear only as
                    noise.

                          Results of this sterilization study are given in
                    Table 1.  It  is  useful to  compare experimental
                    standard deviations with OR, as suggested earlier.
                    Values  from  Table  1   are  summarized below.

                         Here, the three experimental  standard devia-
                    tions  exceed  the values  predicted for random dis-
       TABLE 1.   EFFECT OF ETHYLENE OXIDE STERILIZATION ON FECAL COLIFORM
                  COUNT
Replicate
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Mean,
X
Streak Plate
58
76
68
74
89
120
65
74
80
88
56
81
89
66
65
75
71
77

76.2
Membrane,
Unsterilized
127
128
129
85
115
117
143
142
135
114
129
102
135
125
112
126
102
84

119.4
Membrane, Ethylene
Oxide Sterilized
152
135
139
100
93
121
109
87
143
90
124
118
144
134
120
165
127
130

123.9
      Experimental
      Standard Deviations,*a
14.6
17.3
21.8
       'Computed from s2 = 2 (X-X)2 / (n-1)
                                              29

-------
tribution of bacteria in the sample. This indicates
that additional factors, such as the experimental
variables discussed above, have  contributed to s,
and that randomizing at  each stage of the experi-
ment was a necessary precaution.

     Some contrasting  data on standard deviations
are given  in Table 2, which  is a study of steriliza-
tion by three  methods (1). Comparing s with OR,
here, we see that two s-values agree well with ap,
two are considerably lower  than OR and one  s is
considerably higher than /n
                             95% Confidence
                               Limits on X*
Unsterilized Membranes
Autoclaved Membranes
Ethylene Oxide Sterilized
Membranes
Irradiated Membranes
Streak Plates
5
5
5
5
10
98.8
107.5
103.0
94.3
82.0
4.3
10.5
6.0
9.4
14.2
9.9
10.4
10.2
9.7
9.1
±8.7
±9.1
±8.9
±8.5
±5.6
                                               30

-------
tion is again useful. To randomize this variable it
would be necessary to list all the water sources of
importance for this type  of test and to  select at
random a group for testing. Then at the end of the
program  results  of all   experiments  would be
averaged together to eliminate any bias due to
water source effects. This kind of program  is in-
deed  a  massive undertaking and  requires the co-
operation of several laboratories over a long period
of time. We  should thus regard  the sterilization
studies presented  here, using two sewage sources,
as only a beginning. A fine example of an extensive
study involving three laboratories and a great vari-
ety of sources is given in "An Improved Membrane
Filter Method for Fecal Coliform Analysis" later in
this Symposium (4).

     A second important bacteriological considera-
tion is  the  selection of  a control method. The
spread plate was the non-membrane control used in
the  investigations of  Table  1 and  2. The spread
plate itself, however, is subject to a whole set of
procedural details, one of which will be described
here.  Table 3 presents  fecal coliform counts for
0.1 ml  samples of diluted sewage on two thick-
nesses  of M-FC agar. The same  experiment was
repeated on  seven occasions  using  fresh sewage
diluted in phosphate buffered peptone.  Attempts
to use agar thicker than 0.59 cm were not success-
full  due to the presence of spreaders. Looking at
the  overall  means, there is clearly a significant
increase in count with agar thickness.
                           The agar thickness effect  illustrates that the
                      spread plate control does not necessarily provide a
                      complete measure of the number of viable bacteria.
                      Although the spread plate does  not provide an ulti-
                      mate standard, we believe  that  it is important to
                      include some  non-membrane  standard in  mem-
                      brane evaluations.

                           Table  3 also  shows a possible water source
                      effect. The first six runs show a statistically signifi-
                      cant agar thickness effect,  while the last one does
                      not. If only the 12-27 run had been  performed, we
                      would have missed the effect altogether!

                                      CONCLUSION

                           In  conclusion, some of the factors entering
                      into successful  experiments evaluating membrane
                      filters have been discussed. To provide narrow con-
                      fidence  limits  on  a  mean  count, it  is useful  to
                      stabilize the water sample so that large numbers of
                      replicates can be employed. The scatter in replicate
                      counts,  which  also  affects  the confidence limits,
                      cannot  be  reduced below  a minimum which is
                      characteristic  of  the  random  distribution  of
                      bacteria in  the water  sample.  Procedural  details
                      also affect  the  counts, and it is important to ran-
                      domize  the  order of procedures so that unwanted
                      variables appear as scatter rather than as bias in the
                      means.  The sample  source is  also  an  important
                      variable which should be randomized. Finally, non-
                      membrane control  methods are desirable in mem-
       TABLE 3.  EFFECT OF AGAR THICKNESS ON FECAL COLIFORM COUNTS ON
                  SPREAD PLATES
       Test Date
       Fecal Coliform Count*
On 15 ml Agar         On 35 ml Agar
(0.25 cm thick)        (0.59 cm thick)
        * Each value is a mean of six replicates.
       ** At the 95% confidence level.
Size of Statistically
    Significant
    Difference**
11/12/74
11/14/74
11/15/74
11/20/74
11/21/74
11/30/74
12/27/74
Overall
40.8
30.2
68.7
24.8
48.8
60.8
27.3
43.0
49.7
46.8
94.3
46.7
69.8
82.6
27.7
59.7
7.6
7.0
10.2
9.3
8.7
9.6
5.9
3.1
                                                31

-------
brane evaluation experiments, but these are subject
to procedural variables, too, and do not necessarily
provide a complete  measure of  the  number  of
viable bacteria.
                 REFERENCES

 1.   Sladek,  K.J., R.V. Suslavich,  B.I. Sohn, and
     F.W.  Dawson.  Optimum  Membrane  Struc-
     tures  for Growth of Fecal Coliform Organ-
     isms. This Symposium, 1975.
 2.   Moroney,  M.J.  Facts From Figures, Penguin
     Books Ltd.,  pp. 96-107; pp. 220-1; pp. 261-3,
     1971.
 3.   Bennett,  C.A., and N.L. Franklin. Statistical
     Analysis  in   Chemistry  and  the  Chemical
     Industry. John  Wiley & Sons, pp. 115-8; pp.
     172-5; pp. 601-11, 1954.
 4.   Rose, R.E.,  W. Litsky, and E.E. Geldreich.
     An  improved  Membrane  Filter  Method for
     Fecal Coliform  Analysis.  This  Symposium
     1975.
                                                 Geldreich:   We see it  this way and I  just won-
                                                             dered  if you recognized this problem
                                                             of lot to lot variation.

                                                 Sladek:     I  think, the best  way that I  could
                                                             reply is to say,  if you want to study
                                                             sterilization, you  should  start  with
                                                             things that are identical. I think too
                                                             often  in the past people have found
                                                             differences between membrane filters
                                                             and may  have attributed  them  to
                                                             sterilization whereas in fact the  dif-
                                                             ferences were  from  other  causes.

                                                 Geldreich:   Part of this problem may be an inter-
                                                             reaction between ethylene  oxide  and
                                                             the membranes. Perhaps some of the
                                                             products on the membrane may  give
                                                             a  latent residual  effect.  Is there  a
                                                             problem?

                                                 Dawson:    Mr.  Geldreich  has  mentioned  the
                                                             possibility  of residual ethylene oxide
                                                             or  hydrolysis  products. Karl would
                                                             you  comment   on  residual  ETO?
         QUESTIONS AND ANSWERS

Geldreich:   In these studies on randomization you
            mentioned  18 membranes  that  you
            picked  at random.  Are you also talk-
            ing  about  18  random  lots of mem-
            branes?

            No, 18 membranes.
Sladek:

Geldreich:

Sladek:
Geldreich:

Sladek:

Geldreich:
How many lots are you talking about?    Sims:

This particular study was for the pur-
pose of  evaluating  this one effect,
ethylene   oxide  sterilization.   This
specific study was done with 36 mem-
branes  which  were  as closely  repli-
cated as we could choose.

Out of the same lot?
           Yes, they were identical.

           You  recognize that the variation  of
           membrane  filters  from lot to lot  is
           going to be another great variable that
           we are going to have to work with.  Is
           this correct?
Sladek:     Quite possibly.
Saldek:      I  would  be glad to. We have had
           analyzed  a  number  of  membranes
           that were sterilized by ethylene oxide.
           We had the  membranes analyzed by
           an outside laboratory, the best one we
           could  find.  They  never found any
           residual of ethylene oxide by chemi-
           cal  analysis. The limit of detection in
           that  test  was 4  parts per billion.

            One  problem that  will  be coming up
            over the  next two days is: what is a
            lot? Is a lot  the amount of raw mater-
            ials you mixed, that you manufactur-
            ed,  or that  you sterilized?  One lot,
            number 500, may  refer to each time
            the sterilizer was used;  it can be each
            time  a batch of  raw materials were
            mixed;  it  can  be  each  time  the
            machine  operated.  This word 'lots'
            can refer to at  least three  different
            possibilities.

 Sladek:     I would like to comment on that. Its
            true  that sometimes there are diffi-
            culties in defining a  lot. However,
            once  you  have decided  what a lot is,
            you get a lot of protection by doing
            random sampling  and  using a large
            number of  replicates.  For  example,
                                               32

-------
Litsky:
Sims:
we have often sampled  lots of mem-
brane filters  taking a sample of one
hundred membranes  drawn  at ran-
dom.  This makes certain that if you
didn't  recognize some difference be-
tween  lots when you were  defining
them, by random  sampling you will
pick  up  these  differences  in your
quality control program.

Would you give  us your definition of
a lot? What constitutes  a  lot? Would
you like to give an answer?
I  was actually asking what constitutes
a lot for him. For me a lot is one day's
machine  operation. In  our plant one
solution  is  prepared  per  day.  We
           sterilize about 4  times a day  and we
           have to check each load after steriliza-
           tion.  I  didn't know what they were
           calling a lot but the definition  of a lot
           can be  important when related to all
           of the  data  and  quality control  and
           experimental results.

Cotton:    In our terminology we refer to  a batch
           as that  material produced at any one
           particular period of time during a day.
           For  the lot  numbering purposes  we
           take a  section  of that batch and  we
           may treat it in one fashion or another.
           We may autoclave pack it, or ethylene
           oxide pack  it. That section  of  the
           day's batch  which  is  produced  and
           packaged  in  a particular way  is con-
           sidered  and given  a  lot number.
                                               33

-------
                            EFFECT OF INJURY ON THE RECOVERY
                      OF INDICATOR BACTERIA ON MEMBRANE FILTERS

                                    Review of Author's Research
                                        Alfred W. Hoadley
                                    School of Civil Engineering
                                  Georgia Institute of Technology
                                      Atlanta, Georgia 30332
                  ABSTRACT

     Investigations have  been undertaken  which
demonstrate  that indicator  bacteria,  which  are
injured during aqueous suspension or by exposure
to chlorine, may  fail to form colonies on mem-
brane  filters incubated  on selective  media.  Al-
though  dividing  cells  of  Escherichia coli  were
recovered equally well on Trypticase soy agar and
membrane  filters incubated on M-FC medium at
44.5 C, the efficiency of cell recovery decreased
with time  of exposure  to  stress.   Streptococcus
faecalis, in aqueous suspension, was recovered on
membrane  filters  incubated on  KF agar and on
M-Enterococcus  agar, with  an efficiency of  less
than  50%.  Recovery  of  streptococci  did   not
decrease with time of suspension.

     These findings helped to explain results  ob-
tained  by other workers and suggest  the need to re-
duce the selectivity of membrane filter techniques
against injured  cells. Comparative studies of  the
recovery  of decreasing, as well as growing, popula-
tions on  rich and  selective media ought to be in-
cluded in evaluations of selective media.
                INTRODUCTION

     Most  microbiologists have  been concerned
primarily with growth measurements of bacterial
populations. The food and sanitary microbiologists
are often required to estimate numbers of stressed
indicator bacteria and pathogens. This paper re-
views  recent observations on the enumeration of
declining bacterial populations in water, with spe-
cial reference to the recovery of indicator bacteria
on membrane filters.
Background

     In  recent years, an  extensive  amount of
literature has appeared on the recovery of starved,
heated,  frozen and  thawed  indicator  bacteria,
leakage  and degradation  of cellular  components,
and the repair of  cell  injury. A  limited amount of
literature has appeared on the recovery of bacteria
exposed  to disinfecting  agents.  However,  little
effort has been made to  evaluate the recovery of
stressed indicator  bacteria in aquatic environments.
     In  1961, McCarthy,  Delaney and Grasso re-
ported that "weaker" coliforms failed  to  form
colonies  on  membrane filters incubated on  M-
Endo broth.  They  found  that  pre-enrichment
on  lauryl tryptose broth, prior to  incubation on
LES .M-Endo  agar, improved  recovery from sur-
face water samples.  Klein and  Wu (3)  recently
demonstrated  significantly  higher  recoveries of
bacteria from  stream waters on spread plates than
on pour plates. These results indicate that a signifi-
cant portion of the heterotrophic  bacterial flora
in the streams consisted of injured cells that were
unable to tolerate the secondary stress of exposure
to melted agar at 42, 45, and 50  C. Hoadley and
Cheng (2)  demonstrated  injury  of indicator bac-
teria suspended in  a variety of aqueous suspending
media.  Injury prevented  recovery on  selective
media.

     In 1958, McKee, McLaughlin and Lesgourgues
demonstrated  reduced recoveries of coliforms from
chlorinated  settled sewage on membrane filters
that  were incubated  on dehydrated scheduled
nutrient  pads containing  an  Endo-type  medium.
On the  other hand,  recoveries of coliforms from
unchlorinated  settled sewage on membrane filters,
                                               34

-------
by most probable number tests, were in agreement.
By  applying  the pre-enrichment  procedure of
McCarthy,  Delaney and Grasso  (7),  Lin  (5) was
able  to  obtain  agreement  between  most  prob-
able  number values  and MF counts of total and
fecal coliforms in chlorinated secondary sewage on
membrane  filters incubated on M-Endo and  M-FC
medium.   Braswell  and  Hoadley  (1)  described
injury  to  Escherichia  coli  cells in  chlorinated
secondary sewage.

     It is the purpose of the present review  paper
to describe briefly the recovery of injured and lag
phase  indicator  bacteria  on  membrane filters
employing  standard  selective media.  Our results
suggested  the need  to reduce the selectivity of
membrane  fitter  techniques against injured  cells,
and the  need  to understand  cell  injury better. It
is encouraging that the recovery of injured bacteria
is receiving so much attention at this symposium.

     Hoadley  and  Cheng  (2) examined the re-
covery, on  membrane filters  incubated on stand-
ard selective media,  of E. coli and Streptococcus
faecalis after suspension for varying periods of time
in sterile  stream water,  double  distilled water,
phosphate  buffer (3.125 x  1Q-4M, pH 7.2), pep-
tone  water (0.01%), and  tap water.  Escherichia
coli ATCC 11775 and  Strep, faecalis ATCC 19433
were spread on plates of Trypticase soy agar (BBL)
and incubated for 18 to 24 hr. at 37  C and 35 C,
respectively.  Suspensions  of each organism  were
prepared in each  of the sterile experimental sus-
pending  media  and  compared  with McFarland
barium sulfate standards  to obtain an  inoculum
yielding  a  density of about 1000 organisms/ml
after  addition  to the test  suspension.  One liter
volumes  of test suspensions  were maintained at
20 C  in  water jacketed stirred flasks and samples
were  removed at intervals  of up  to 24 hours for
counting.

     Upon  removal,  samples  were spread immedi-
ately in triplicate on  Trypticase soy agar plates and
were   filtered  in triplicate  through  membrane
filters  (type  HA,  Millipore  Filter Corp.). Spread
plates  inoculated with  E. coli were incubated at
37 C and were counted at 24 and 48  hours. Mem-
brane  filters carrying  E. coli were incubated on
pads saturated with M-FC  broth (Difco) at 44.5 C
for 22±2  hours.  Spread  plates  inoculated  with
Strep,  faecalis were  incubated at 35  C and were
counted  at  48  hours.  Membrane  filters  were
incubated on KF broth (BBL) and M-Enterococcus
agar  (BBL) at 35 C  and were also counted at 48
hours.
     Counts  of  E.  coli in double  distilled  water
are presented in Figure 1. As a rule,  counts on
membrane filters initially resembled counts on the
rich,  non-selective  control medium.  But  as the
age of the suspension  increased, so  did the dis-
crepancy  between   counts on  the  two  media.
Counts in double  distilled water  are  presented
because the discrepancy between recovery on the
two media  is greater than in phosphate  buffer.
Many tap waters may be highly toxic, even  in the
absence of residual chlorine. As a result, recovery
on  M-FC  medium was  very poor (Figure 2). Re-
coveries from peptone water and  stream  water
(each of which  contained  between 545-550 mg/
liter total organic carbon) were  nearly identical.
In  each  suspending medium, growth followed a
lag  period (Figure 3). During the lag period, the
discrepancy between counts on selective and rich
media increased  until about 10 hours, after which
cells  recovered.   Following recovery  (18 hours),
counts were  nearly  identical on  the two  media.
Poor  recovery of E. coli on  a  selective medium
during  the  lag  phase  was  observed  also  by
Scheusner, Busta, and Speck (9).

     The  implications of  these observations are
clear. Cells of E. coli were injured  during suspen-
sion in water, either  as  a result  of leakage  or
degradation of cellular components. As  a result, a
substantial portion  of  the population  failed  to
produce colonies on membrane filters  incubated
on selective media.

     In contrast,  although recoveries of S. faecalis
were at all times low on  both selective media, there
was no evidence of progressive cell  injury in most
suspending media.  Recoveries of S. faecalis sus-
pended  in distilled  water  and phosphate  buffer
were  identical, and  remained  constant over a  24
hour  period  (Figure 4). However,  recoveries  on
both  KF  medium and M-Enterococcus  agar were
about one  half those on Trypticase soy agar. Re-
coveries  from  toxic tap  waters on  membrane
filters were very  much lower (Figure 5).  S. faecalis
grew both in stream  water  and peptone  water fol-
lowing a short lag (Figure 6). Counts on each selec-
tive medium  again were lower than those on the
rich medium,  but they reflected the behavior of
the population  as  a whole.  The  consistent  re-
coveries of  S. faecalis  from  most aqueous sus-
pensions on selective media is a desirable attribute
in a recovery medium.

Recovery of E. coli from Chlorinated Sewage

     Braswell  and  Hoadley  (1)  investigated  the
recovery on  membrane filters of  E. coli ATCC
                                               35

-------
    200
    100
O ^  1
o <   '
LU —
cc >
    0.1
I    '
            i    I    T
                               I    i    I    i    I    i    I    i
                       8      12      16      20
                      AGE OF SUSPENSION (hr)
                       24
28
32
  Figure 1. Recovery of E. Coli ATCC 11755 Suspended in Double Distilled
          Water on Trypticase Soy Agar (Circles) and on m-FC Medium
          (Triangles) (after Hoadley and Cheng, 1974).
     100
                             35
                            AGE OF SUSPENSION (hr)
      Figure 2. Recovery of E. Coli ATCC 11755 Suspended in Tap Water on
              Trypticase Soy Agar (Circles) and on m-FC Medium (Triangles)
              (after Hoadley and Cheng, 1974).
                                   36

-------
    1000
ig

SJ2

N? z
b=)

5- O
£ o

> %
O OQ
O <
UJ —
DC >
o
o
— »
o
                           — ^r
                                                i
8      12      16      20


  AGE OF SUSPENSION (hr)
                                                       24
28
               32
        Figure 3. Recovery of E. Coli ATCC 11755 Suspended in Stream Water

                on Trypticase Soy Agar (Circles) and on m-FC Medium

                (Triangles) (after Hoadley and Cheng, 1974).
     200


     100
   CO
 >• o
 £ 0
 O 03

 O <
 UJ —
 cc >
     0.1
                        8      12      16      20


                         AGE OF SUSPENSION (hr)
24
                                       28
       32
       Figure 4. Recovery of Strep. Faecalis ATCC 19433 Suspended in Double

                Distilled Water on Trypticase Soy Agar (Circles), KF Medium

                (Squares), and m-Enterococcus Agar (Triangles) (after

                Hoadley and Cheng, 1974).
                                   37

-------
    100
     80
u_ co 60
LU
     40
o 
     20
               1
   3              5
AGE OF SUSPENSION (hr)
     Figure 5. Recovery of Strep. Faecalis ATCC 19433 Suspended in Tap
              Water on Tryptlease Soy Agar (Circles), KF Medium (Squares),
              and m-Enterococcus Agar (Triangles) (after Hoadley and
              Cheng, 1974).
   1000
~°100
o£
O m
0<  10
11 I —  • v
8      12      16      20
  AGE OF SUSPENSION  (hr)
                                                       24
                                      28
32
    Figure 6. Recovery of Sterp. Faecalis ATCC 19433 Suspended in Stream
             Water on Trypticase Soy Agar (Circles), KF Medium (Squares),
             and m-Enterococcus Agar (Triangles) (after Hoadley and
             Cheng, 1974).
                                   38

-------
27622 suspended in secondary sewage and exposed
to chlorine. A suspension of cells was prepared, as
previously described, and was added to 500 ml of
sterile sewage in a 1000 ml beaker to yield approx-
imately  2 x 10^ organisms/ml. A sodium hypo-
chlorite solution was added to yield a dosage of 2
to 3 mg/liter and a total residual of 0.3 to 0.5 mg/
liter after 30 minutes. Ten  ml samples were  re-
moved after  1, 5, 10, 20, and 30 minute intervals,
and were placed immediately into 10 ml of sterile
sodium thiosulfate,  after which appropriate dilu-
tions were made. Diluted samples were spread in
duplicate on  Trypticase soy agar plates and filtered
in triplicate  through membrane filters (type HA,
Millipore Filter Corp). Spread plates were incubat-
ed at 37 C for 24 hours. Membrane filters  were
placed on pads saturated with M-FC broth (Difco)
and incubated at 44.5 C for 24 hours. In one ex-
periment, filters were  placed  on Trypticase soy
agar and  on pads saturated  with Trypticase soy
broth.

     Counts  of  E.  coli  in chlorinated  secondary
sewage are presented in Table 1. Counts on mem-
brane  filters  generally  were close to those  on
spread plates initially. With exposure of sewage to
Table 1.  Recovery of E. coli ATCC 27622 from
          chlorinated secondary sewage on Trypti-
          case  soy  agar  and  on M-FC  medium
          (after Braswell and Hoadley (1)).
Chlorin
contad
time
(min)
1
5
10
20
30
e
[
Expt
TSAC
16,000
1,600
1,000
10
6
Counts/ml
1a
M-FC
15,400
17
0
0
0
Expt
TSA
40,000
26,200
9,600
1,000
100
2b
M-FC
14,000
1,400
460
0
0
aTemperature, 21  C; pH  7.0; chlorine dosage,
3 mg/liter; chlorine  residual  after 30 min, 0.75
mg/liter.

bTemperature, 21  C; pH  7.0; chlorine dosage,
2mg/liter;  chlorine  residual after  30  min, 0.35
mg/liter.

CTSA, Trypticase soy agar.
chlorine, however, recovery of E. coli on filters was
decreased more rapidly  than recovery  on spread
plates.  In  each of  the  experiments reported in
Table 1, counts on the rich medium were still 1000
bacteria/ml  after colonies failed to form  on the
selective medium.  Scheusner et al (9) and Maxcy
(6) reported that  E. coli exposed to chlorine and
other disinfectants did not form colonies on violet
red bile agar.  Poor recoveries of coliforms from
chlorinated  settled  sewage on membrane  filters
reported by McKee, Mclaughlin, and Lesgourgues
(8), and enhanced recoveries following pre-enrich-
ment reported  by  Lin  (5), can be explained if it is
understood  that injury occurs during exposure to
chlorine,  thus preventing  recovery  on selective
media.

     A  single experiment was undertaken to deter-
mine whether injured cells form colonies as readily
on membrane filters as they do on  agar surfaces.
Spread  plates were prepared as usual and samples
were  filtered  through  membrane  filters,  then
incubated in triplicate  on Tyrpticase soy agar, and
pads  saturated  with  Trypticase soy broth.  All
plates were incubated  at 37 C for 24 hours.  Re-
sults are presented in Table 2.

     As in previous  experiments, initial counts on
membrane filters closely resembled those on spread


Table 2.  Recovery of E. coli ATCC 27622 from
          chlorinated secondary sewage on Trypti-
          case  soy  agar  (Braswell  and  Hoadley,
          unpublished)^
                                                                               Counts/ml
                                                   Chlorine contact
                                                      time (min)
                                Membrane filter

                  Spread Plate    TSA    TSBC
1
5
10
20
30
33,670
31,630
20,700
500
30
33,570
28,070
14,630
45
2
30,600
22,370
4,800
1
	
aTemperature, 21  C; pH  6.9; chlorine dosage,  2
mg/liter; chlorine  residual  after 30 min, 0.4 mg/
liter.

"TSA, Trypticase soy agar.

CTSB, Trypticase soy broth.
                                                39

-------
plates. However, after the initial sample, recoveries
were  lower on  membrane filters, and  again the
discrepancy increased as  time  of  exposure to
chlorine  increased.  Furthermore,  recoveries on
membrane  filters were poorer when  incubated on
pads  saturated with  Trypticase soy broth  than
when incubated on Trypticase soy agar.

Conclusions

     It is clear from the above observations that we
must pay more heed to the effects of injury on the
recovery of indicator bacteria from water. While
healthy,  multiplying E. coli cells recover  well on
membrane  filters  incubated  on selective  media,
cells exposed to stress may not  form colonies on
filters or selective media after suspension in surface
waters, chlorination of wastes, or chlorination and
passage of  potable  waters through distribution
systems. Furthermore,  it  might  be possible to
devise media,  superior  to those employed, on
which all viable cells are able to form colonies. The
significance of the discrepancy between  counts on
rich and selective media is  suggested  by the obser-
vations of  Speck  and Cowman  (11), that freeze-
injured  salmonellae  may be as pathogenic as un-
injured cells.

     Evaluations  of  media and techniques should
include examination  of  the  recovery  of  stressed
bacteria  in pure  culture.  Cells  of  Pseudomonas
aeruginosa  stressed in fresh waters  and estuarine
waters were  employed  by Levin  and  Cabelli (4)
to evaluate their M-PA procedure. Such evaluations
might be applied  profitably to existing, as well as
newly developed procedures.
                 REFERENCES

 1.   Braswell, J.R.,  and A.W. Hoadley.  Recovery
     of Escherichia coli  from chlorinated second-
     ary  sewage.  Appl.  Microbiol.  28:328-329,
     1974.
 2.   Hoadley, A.W., and C.M. Cheng. The recovery
     of indicator  bacteria on selective media. J.
     Appl. Bacteriol. 37:34-57, 1974.
 3.   Klein,  D.A., and S.Wu.  Stress: a factor to be
     considered  in  heterotrophic  microorganism
     enumeration   from   aquatic  environments.
     Appl. Microbiol. 27:429-431, 1974.
 4.   Levin,   M.A.,  and  V.J. Cabelli. Membrane
     filter technique for enumeration of Pseudo-
     monas aeruginosa.  Appl.  Microbiol. 24:864-
     870, 1972.
5.   Lin, S. Evaluation of coliform tests for chlor-
     inated  secondary  effluents. J.  Water Pollut.
     Control. Fed. 45:498-506, 1973.
6.   Maxcy, R.B. Non-lethal injury and limitations
     of recovery of coliform organisms on selective
     media. J.  Milk Food Technol. 33:445-448,
     1970.
7.   McCarthy,  J.A.,  J.E.  Delaney,  and   R.J.
     Measuring coliforms  in water. Water  Sew.
     Works. 108:238-243, 1961.
8.   McKee, J.E., R.T. McLaughlin, and  P.  Les-
     gourgues.  Application  of  molecular  filter
     techniques to  the bacterial assay  of  sewage.
     III. Effects of physical and chemical disinfec-
     tion.  Sew.  Ind.  Wastes. 30:245-252., 1958.
9.   Scheusner, D.L., F.F.  Busta, and M.L. Speck.
     Injury of bacteria by sanitizers. Appl. Micro-
     biol.  21:41-45, 1971.
10.  Scheusner, D.L., F.F.  Busta, and M.L. Speck.
     Inhibition of injured Escherichia coli by sev-
     eral agents. Appl.  Microbiol. 21:46-49, 1971.
11.  Speck, M.L., and  R.A. Cowman.  Injury and
     recovery  of  frozen  microorganisms.  J. Milk
     Food Technol.  34:548, 1971.
                   Discussion

Brezenski:  I would like to ask you one question
           with respect to the chlorination experi-
           ments. Do you have some of the other
           chemical characteristics of the second-
           ary effluent that you  use with respect
           to  solid and  other organic  removals?

Hoadley:   We don't  have parameters like  TOC,
           TSS,  etc.  These   samples  were just
           autoclaved  and  we  used  the  same
           batches as much as possible.

Brezenski:  We did some work several years ago in
           our  laboratory  comparing  the  re-
           coveries of membrane filter and MPN
           systems  from  primary chlorinated ef-
           fluents  of sewage treatment plants. We
           experienced atrocious recoveries. The
           MF's , in every case, were many  magni-
           tudes  lower.  But,  as we  progressed
           through  various stages  of  treatment,
           e.g., when  we went up to secondary
           treatment  and to  effluents  from good
           activated sludge plants, we  found that
           the results were coming closer together.
           When we came to  a tertiary plant, we
           found  that we  almost experienced
           comparable results. These results make
                                                40

-------
Hoadley:
Alico:
Hoadley:
           me wonder if we are dealing solely with   Alico:
           the  recovery of injured  cells. I  agree
           with  you  100%  that  we  do have  a   Hoadley:
           problem  of  cells  being stressed in the
           environment, and  also  in  chlorinated
           waters. I  also believe that we have to
           take  into consideration  the physical
           factors,   for  example,  the  level  of   Lane:
           suspended  solids.  It is quite evident
           from this data  that  if you  do  have a
           high  level  of  suspended  solids,  the
           membrane filter is not going to give a
           good recovery.  There  is no question
           that McKee, in his early work in 1958,
           showed that chloro-organic complexes
           form on  the membrane. And this  does
           have a growth-inhibition effect on these
           colonies.  I   believe we are not  only
           talking about a  stressed system,  we are
           also talking about a system  of physical
           characteristics coming into play. I think   Hoadley:
           the analyst, in this case, is going to  have
           to determine when he  sees the sample
           concerned,  how he is going to treat it.
           It becomes  difficult because now we
           have some subjectivity involved.
I agree. I  think there are many factors
that are  going to  influence our re-    Dawson:
coveries here. All I  can do is demon-
strate  a phenomenon. There  is an ef-
fect. Initially our recoveries are much
the same.  In all of this work we would
grow the  bacteria on plates, suspend
the  cells,  adjust the  turbidity  and
inoculate  our  samples.  This  took  7
minutes usually.  Then  we would run
those  things, initially the counts are
about the  same.

Two questions, the first is, how many
replicates  did  you run  and  secondly
what were  the means  of  sterilization
of the filters that you used?

We used the filters  from the  presteri-
lized packages as  they  came. Usually
tests were performed in triplicate. We    Hoadley:
have run  various experiments on dif-
ferent  strains from time to time, and
you  get variation  from  experiment to
experiment;  but  the phenomenon  is
there.
What  membranes  were  you   using?

I  don't think  the  brand is pertinent
here, because we are looking  at the
phenomenon;  but they were ethylene
oxide-sterilized.

I  would just like to point out here the
importance of osmotic pressure and its
effects on recovery of injured cells. For
example,  in clinical  microbiology and
blood cultures, if you take  one broth,
use it as is and add  10% sucrose to an-
other bottle of the same broth, you will
get a increased  recovery of injured cells
in the broth with the 10% sucrose. This
has been demonstrated repeatedly, and
I  think you may have a similar situation
here.

I  think there  are a  lot  of questions
about  the environmental conditions.
I  don't think we have good control of
that  in   our   experiments,  and  I'm
ready to admit that. I  imagine we were
not  doing  things  much  differently
than  other people, so this  is  the way
our techniques are applied.

Just a comment. I think  that osmotic
pressure may play an important factor
here but I think we also should look at
the quality of the water we use for pre-
paring dilutions in reagents. The gentle-
man from BBL indicated  that he would
like  to see DSP purified  water used as
reagent grade  water. Purified  water
USP-18 is not  all  that  good in the
light of modern-day technology. Single
distilled  water   can  contain a high
quantity of amines and we  know that
amines come over with a single distilla-
tion.   There   are  numerous  reports
published  in the  literature that amines
are toxic to both bacteria and to tissue
culture  cells  and I  wonder if  we are
seeing something here?

We  would never use single  distilled
water at all. It becomes  very clear in
any  of these  comparisons that the
better the quality of distilled water the
better your  recoveries are after sus-
pension.
                                               41

-------
                      EFFECTS OF TEMPERATURE ON THE RECOVERY OF
                                      FECAL COLI FORMS

                                       J. B. Hufham, Ph. D.
                                    Department of Life Science
                                   University of Missouri at Rolla
                 ABSTRACT

    The theoretical  concept of cell and  culture
death  resulting  from  increased  temperature  is
important to an understanding of the  inaccuracy
of the fecal  coliform  procedure. Results  of our
studies on the differential effects of temperature
on various types of membrane filters show that the
type of filter is  an important variable.  Recoveries
of known  densities  of pure  cultures  of  E. coli
varied  from  10-70  percent  from  this variable
alone. The importance of selection  of  the proper
temperature for membrane  evaluation  cannot be
stressed too strongly.

     Results  show  that  the  interaction  of the
membrane with  the  cell,  resulting  in loss of cell
viability,  is  temperature dependent  and varies
from  brand  to brand. A modified  fecal coliform
procedure has  been developed  and field tested.
This new procedure lessens the detrimental effects
of the  membrane on  recovery  of  the organism.

     While this  symposium  is entitled, "The Re-
covery  of Indicator Organisms  Employing  Mem-
brane Filters", I sense  that it is really an effort  to
re-establish the M-FC method, or a modification  of
that method, as recommended technique.  I sin-
cerely hope  that this  symposium and  the discus-
sions that follow will, in  fact, provide investiga-
tors with a workable method.


     It is no secret that there  have been several
 reports over the past  year and a half  which have
demonstrated poor  results with the M-FC method
 under some conditions. Most of these reports have
 indicated that the fault  lies in the quality of the
 membrane filter. As the author of one of those
 reports,  I am certainly not going to deny that con-
clusion.  I will modify it and say  that the  mem-
 brane quality gives  some of the error and  most
 of the variability.
     I  will  limit my presentation  to  the M-FC
method. The  subject of  my  presentation is the
effects of temperature on recovery, but in order to
fully  develop  my topic I  would  like to discuss
several  aspects of our evaluations  concerning the
procedures that were used. I  hope that I will not
tread on the data of our other speakers.

     In order  to evaluate the problems with the
M-FC procedure, one must first develop a method,
based on sound reasoning, which restricts the vari-
ables  with which one has to  deal.  I would like to
begin by  discussing  some of  these variables and
describing  our selection of a defined evaluation
procedure. (Tab. 1).

The organism

     Any organism selected  as a test  organism  in
evaluating  this method must  meet  several criteria:

     1) The organism must be a pure culture. This
is not to deny the use  of field trials; nor of their
importance in  the total evaluation of any analytical
procedure. It  is necessary, however, to first estab-
TABLE 1.  VARIABLES


A.   Organism
     1.   Must be a pure culture.
     2.   Must be easy to obtain by the investigator.
     3.   Must be E. coli.
     4.   Must not be selected for by the M-FC
         Procedure.

B.   Medium

C.   Filter membrane

D.   Temperature
                                               42

-------
lish the  validity  of  the test  procedure using a
known, pure culture of the organism which the test
is  supposed to quantitate.  I  stress  this because
some  have  proposed  using mixed cultures. Mixed
cultures  have  a way of changing their makeup
depending upon how  they are cultured.
     2) The  organism must be  easy  to  obtain.
Specifically it should be on deposit with a type
culture collection and the strain number identified
so that each investigator is not introducing his own
variable.  We have often found that we could not
compare  our  results with  other  investigators
because  of the organism employed  and the  im-
possibility  of isolating the same culture under  the
same conditions.

     3) The organism must  be E. coli. This  point
should be obvious. Notice, I did  not  say it should
be IMVIC positive. True, E.  coli  is, but that test is
not sufficient by itself to  identify the organism.
Such  studies have never been done. It does differ-
entiate the organism from Enterobacter aerogenes
and so far that  is the only use of the test. Nor did I
mention  other  criteria such as lactose fermenta-
tion at 44.5 C, etc. There may be other organisms
that meet  that test.  If it is  E. coli, it meets  all of
the criteria. But  if it meets all of the criteria, it
may not  be E. coli.

     4) The organism must not be selected for by
the M-FC Procedure. By using the M-FC Procedure
to obtain the test organism,  it is possible that one
might  obtain a  mutant more resistant to the inhibi-
tory effect. This would lead to  erroneous results
since the procedure itself is not designed to recover
only  those resistant types.  We  isolated  such a
colony and it gave 2.5 times  the recovery as our
stock culture.

     In our work, we selected  ATCC strain 11775
because  it  is the  neo-type strain, and we felt that
this organism best  represented the  organism  for
which  coliform tests were  originally  developed.
Cultures  were  grown in nutrient broth  for  16
hours at  35 C.

The medium

     The original investigators made the medium
from   scratch.  Most  people probably employ
commercially available products.  The question is,
are all products equal in  quality  and  do they vary
from  batch to  batch. We had  not had time to ex-
pand  our studies into this area, but we feel that
someone should.  I  am  also concerned  about  the
sources and quality  of the rosolic acid employed.
 We have used both Difco and BBL media with no
 apparent  difference. Our rosolic acid has always
 been from  Difco. In order to remove a variable
 we are using Difco  medium only, at the present
 time.

 The filter membrane

     We did some work  comparing  the different
 brands of membranes.  The early work  by  Press-
 wood  and  Brown, plus some  of our own  work
 comparing  Gelman and Millipore  filters, showed
 that while the brand of membrane has a decided
 effect  on  the recovery  in the M-FC  Procedure, it
 might  not be the only  effect. First we compared,
 Then we  chose Millipore because it was the worst
 and we wanted to know why. Then we went to
 Gelman because  it gave the best  results and we
 would  not  have  a second experimental  variable..
 And finally  we decided to use none because we
 couldn't  trust any of  them.  One gets that way
 when  he  opens a  new package and finds nothing
 between the blue paper.

 The temperature

     Temperature in  this method is very import-
 ant. 44.5  C is not hard to obtain, but most incu-
 bators and waterbaths have a hard time maintain-
 ing ± 0.2  C throughout  the chamber. For this rea-
 son  we used a homemade circulating waterbath
 that we had been using for our enzyme work. We
 adjusted and checked it  with a quartz thermometer
 at first, but we now  use a thermometer which has
 been checked against an NBS standard thermom-
 eter. I  have seen many people who simply reach in
 the  drawer,  pull  out a thermometer and use it.
 Perhaps we  need to stress standardization  more
 than we do.

     We have concentrated on  this variable, tem-
 perature, and it is these results which  I would like
 to discuss  today.

    What  effect does  temperature  play in the
recovery of  coliformsby the M-FC method?  Let's
look at the results of  several investigations compar-
ing brands of  filters. Presswood and  Brown  com-
pared Millipore and Gelman filters. They found a
mean loss  on Millipore  of 53 percent when  com-
pared with the Gelman  filter. In their study, they
concluded that temperature itself was not detri-
mental, but that the Millipore filter was. We dis-
agree  with  those conclusions.  Their organisms
were isolated with the M-FC Procedure and their
results  were compared to cultures grown on M-FC
agar pour  plates at  35 C.  Since  the difference
                                               43

-------
showed  up only  at  the  elevated  temperature  it
would seem that temperature did  have an effect.
The question  is, is the effect on the membrane, on
the cells, or both.  Recently Dutka, Jackson, and
Bell  reported  on  a comparison of  six different
filters.  None  of these filters gave  better than 38
percent  recovery at 44.5 C compared to a pour
plate control at the same temperature.

    When we compared Millipore with Gelman we
found that the Millipore recovered only about 10
percent  of the cells known to be in the sample.
(Tab. 2) The  Gelman  filter,  however,  recovered
about 55 percent.  Our standard was the number of
cells which grew on plate-count broth at 35 C. We
feel this is the only reliable criteron for knowing
how many viable  cells there are  in the sample.
TABLE 2.  RELATIVE ERROR IN THE FECAL-
           COLIFORM METHOD ASA
           FUNCTION OF THE BRAND OF
           MEMBRANE FILTER EMPLOYED.

                   No. of E. coli cells
                   per 100ml
Experi-
ment
No.
1

2

3

a Rplat
Filter
Brand
Millipore
Gelman
Millipore
Gelman
Millipore
Gelman
i\/p orrnr =
Total
Coliform M-FC
Broth Broth
(35 C) (44.5 C)
50
48
46
48
79
73
count 35
6
26
5
21
6
50
C-count 44.5
Relative
Error3
(%)
88
46
89
56
92
32
C
                      count 35 C
     One can  improve the  results with Millipore
filters by  incubating the stock culture at 44.5 C
for an extended period of time. (Tab. 3) After 120
hours at 44.5 C the  recovery on Millipore went
from 13 percent to 60 percent. Please notice that
the number of viable cells decreases with  time but
the ratio of those  countable at 35 C to  those  at
44.5 C  changes. It was thought  that by serially
diluting the culture, we could select for  100 per-
cent recovery.
TABLES.  RECOVERY DATA FOR E. COLI
           GROWN AT 44.5 C.
                  No. of E. coli cells
                  per 100 mla
Tube
no.b
1



2


3

4
Incubation
Time (1 hr)
(44.5 C)
24
72
96
120
48
72
96
24
48
24
Total
Coliform
1,180
300
20
30
920
240
110
1,780
510
1,440
M-FC
(44.5 C)
158
30
9
18
85
62
22
490
113
540
<%>
13
10
45
60
9
27
20
28
22
38
a.   Counts are given as cells per 100 ml of a 10
     dilution.
b.   Each tube is a serial inoculation of the
     previous tube.
     If temperature  alone  affects  the  recovery,
then it should be noticable without any filter. We
ran standard plate counts at various temperatures
and found that the count dropped as the tempera-
ture was increased  above  40 C. (Tab. 4)  Please
note that the recovery  is about 50-60 percent of
that obtained at 35 C.

     We have tried  to  improve  the technique by
lowering  the incubation temperature.  At  35 C
good differentiation was obtained between E. Coli
and E. aerogenes. (Tab. 5)  The same would prob-
ably be true at 40 C  with Gelman membranes
and we recommend  that this be field  tested by
those who  are set up to perform extensive field
trials.

     Let  me briefly summarize our conclusions.

     1.   Incubation of E.  coli at 44.5 C destroys
         the viability of at least 40 percent of the
         cells, even  in the absence of a membrane
         filter.
                                              44

-------
TABLE 4.  COMPARISON  OF POUR PLATE
           RECOVERIES OF E. COLI* ON
           PLATE COUNT AGAR AT
           VARIOUS TEMPERATURES.
Temp
30 C
35 C
41. 5 ± 0.5 C
44. 5 C
35 C
44.5 C
Average
Count
142
142
94
70
640
470
No. of
Plates
2
2
2
2
4
4
 * E. coli#11775

TABLE 5.  DIFFERENTIATION OF  E. COLI
           AND E. AEROGENES BY M-FC
           METHODS AT 35C.

                         No. of cells per 100 ml
Experiment Culture


1
2
3
4.


5.




E. coli
E. coli
E. aerogenes
E. coli
E. aerogenes
Mixture (l:l)
E. coli
E. aerogenes
Mixture (l:l)
TGY
Broth
(35 C)
58
51
59
—
—
—
—
—
—

M-FC Broth
(35 C)
59
48
58a
73
154a
230a
156
253a
389a
a.    E. aerogenes colonies were cream to light
     green in color.

     2.   Use of the Millipore  membrane inhibits
         another 50 percent of the cells. This is
         probably a heat soluble factor.

     3.   Previous  incubation  on Millipore mem-
         branes  at 44.5 C  would  select  for
         mutants   capable  of  growing  under
         these conditions.
     I believe that a test strain of E. coli should be
established. I  also believe that counts of this cul-
ture on  plate count  medium  at  44.5 C should
equal the counts on the same medium at 35 C.
Until they do,  I do not feel that the M-FC Pro-
cedure should even be considered.

Question and Answer Session

Grasso:   You made mention of your comparison
         studies of the fecal coliformson Gelman
         and Millipore  filters,  and mention  of
         numbers that were recovered. You didn't
         make  any mention  of colony size  or
         characteristic. Did you detect any  dif-
         ferences on the two filters at the higher
         temperatures?

Hufham: Yes. There is a tremendous difference.
         At 35 C, colonies were about the same
         size. When you get to  44.5 C, you find
         a lot  of variation. You  find pinpoint
         colonies and,  if you  let the  culture
         incubate a  little  more, you sometimes
         can  see additional  colonies  to count
         that weren't there  previously.  I  don't
         know what causes that.

Grasso:  The reason that I mention it is because,
         as I will state in my paper, we found
         marked differences at  44.5 C  in  the
         characteristics of fecal coliform colonies
         on  the Gelman and Millipore filters, as
         far  as size and color are concerned.  I
         think  that in addition to all the physical
         characteristics   and  media  problems,
         there  is also another problem; the actual
         definition of a fecal coliform.

Hufham: Let me answer that by saying we came
         across  so  many variable  we didn't  feel
         we  could  make a  good solid study.
         We  tried to  eliminate  as many variables
         as possible. We tried to get rid of media
         that seems  to  have  a  problem.  We
         tried to get rid of the membrane and we
         tried various cultures to see if they grew
         at  one temperature  or another.  We
         ended up in frustration.

Ginsburg: I want to make a comment about the
         rosolic acid. We found that early in our
         use  of M-FC tests we also had a problem.
         But now we do not incorporate rosolic
         acid and get comparative  results.
                                              45

-------
                     OPTIMUM MEMBRANE STRUCTURES FOR GROWTH OF
                                 FECAL COLIFORM ORGANISMS

                       K.J. Sladek, R.V. Suslavich, B.I. Sohn, and F.W. Dawson
                            Millipore Corporation, Bedford, Massachusetts
                 ABSTRACT

     The purpose of this study was to determine
the optimum membrane filter structure and charac-
teristics  for  recovery  of  coliform   organisms.
Additionally,  other  factors such as sterilization
method  and  membrane  composition   were ex-
amined.  Fecal coliform growth tests with varied
samples indicated that the  most critical factor in
recovery was  surface pore morophology and not
other factors previously suspected. Fecal coliform
counts showed a dramatic  increase with  increasing
surface  opening  sizes.  Membrane structures with
surface  openings large enough to surround the en-
trapped bacteria  are required for optimum growth
of fecal coliform organisms. Maximum fecal coli-
form  recoveries  are  obtained  using  membranes
composed of mixed esters  of cellulose exhibiting a
surface opening diameter of 2.4 ^m and a retention
pore size of 0.7 /im.
               INTRODUCTION

     Since its introduction as a  tentative method
for coliform enumeration in the 10th  Edition of
Standard  Methods in 1955,  the membrane filter
has  gained  wide acceptance  not  only for total
coliform,  but also for fecal coliform, total bacteria,
and  a wide variety  of  other  bacterial tests. The
unique advantage of  the membrane over other test
methods is  its ability to concentrate and  localize
bacteria from large samples. Hence, the membrane
increases  the  sensitivity of  quantitative  bacteri-
ology  into the range well  below one organism per
ml. Once  the bacteria are localized, the membrane
provides a  structure  for counterdiffusion  of nu-
trients and  metabolic products as well  as "hospit-
able" growth environment. In these functions, the
membrane differs little  from the earlier pour and
streak plate methods.

     The  earliest technique  for bacteriological
analysis  with membrane  filters involved  direct
microscopic  examination  of  bacteria  trapped
on  the membrane  surface. Here, the  optimum
structure required pores smaller than the organisms
being trapped for examination so that they would
lie  in  a single  microscopic  plane.  This surface
planar retention facilitated finding the organisms
under high power microscopy. The above require-
ments evolved naturally to the practice of retaining
organisms on  the membrane surface for various
culture  techniques.  At   that  time, not  much
thought was given to developing an optimal mem-
brane structure for colony growth.


     The ideal  characteristics of a membrane for
quantitative bacteriology  would appear to be pores
small enough  to retain bacteria but open enough to
provide paths for liquid transports, and a "hospit-
able"  surface  growth.  However,  upon examin-
ing  the  variety  of bacterial  methods utilizing
membranes, one finds  a  considerable  range of
bacteria sizes, types, and metabolic requirements.
These considerations led us to wonder if it was pos-
sible  to develop  membranes  which  would  be
especially favorable for the growth  of  particular
types of organisms,  such as  the  coliform group.
     The critical step in development of a colony
from a single bacterium is  the  onset of  cellular
division, and it is not unreasonable to expect that
this delicate process could  be affected by the ex-
tent and nature of  the contact of the organism
with the solid, and the extent and thickness of the
nutrient film surrounding the organism. Further,
nutrient supply  by  diffusion of medium  and re-
moval  of  subsequent metabolic waste products
must be a function of membrane  structure and
pore morphology.


     With  these factors  in  mind, we began this
study with the objective of  defining the optimum
membrane  structure for growth of coliform bac-
teria.
                                               46

-------
         MATERIALS AND METHODS

     Membranes were  obtained from  a variety of
commercial sources and  from our  experimental
membrane   development  activities.  The surface
structures  of these were characterized  using  a
Coates & Welter CWICSCAN 100-4 scanning elec-
tron  microscope.  Before observation,  the  mem-
branes were coated with  a  100 - 200 A  layer of
gold.

     Fecal  coliform and total coliform determina-
tions were  performed in accordance with Standard
Methods (1), Sections 408 A and B with the fol-
lowing modifications:  To achieve the closest pos-
sible  similarity between  membrane  tests  and
streak plate  control,  the  membranes were plated
on a 0.34 cm thickness of agar medium in 47 mm
petri  dishes; each  streak plate was prepared by
spreading a 0.1 ml aliquot of sample onto a 0.34
cm  thickness of agar in a  90 mm dish. The reason
for  using a controlled  thickness of agar is that we
had  found, in earlier  experiments, that fecal coli-
form recovery is a function of agar thickness  (2).


     M-FC  Agar and M-Endo Agar were obtained
from  the BioQuest Division of Becton, Dickinson,
and Company. Plates were stored at 5 C and were
used within 48 hours of preparation. Fecal coli-
form  plates were  incubated  at  44.5 C ± 0.2 C in
Blue  M  waterbaths equipped with  calibrated re-
cording  thermistors. Total coliform  plates were
incubated at 35 ±  0.5 C in circulating air incuba-
tors.

     Most  of the  water samples were untreated
sewage, obtained from the masher section of  the
Billerica, Massachusetts, Sewage Treatment Plant.
River samples were also used. Samples were stored
at 5 C and were used within 30 hours of collection.
     Some refinements of technique were needed
to allow  us to run  experiments involving large
numbers of samples.  Initially,  it was  found that
noticeable die-off occurred in 15 minutes when the
source water was  diluted with phosphate buffer.
The  use  of  0.1% buffered  peptone, however,
stabilized  the  count for a period of one hour  (2).
We also found that it was important  to restrict
the time  between  plating  and  incubation to 15
minutes or less. The complete procedure was then
as follows: A preliminary count was obtained when
the sample was taken. The following day, a dilu-
tion was prepared to give a count of 200 —  1,000
bacteria/ml, using  buffered  peptone diluent. The
 diluted sample  was mixed for  30 minutes  on a
 mechanical shaker. Then groups of about 18 mem-
 branes and 9 streak plates  were  prepared from 0.1
 ml aliquots,  plated, and  incubated. This was re-
 peated  throughout the  experiment.  Using  this
 method,  up  to  100 membranes  plus associated
 streak plate controls could be run within the one
 hour limit. To confirm  fecal  coliforms,  typical
 blue  colonies were transferred  into Lauryl Tryp-
 tose broth and then into EC broth.

 Surface Pore Morphology

     Membrane  filter structure  can  be character-
 ized by several  parameters. The retention pore size
 is a measure of the smallest particle which is  re-
 tained by the structure, and is  best measured  by
 direct determination  of passage of particles (or
 microbes) of known  size.  This  technique  is des-
 cribed by Rogers and Rossmoore (3).

     In  the present investigation,  we  were inter-
 ested not only  in  bacterial retention  but also in
 how  the  bacteria  are situated on  the  membrane.
 It is reasonable to expect that the  environment of
 retained  bacteria depends  on the retention  pore
 size as well as the structure of the  surface layer
 in which they are retained.

     Figure  1  gives scanning electronphotomicro-
 graphs of a series of eight membranes made from
 mixed esters of cellulose.  The photomicrographs
 show similar  structures which differ only  in the
 size of  the openings.   In  each  photomicrograph
 relatively  large surface openings were characterized
 by the surface  opening diameters reported on the
 Figure. These were determined by  direct measure-
 ments on each photomicrograph, or in  the case of
 the smaller  size  openings  by measuring enlarge-
 ments of the  photomicrographs. The retention
 characteristics of  these membranes  for coliform
 organisms were  determined  by  passage tests,  as
 described  in the following section.

     In  summary, the  way in which  bacteria are
 situated on a membrane is determined by  a new
 parameter, the surface opening diameter, which is
 observable from  scanning  electron  photomicro-
graphs. The retention of bacteria is determined by
the more familiar  retention pore size, which is
found from passage tests.

 Results

     Figure 2 shows fecal coliform counts on the
series of membranes described above. There is a
                                               47

-------
•j^sK.
 «m«-
*^te$
    &ESti,

        M

s£2?«&
  0.7,um
  2.0 ftm
          3.0//m
                                      0.8,um
                                     2.4yum
                              4.0yum
 Figure 1. Scanning Electron Micrographs of a Series of Mixed Ester of Cellulose

       Membranes. Numbers shown are Surface Opening Diameters.
                        48

-------
I^U
1—
o
o
I 80
o
LJ_
5 60
o
0
< 40
o
LLt
LL.
or>
A)
n
T
/^\
/ \
f
/
tL
I
I
MEAN OF 5
REPLICATES AND
95% CONFIDENCE
LIMITS. SEWAGE
SAMPLE.
SB \G^X'*'
' . . ^bS-'-"'' ,


•







   0     1.0   2.0    3.0    4.0
SURFACE OPENING DIAMETER, /u m
                                      STREAK
                                      PLATE
  Figure 2. Fecal Col i form Count versus Surface
           Opening Diameter.
remarkable increase  in counts at surface opening
diameters between 1.0 and 2.0 /im. The decrease
in counts at the largest  opening  size is evidently
due  to passage  of organisms through  this  very
coarse structure. The dotted line labeled "passage"
was obtained by re-filtering the effluent through a
bacterial  retentive  membrane and  plating  this
membrane on M-FC Agar in the usual way.  On the
basis  of both growth and passage tests, the  opti-
mum membrane structure was determined to  have
a 2.4  jum surface opening diameter with  smaller
(fecal-coliform  retentive) voids of  approximately
0.7 urn internally. Results of this plus three other
fecal coliform runs are given in Figure 3. In all four
runs,  the abrupt increase in recovery at a  surface
opening  diameter  of  1.0 to  2.0 Aim  is evident,
with  the optimum structure  -  i.e., zero passage
and optimum growth occuring with  2.4 /im  surface
openings.


     Typical  blue colonies were picked for confir-
mation from  the 0.7, the 1.4,  and the 2.4 /im sur-
face opening  membranes. The  ratios of confirmed/
picked were  18/20 for the  0.7, 19/20 for the 1.4,
and  17/20 for the 2.4 Aim surface  opening mem-
brane.
• SEWAGE SAMPLE
T SEWAGE SAMPLE
A MILL RIVER
D CHARLES RIVER
EACH POINT IS AN
AVERAGE OF 5
REPLICATES







FFpAl 1 1
ru^rtL ' * ~
COLI-
FORM 1<2
COUNT 1 Q
FECAL
COLI- 0.8
FORM
COUNT °-6
ON 2.4 Q4
jum
OPEN- 0.2
ING
MEM 0.0
o
n
X""?\
/* \y
T *.
/•
/
; /
n/D
- A6 PASSAGE ^ TX*
- 3* >•
* ' ' ->^*f '



T








1.0 2.0 3.0 4.0 STREAK
bKAINt RLATE
SURFACE OPENING DIAMETER, urn
                                              Figure 3. Normalized Fecal Coliform Counts versus
                                                       Surface Opening Diameter.
                                                   To compare  the  total  and  fecal coliform
                                               effects more directly, a culture was prepared from
                                               a typical fecal coliform colony from the Mill River
                                               source. This was  run  on M-FC agar and M-Endo
                                               agar using the same series of samples, all from the
                                               same dilution. Results  are presented in Figure 5.
                                               Here,  the total coliform test  shows only  a very
                                               slight  surface opening size effect while the effect
                                               is considerably more evident in the fecal coliform
                                               test.  Evidently, while the fecal  coliform  test re-
                                               quires a surface  opening diameter of 2.4  ^m for
                                               optimum growth,  the  total coliform test is less
                                               demanding and is performed well on membranes of
                                               surface opening diameter in the range, 1 to 3
     Figure  4 presents the results  of  two total
coliform experiments on the same series of filters.
Here, a  light effect may  be  observed occurring
only at the  smallest and  largest surface  opening
sizes.
                                                   At this pointjt appeared that surface opening
                                               diameter  was definitely  a  primary determinant of
                                               fecal coliform recovery.  However, other factors,
                                               such as  chemical  composition and  methods of
                                               sterilization, remained to be investigated.
                                               49

-------
               TWO SEWAGE
               SAMPLES. EACH
               POINT IS AN
               AVERAGE OF 5
               REPLICATES.
   50
o
°40
O
IJ 30
o
o
^ 20
   10
    0

     0     1 0   2.0    3.0  4.0   STREAK
                                  PLATE
   SURFACE OPENING DIAMETER,// m

Figure 4. Total Coliform Count versus Surface
         Opening Diameter.
              INCUBATION ON
             MF-ENDO AGAR, 35°
             INCUBATION ON
             M-FC AGAR, 44.50
   40
   30
 ID
 o
 o
   20
 o
 o
   10
       0   1.0    2.0    3.0    4.0   STREAK
  SURFACE OPENING DIAMETER,yw m  PLATE

Figure 5. E. Coli Counts versus Surface Opening
        Diameter. A Comparison of Total and
        Fecal Coliform Tests.
Effect of Chemical Composition

     In  the  foregoing  set of  tests,  membranes
employed were composed of mixed esters of cellu-
lose. A second series of experiments were designed
employing cellulose acetate membranes. Cellulose
acetate has a much smaller affinity for proteins,
and  presumably bacteria, than  does  the mixed
esters material  used  in the previous tests. Thus, if
surface adhesion affects growth, a difference be-
tween the acetate and mixed esters results should
be evident.
     In the next experiment, recovery on the 2.4
Aim (surface opening) mixed cellulose esters mem-
brane was compared with that of a 3.8 fj.m (surface
opening)  cellulose acetate membrane. In addition,
an experimental non-cellulosic membrane composi-
tion having a 3.0 ;urn diameter surface openings was
included.  Results are given in Figure 6. Here, we
have plotted the actual counts on each of five repli-
cates, and the passage count obtained by re-filter-
ing the effluents from each. The results show very
little difference in count between the three mem-
brane compositions.  These  results suggest that
                                                               • COUNT ON EACH
                                                                 MEMBRANE
                                                               x PASSAGE ON EACH.
                                                                 MEMBRANE
   o
   o
      50
                                                    30
                                                    20
   <  10
   o
   LLJ
   "-  o
REPLICA
NO.
     MIXED
      ESTER
     OF
     CELLULOSE
•

xx x x
X
•

•%••

12345 12345 12345 12345
/ / / /
                                                              CELLULOSE NON-       STREAK
                                                              ACETATE   CELLULOSIC  PLATE
                                               Figure 6. Fecal Coliform Count on Membranes of
                                                        Three Different Compositions.
                                            50

-------
membrane composition is not an important factor
in fecal coliform recovery.

Effect of Sterilization Method

     Several authors (4, 5) have  suggested that
bacterial  recoveries may be affected by the method
of sterilization.  They did  not, however,  present
data derived from comparing identical membranes,
where the only variable was the method of sterili-
zation. To test for possible  sterilization  effects,
membranes were selected from  the group exhibit-
ing optimum growth characteristics  (2.4 /im sur-
face   openings). These  membranes  were  then
divided into four groups using random sampling
techniques. One group was left unsterilized, one
was autoclaved  at 121  C for 15 minutes, one was
exposed to ethylene oxide using a standard sterili-
zation cycle * and was  aerated three days, and the
fourth group was sterilized  by irradiation at a dose
of 1.0 megarads using gamma  rays from  a cobalt
60 source. Mean counts and 95% confidence limits
on the means are given  in Table 1, and are discus-
sed in more detail in our other paper (2). There are
no significant differences between  counts on the
unsterilized membranes  and counts on the mem-
branes sterilized  by  the  three  methods used.
TABLET    EFFECT OF STERILIZATION ON
            MIXED CELLULOSE ESTER
            MEMBRANES HAVING 2.4;um
            SURFACE OPENING DIAMETER
Total Coliform
Count**
Unsterilized
Ethylene Oxide
Sterilized
Autoclaved
Irradiated
Streak Plate
42 ±6
44 ±6
38 ±6
40 ±6
50 ±6
Fecal Coliform
Count**
99
103
108
94
82
±9
±9
±9
±8
±6
    Two different sewage samples were used. Each
    mean is an average of five replicates.
  The cycle used a two  hour exposure of 12%
  ethylene  oxide at  130  F and  60%  relative
  humidity  (6).
Discussion

     The data  collected  to  this  point  strongly
suggest  that neither chemical  composition nor
method of  sterilization has any  significant effect,
but that the primary determinant of fecal  coliform
growth  on a membrane filter  is that of the surface
pore morphology (specifically with respect to the
size of upper surface openings).

     We speculated that since surface effects  are
strongest at surface void sizes which  are close to
coliform dimensions, some sort of fit of the organ-
ism  into the pore might  be required for optimum
growth. In  particular, the mechanism of the effect
could be that  organisms which  are deposited  on
very fine surface structures are  incompletely sur-
rounded by nutrient, while ones that fit  into sur-
face openings can be cradled below  the level of
nutrient that  is drawn  up  by  capillary forces.
Because of evaporation, an incompletely surround-
ed  bacterium  might be  subjected  to  a locally
hypertonic   solution,  with  resulting  plasmolysis
and  death. This effect would be particularly evi-
dent at the elevated temperature (44.5 C) of the
fecal coliform test.

     To test this hypothesis, three methods of sup-
plying nutrient were  compared.  The  0.7 ;um and
the  optimum  2.4 p.m surface opening  cellulose
ester membranes were used.  One  set was plated
in the standard manner, one set was plated face
down on the  M-FC  Agar, and the third set was
plated right side up with 2.0 ml  of M-FC  Agar
overlayed onto each membrane.  Results are  sum-
marized in Table 2.

     Due to the confluence of colonies,  accurate
counts could not be obtained  from the membranes
placed face down. However,  it was clear  that the
number of  colonies  on  membranes having the
smaller surface openings (0.7 /jm) was substantially
increased by placing face down.  Overlaying  these
membranes gave a dramatic increase in counts. The
increase  in  growth thus seen from  inverting the
filter, plus  the close  agreement  in counts of the
two  membrane  groups when the  lower yield filters
were overlayed  with nutrient,  gives strong  evidence
that complete nutrient coverage  of the organisms
is  required  and that this is  achieved only  with
larger surface opening sizes.


     During the comparison  testing of the mem-
branes for the 0.7 /zm and 2.4 jum surface opening
groups,  some additional  benefits were noted rela-
                                               51

-------
           Figure 7. A Fecal Coliform Organism Cradled within the Surface Opening
                   Diameter of a Type HC Filter (15,500X).
TABLE 2   EFFECT OF PLATING METHOD AND PORE SIZE FECAL COLIFORM TEST,
           SEWAGE SAMPLE
Membrane Plated and
 Overlayed

Streak Plate
                                     Mean Counts and 95% Confidence Limits
Method of Plating
Membrane Plated in
Standard Manner
Membrane Inverted on Agar
Membranes with
0.7 MTI Surface Openings
14±3
Approx. 30
Membranes with
2.4 nm Surface Openings
44+ 10
Approx. 45
46 ±7
53 ±8
               35 ±5
                                        52

-------
tive to the latter. These predictable, but nonethe-
less important, phenomena were an increase in the
flow rate through the membrane, an increased dif-
fision rate of media to the membrane surface and,
significantly,  increased  capacity  to filter  large
volumes  of water particularly  those where algae
or other collodial turbidity would otherwise  limit
the sample size.

     In  summary, the factors expected to have an
effect  on  fecal coliform  recovery  were  investi-
gated.  The only  one showing  a significant effect
was that of surface pore morphology. The evidence
suggests  that  unlike  other organisms, fecal  coli-
forms specifically must be cradled slightly below
the membrane surface for optimum recovery. This
suggests  an  optimum membrane structure  with
surface pores slightly larger than the fecal coliform
organisms,  but  with internal  bacterial  retentive
pores. See Figure 7.
     Until  now,  membranes   recommended  for
bacterial testing have been specified by a retention
pore size of 0.45 Aim. Typical  0.45  Aim retention
membranes  have surface  opening  diameters  of
1 to 2 jum. As can be seen in  Figures 2 and 3, a
slight shift of position on  the curve in the range of
1 to 2 /zm surface openings can  have a large and
significant effect on recovery.

     Since membranes of  different manufacturers,
all having  0.45 jum  retention  size, may exhibit
differences in  surface morphology (i.e.,  in relative
surface opening diameters) they may also exhibit
considerable differences in fecal coliform recovery.

     A  change to the optimum 2.4 /im surface
opening  size  will  not only provide higher  fecal
coliform counts,  but will  also lead to a smaller
sensitivity to small differences  in surface morpho-
logy.

     For the  total coliform test (Figure 4 and 5),
however, membrane  performance  is not sensitive
to surface morphology (except  in the range below
1 /urn surface opening  size). The new 2.4 Aim  sur-
face opening/0.7 ^m  retention pore size membrane
developed in  this work should  be regarded as an
improvement  for fecal coliform  tests, and may also
be used for total coliform with results equivalent
to 0.45 Aim retention  membranes.


                REFERENCES

1.   Standard Methods for the  Examination  of
     Water and  Wastewater.   13th  Ed.,  APHA,
     1971.
2.   Sladek,  K.J., C.F.  Frith,  and  R.A.  Cotton.
     Statistical Interpretation of  Membrane Filter
     Bacteria Counts. This Symposium, 1975.
3.   Rogers,   B.G., and   H.W.  Rossmoore.  Deter-
     mination  of  Membrane  Filter  Porosity  by
     Microbiological  Methods.  Dev.  Ind. Micro-
     biol. 11,453, 1970.
4.   Presswood,  W.C.,  and  L.R.  Brown.  Com-
     parison  of Gelman and Millipore Membrane
     Filters  for  Enumerating   Fecal   Coliform
     Bacteria. Appl. Microbiol., 26, 332, 1973.
5.   Dutka,  B.J.,  MJ.  Jackson, and  J.B.  Bell.
     Comparison  of  Autoclave  and   Ethylene
     Oxide-Sterilized  Membrane  Filters  Used  in
     Water  Quality   Studies.   Appl.  Microbiol.,
     28,474,  1974.
6.   Kereluk, K., and R.S. Lloyd. Ethylene  Oxide
     Sterilization.  J.  Hospital   Research,  7,  7,
     1969; Cycle  is shown in  Figure  32, p.67.
         QUESTIONS AND ANSWERS

Hufham:    We did a study which we are not quite
            through with  yet, in which we washed
            membranes  in  weak sodium bicar-
            bonate at 50 C, then in distilled water,
            and then reautoclaved. We tried to see
            if we  could get an inhibitor out, and
            got an increase of 50% on our counts
            with your particular filter. We were a
            little  afraid  however, that we were
            distorting the filter. I wonder if you
            would  comment on a procedure like
            this;  might we be  making the pore
            sizes bigger as you  describe  in your
            paper?

Sladek:     It's  possible  that something like boil-
            ing  or strong autoclaving can have a
            slight  effect  on- the  surface morphol-
            ogy of the membrane.

Ginsberg:    I, and  perhaps  others,  don't  quite
            understand what the difference is be-
            tween  the pore size and the surface
            opening size?

Sladek:     I am glad that you asked that, because
            I  would  like to  make  it absolutely
            clear.   With   filtering    to   remove
            particles or bacteria  we are interested
            in retention of particles  and bacteria.
            That is  the way we characterize our
            material,  by  retention   pore  size.
                                               53

-------
Bordner:
Sladek:
Bordner:
                                                  Sladek:


                                                  Grasso:


                                                  Sladek:
The   way  you  measure  retention   Grasso:
pore  size is  by retaining something.
That  is the only certain way to char-
acterize a membrane for retention, by
performing a passage test. Generally
these are done with microorganisms
since the size  distribution  is quite
narrow.  I can  give  you some refer-
ences to the procedure. This is the re-
tention  pore  size; it's something that
is  experimentally  determined.  You
can  only speculate  how  this relates
to  the structure.  You  can't  really
make a  microscopic study  using  a
scanning  electron   microscope   and
say what the retention pore  size is.
I  would like to  view the retention
pore size and the surface morphology
characterized  by   surface   opening
diameter as two  separate and inde-
pendent  parameters  characterizing  a
membrane. In  fact,  they aren't  en-
tirely independent. As you see, one  is
constrained  as  to  what  you   can
manufacture, so that as you  increase
the  surface  opening diameter   you
also   increase  the  retention pore
size, generally speaking, for any given
type of membrane.
Your remarks are certainly interesting
and I think you have given us a lot of
new  information.   I   don't  know
whether  to go  home and pour selec-
tive  agar over  all  my  membranes or
throw them all out and buy this new
proposed material that you describe.
Do I  understand that these are experi-
mental materials with  the larger  sur-
face  porosity  that you  are  talking
about, not the ones that I have been
buying recently?
                                                  G rasso:
                                                  Sladek:
           If you give the  right answer I  might
           not  come  back tomorrow  and give
           my paper.  Do you feel that with this
           increased  pore size  the  temperature
           effect and all the other variables men-
           tioned today are overcome? In other
           words, you could  use this new filter
           with the larger surface pore size and
           with the M-FC  broth and overcome
           these difficulties and problems?


           Yes, I think you have asked me if this
           is the cause of all  the controversy?

           Right.


           I  think to  a  large  extent, yes. In the
           figure that I gave on the fecal coliform
           count  versus  the   surface opening
           diameter, this slope was very steep on
           the  left.  This means that different
           membranes that are  manufactured by
           various  companies  striving for  the
           same retention pore size were not con-
           trolling  the surface  morphology  dir-
           ectly.  The surface  morphology was
           down in  a range where  it was a very
           sensitive parameter with regard  to the
           fecal  coliform test.   I think we  can
           look  forward to a  period  when all
           those differences will go  away,  be-
           cause now we move  up further on the
           curve towards the peak.
           Are  these experimental  membranes
           going to be available?

           As far as  I  know  they  will be very
           shortly.
Sladek:
These  materials that  I  showed  you
were made specifically for this test.
Now they do have a lot of similarities    Sladek-
to   present  0.45  /inn  membranes.


 Then  are we talking about the possi-    Dawson:
bility of a membrane filter formulated
specifically for fecal coliforms?
 Yes, indeed we are.
Seidenberg:  How much vacuum did they use to
            pull those samples through, and what
            is your recommended vacuum?
            Let  me refer  this  question  to our
            bacteriologist, Mr. Dawson.

            I don't think vacuum has any tremen-
            dous effect on whether or not an or-
            ganism  is   impinged deeper  into  a
            smaller pore or  remains near the sur-
            face. It's been known for  a long time
            that the  gram  negative  microorgan-
                                               54

-------
           isms are extremely sensitive to hyper-
           tonic  solutions.  Indeed this is  the
           recommended method  for  preparing
           protoplasts from gram negative organ-
           isms. I believe in our laboratory we
           normally  use  something   like   14"
           vacuum for running samples. I've been
           working with membranes since about
           1955 and  I  haven't seen any effects
           that were due to vacuum.


Brezenski:  I  think  I  recall  one paper that dis-
           cusses  cavitation  and  the  effect of
           vacuum on coliforms, so I don't quite
           believe what  you said.


Levin:     Two  questions really — one is on the
           steep  part of the  curve, where  you
           were talking  about the surface effects.
           I  didn't  really  catch it. When  you
           boil or autoclave you go up in terms
           of improving recovery or do you go
           back  down  in  terms  or  decreasing
           recovery after treating them in your
           own laboratory?

Sladek:    We  have  found  very  little effect of
           either procedure. I showed  you some
           data on autoclaving.


Levin:     It really doesn't make any difference?

Sladek:    No.

Levin:     The second question is only indirectly
           related but since you're a statistician
           I  thought that  I  would  put  you
           on a  spot, if I could. You've given
           data  in two  papers and your counts
           have ranged from as low as 9 or 10 up
           to about  120 for the  average count
           and yet we think of 20 to 80 as being
           the optimum numbers. I'm wondering
           where you would draw the line? For
           instance,  if  I'm doing  a  one hun-
           dredth, a tenth, one  ml and 10 ml,
           and my 10  ml  averages 12, and my
           1  ml  averages 100, which  one do  I
           believe and  have the  most faith in?
Sladek:      Yes, let me comment on that. I'm sure
            that you realize in terms of the statis-
            tics  that you  want  to  avoid  low
            numbers, so it  is to your advantage
           to  get up  into  the higher numbered
           range. We have  not  established for
           certain  whether there is an  upper
           cutoff beyond which  you should go.
           We have had  good results  going up to
           about 140.  I  mean  "good" in the
           sense that  the scatter  didn't increase
           extraordinarily.   I've   seen  data up
           beyond  200 where the  scatter  was
           really terrible and I am not sure how
           safe it is to conclude that above 200
           something  goes  wrong .  . . perhaps
           someone else has some information
           on that.

Litsky:     Did  you examine brand X for cavita-
           tion and if so, what did you find?


Sladek:     You are asking  if we examined other
           brands of  membrane  filters so far as
           their surfaces go? Let me  answer that
           in this way.  I would  like to refer to
           my  earlier  paper. I stressed there the
           idea  of  random sampling.  Random
           sampling is a very important concept
           in not producing bias in your data. We
           have tested  what  I will  call  a  few
           boxes of competitor's membrane fil-
           ters. I will  not tell you the results be-
           cause I  know that this does not repre-
           sent a random sample of their produc-
           tion. I can tell you the results that we
           have obtained on Millipore  filters be-
           cause we are very  careful to take a
           representative random sample of our
           production and  we base  results that
           we report on that. I can't really com-
           ment in  public about other manufac-
           turers' filters because  I don't have a
           proper random  sample of their  pro-
           duction.

Litsky:     May I  take my prerogative as the
           chairman and ask any  other represen-
           tatives,  Gelman,  Johns-Manville or
           anyone else, if  they  examined their
           filters  and  observed  the cavitation
           effect?

Sladek:     Please don't call it a cavitation effect.

Litsky:     Litsky stands corrected.

Brezenski:  I was going to mention that cavitation
           isn't what we are talking about here. I
           want to get  back to surface morpho-
                                               55

-------
Sladek:
Brezenski :
Sladek:
logy and  the  reasons that you  gave
why there was no  effect on the re-
covery of the total  coliforms,  yet
there was  a decrease in recovery. You
assumed this because of the surface
morphology and that fecal coliforms
stay on  the  surface  and  don't get
deep down into the layers where they
can get more medium. I  haven't seen
any data that you've presented which
shows  specifically  that  surface  mor-
phology  effects the  fecal  coliform
and not the total  coliform. Do you
have any  data which shows the  total
coliforms  in the upper layers and the
fecal coli,  or  E. coli  in  the bottom
layers? This seems to be the crux of
your explanation.


No. This is not the crux of my explan-
ation.
Oh,  I
that?
am  sorry. Would you clarify
The organisms are of course the same
size. They are  from the same group.
We are speaking  of what happens to
an  organism that is filtered on a very
fine surface  structure,  and ends  up
"on a moutain  top". The explanation
was that because of evaporation, hav-
ing to do with dehydration rates,  va-
por pressures, etc. a fecal coliform  or-
ganisms may be in contact with a pool
of  nutrients, which will shrink in size
and consequently become  concen-
trated, or locally  hypertonic, causing
plasmolysis of the organism. If  this is
indeed  the  mechanism, you  would
expect it to be much  less pronounced
at  a  lower temperature where  the
evaporation  rate is lower. In trying to
tie all of these  things  together,  we
have shown  rather conclusively that it
is the way  in  which  nutrient is sup-
plied  that is the  origin  of the effect.
We did this by  turning the membrane
over and  pouring agar  on top of it,
etc. We have  also  shown  that  the
effect  is very strong in  the fecal coli-
form  tests,  but  rather  weak  in  the
total  coliform  test.  The  way   we
bridged the gap and put all this back
           together has to do with the evapora-
           tion rate.

Winter:     I just have one quick question. I can't
           really quite understand evaporation as
           the  reason, or  should I  say the hypo-
           thesis,  because we are dealing with a
           saturated humidity chamber whether
           you are on M-FC in a bag or in a super
           tight fitting plate.  I wonder whether
           this  really  has  an effect.   Aren't
           we  really  reaching  a  bit to  explain
           why something happens that we don't
           know anything about?

           Secondly, although you may not have
           tested many Gelman or other filters,
           it would  seem ironic  that  some of
           the  other manufacturers are ahead of
           you, because they, by some process,
           were able  to  prepare  a  membrane
           with larger  surface openings, hence
           large recoveries. If we just look at the
           summary of results they seem to con-
           clude that under present manufactur-
           ing techniques, Gelman  seems at least
           to have given higher recoveries by the
           majority of experimenters and no one
           has  explained this fact. We may have
           some random error  occuring, and  we
           know  we have  systematic  error in
           everyone's work, but this is random-
           ized, by the number of manufacturers
           putting  out  membranes   and  the
           number of people  doing  the work,
           when we have  had perhaps as many as
           50  people  coming  up  with  results
           which all  tend to  point in the same
           direction.  I  am wondering aren't  we
           reaching a bit at this time?


Sladek:     Let me respond to your first question.
           Aren't  we reaching  a little  in the ex-
            planations? Yes, indeed we are reach-
            ing  in   the  explanation. Concerning
            your comment  that it is  a  closed
           system, with respect to water vapor,
           there are always some small gradients
            in these systems and water vapor does
            not move around inside the plate, so
           that is  about the only support that  I
           can give to the idea that small changes
            in concentration can occur.

            Again,  with  regards to  your other
            question,  I  would have to  defer any
                                               56

-------
comments on what the  real status of
different  membrane  manufacturers'
membranes are, pending a  study.  I
believe  the D-19 round  robin study
has been  completed and  I  certainly
would not accept your statement that
everyone agrees that this kind of filter
is  better  than  that  kind, because
everyone doesn't.
                                   57

-------
                     A COMPARISON OF MEMBRANE FILTERS AND MEDIA
                         USED TO RECOVER COLIFORMS FROM WATER

                                 M.H. Brodsky and  D.A. Schiemann

                                    Ontario Ministry of Health,
                                    Laboratory Services Branch
                               Environmental Bacteriology Laboratory
                                Toronto, Ontario. Canada M5W 1R5
                 ABSTRACT

     Many  laboratories involved in water analysis
are using  membrane  filtration  methods for  the
enumeration  of  pollution indicator organisms in
water. The  Ontario Ministry of Health, Laboratory
Services Branch, analyzes approximately 350,000
water specimens annually, almost  exclusively by
membrane  filtration. We  have observed that there
are pronounced differences in the  abilities of the
filters produced  by various companies to recover
coliforms. This paper reports the results of an in-
vestigation  which evaluated three brands of mem-
brane filters. Seven of our laboratories participated
in this study.  Parallel analyses for  total coliforms
from routine water samples were performed using
filters supplied by the Johns-Manville Company of
Canada (045 MO 47SG), the Millipore Corporation
(HAWG  47SO)   and  the  Sartorius  Company
(11456).  Statistical   evaluation  of  the  results
indicated that the  Johns-Manville  and  Millipore
filters were equivalent and much  superior to the
Sartorius filters for the enumeration of coliforms
from water.

     LES Endo agar is the only solid medium re-
cognized  by Standard Methods for the  Examina-
tion of Water and Wastewater (13th ed. 1971) for
the direct  recovery of coliforms from water. M-
Endo broth media are also recognized by Standard
Methods  for use in  membrane filtration; however,
it is recommended that these broth media prepara-
tions be used  with  sterile pads. This latter proce-
dure as outlined in Standard Methods, adds an
additional  time factor to the processing of  each
water specimen which would present difficulties to
a high volume laboratory such as ours.

     The relative cost of  these two types of Endo
preparations as well as the problems created by the
recently  experienced  shortages  of  Endo  based
products prompted us to compare LES Endo agar
with various  M-Endo  broths with agar added for
coliform analyses.
               INTRODUCTION

     Specifications for membrane filters used for
the bacteriological analysis of water are presented
in the  13th edition of "Standard Methods for the
Examination of Water and Wastewater".

     All manufacturers of membrane filters claim
or  intimate  in  product  advertising  that  their
filters  meet  the  criteria  specified in  Standard
Methods for  bacterial recovery.  Recent  investiga-
tions indicate that, despite manufacturers' claims,
there exists  considerable  variation among com-
mercial  brands of membrane  filters in their ability
to recover coliform and faecal  coliform organisms
from water. There is, however, some disagreement
concerning  experimental designs  and statistical
evaluations used  in  these studies. Consequently,
any general conclusions to be drawn from such in-
vestigations must be guarded.

     The  Laboratory  Services  Branch  of  the
Ontario Ministry of  Health analyzes  more  than
350,000 water  specimens  annually,  almost  ex-
clusively by membrane  filtration. Our laboratory
personnel  had  also  observed  inconsistencies in
coliform and faecal coliform enumeration on mem-
brane filters produced by various companies. As a
result of these observations we carried  out a series
of  investigations  to  quantitatively compare coli-
form  and  faecal  coliform  recoveries on  three
brands of membrane filters — Johns-Manville, Milli-
pore and Sartorius.
                                              58

-------
         MATERIALS AND METHODS

     The comparative evaluations  were done in
three phases. Phase 1  was a preliminary field study
involving five regional laboratories. Phases 2 and 3
were more  rigidly controlled  investigations con-
ducted solely in our central  laboratory.

Membrane Filters:

     The Millipore filters (Catalogue  No. HAWG-
047SO) and  the Sartorius  filters (Catalogue  No.
11456)  used throughout the  investigation were
obtained from the  Ministry of Health Laboratory's
stock supplies.  The Johns-Manville filters  (Cata-
logue Nos.  045M047SG and 045M047LG) were
supplied by the company.

     Three different lot numbers of the three filter
brands,  pre-sterilized  by ethylene oxide, were in-
cluded for  comparison in   the  first  and  second
phases of this investigation.  For the  third phase,
unsterile Johns-Manville filters were obtained  and
autoclaved in our laboratory.

Source of Cultures:

     In  the  preliminary phase  1 of the study we
used routine water samples. Appropriate dilutions
of the samples  were filtered  in duplicate for total
coliforms only. All the  participating laboratories
did  not  stock the same brands of filters.  Three
laboratories  compared  Johns-Manville  with Sar-
torius  filters and   two  others  compared  Johns-
Manville with Millipore filters.

     Similar routine water  samples  were used for
the  second   phase of  this study.  After  being
analyzed by our routine procedure for total  and
faecal coliforms, these water samples were  refrig-
erated overnight. The  following day, those samples
having at least  20 faecal  coliforms  per 100 ml
were selected for  further  processing.  Ten  ml of
each of  these samples  were added to 10  ml of
double-strength MacConkey broth in screwcapped
fermentation tubes. Following incubation for 24
to 48 hours at 35C, two loopfuls of each positive
broth culture were subcultured into EC broth. The
EC broths were  incubated  at 44.5C for 22 to 24
hours.  Five  replicate  filtrations of a  dilution of
each  EC broth  culture, standardized by  optical
density, were performed for both total and faecal
coliforms for each of the three brands of filters.

     In  phase 3  of the study, we collected eight,
one-litre  samples of water from a known polluted
surface source, the  Number  River, over  a  one
week  period. On the  day of  collection,  a  pre-
screening  membrane  filtration was done on each
sample, to determine coliform and faecal coliform
densities.  Based on the screening densities,  appro-
priate test  dilutions of the  refrigerated samples
were prepared  to  provide 20  to 80 total coliform
colonies and 20 to 60 faecal coliform colonies per
filter. Ten  replicate  filtrations per sample  per
brand of  filter were completed  on each  water
sample for total and faecal coliforms.

Cultural Techniques:

     Throughout  the  study,  M-Endo  MF  broth
(Difco) with 1.5% agar added was used for total
coliform recovery, and M-FC  broth base (Difco)
with  1.5% agar added was  used to culture  faecal
coliforms.   These  solid  media  were  prepared
in 15 x 150 mm plastic petri plates, which accom-
modate 5  filters.  Incubation  times  and tempera-
tures were as specified in Standard Methods.  The
M-FC plates were  heat sealed in waterproof plastic
bags  before  being immersed  in a constant tem-
perature water bath at44.5C.

Statistical Analysis:

     The Students t  test for comparison of means
was used to evaluate the results of the preliminary
field  study. Eighty-two comparisons between  the
Johns-Manville and Sartorius filters were tabulated.
Sixty  eight comparisons  between Johns-Manville
and Millipore filters were analysed.

     Analysis of variance (ANOVA) was applied to
the results of phases 2 and 3 of this investigation.
When the  F ratio indicated a significant difference
in the means at the 5% significance level, a  multi-
mean comparison  test (the  Tukey Test) was used
to determine where the difference occurred.
         RESULTS AND DISCUSSION

     Our results  clearly demonstrate  that  varia-
tions in experimental design  can lead to very dif-
ferent conclusions regarding the superiority of one
MF brand over others. Tables  1 and 2, summarizing
the  statistical  analysis of the  preliminary  field
study, show  that  the  Johns-Manville filters  were
superior to Sartorius filters but equivalent to Milli-
pore filters for  total coliform recovery. But the re-
sults  of phase  2  conflicted with this conclusion.
Recovery of  faecal coliform  isolates on  m-Endo
medium (Table 3)  concluded  that the three brands
                                               59

-------
Table 1:    Total coliform recovery from routine
            water samples in three laboratories with
            Johns-Manville and Sartorius membrane
            filters.
No. Comparisons

Means (x)

Standard
  Deviation

  t-Test
   Analysis
Johns-Manville      Sartorius

       82              82

      14.9             8.1


      15.50          11.94

  t = 3.17
  (t .05, 162= 1.92}
of filters were equivalent; but, recovery on M-FC
medium  concluded  that  Johns-Manville  filters
were  superior to both the  Millipore  and the Sar-
torius filters (Table 4).  There was no significant
difference  in faecal  coliform recovery  between
Millipore and Sartorius filters.

     We attempted to resolve this conflict of data
by the experimental design of phase 3  using natural
water samples. Statistical  analysis of total coliform
recovery by the three filters (table 5) supports the
findings of  the  preliminary study, i.e. the Johns-
Manville  filters  were  superior  to  Sartorius filters
but equivalent to Millipore filters for total coliform
recovery. As in the preliminary study, these results
conflict with the results of phase  2.  Similarly,
statistical  analysis of  faecal  coliform  recovery
rates  from natural water samples (Table 6) indicate
that  there  was no difference  among the  three
Table 2:  Total  coliform recovery  from  routine
         water  samples in  two laboratories with
         Johns-Manville and Millipore  membrane
         filters.
Johns-Manville
No. Comparisons
Mean (x)
68
192.3
Millipore
68
184.5
Standard
  Deviation

  t-Test
Analysis
      499.60

  t= .05
  (t.05, 134= 1.98)
335.80
                                Table 3:   Recovery of faecal coliform isolates on
                                          M-Endo medium with  Johns-Manville,
                                          Sartorius and Millipore membrane filters.
                                                              Johns-Manville   Sartorius   Millipore
          No. Comparisons 100

          Mean (x)          52.0
          Standard
           Deviation
                   30.70
100

 44.5


 30.87
100

 43.7


 32,28
                                         Analysis of Variance (ANOVA)

                                Source          SS       df     MS     F ratio3

                                Within    290885.70  297     979.31

                                                                           2.14

                                Between    4194.06     2    2097.03


                                a F .05°° = 3.00
brands of filters at a significance level of .05. This
finding  also disagrees with phase 2  of  the  study,
which used  laboratory cultures  of faecal coliform
isolates. Similar disagreements have been described
by other  investigators. Presswood and Brown (4),
and  Harris (3)  concluded that Gelman  membrane
filters were  superior to Millipore membrane filters
for  recovery of E. coli. Schaeffer et al  (5) dis-
agreed with their statistical conclusions. Schaeffer's
group found when  using natural water samples,
that  Gelman and Millipore filters were  equivalent
for  faecal  coliform  recovery, but,  that  Gelman
filters were  superior  to  Millipore filters for total
coliform recovery.

     In  a recent paper,  Dutka  et al  (2) reported
conflicting  results in two  studies employing the
same experimental design. Field  samples and broth
cultures  of  E. coli  ATCC 25922 were used for
comparative recoveries with autoclaved and ethy-
lene  oxide sterilized filters. The results of their first
study (March  1973) concurred  with  findings of
Schaeffer et al (5). At a significance level of .01,
Gelman and Millipore filters were equivalent and
superior to Sartorius filters  for faecal coliform
recovery;  however,  Dutka et al  (2), also reported
that  both  Millipore and  Sartorius  filters were
                                                60

-------
Table 4.   Recovery of faecal coliform  isolates on
          M-FC medium with Johns-Manville, Sar-
          torius, and  Millipore membrane  filters.

             Johns-Manville  Sartorius  Millipore
                                                   Table 5.  Total coliform recovery from  Humber
                                                            River water  samples with  Johns-Man-
                                                            ville, Sartorius, and Millipore membrane
                                                            filters.
No. Compariso
Mean (x)
Standard
Deviation

	 JUIIIIb-IVIdllVII
ms 100 100 100
No. Comparisor 80
40.5 30.3 33.1
Mean (x) 49.1

32.47 26.75 28.19 Standard
Deviation 25.81
lie oariurius iv 	 _uuie
80 80
34.7 42.7


22.01 27.12
         Analysis of Variance (ANOVA)

Source     SS          df    MS        F ratio3

Within  25387.02    297      854.78

                                        3.22

Between  5496.17       2    2748.08


     Multimean Comparison Test (Tukey Test)

Tukey          Johns-Manville      Millipore vs
Calculation13  vs Millipore vs Sartorius  Sartorius

(Xl-x2)±Tc   +17.Q3 to    +19.83 to +13.48 to
              - 2.33       + 0.47     - 6.88

Conclusion      Not                     Not
 (


-------
TableB.  Recovery  of  faecal  coliforms  from
         Humber  River   water  samples  using
         Johns-Manville, Sartorius and Millipore
         membrane filters.
            Johns-Manville   Sartorius  Millipore

No. Comparisons  70          70        70

Mean(x)       30.1        26.7       26.0

Standard
 Deviation    12.39       10.15      10.40


          Analysis of Variance (ANOVA)

Source        SS      df        MS     F  ratio3

Within   25150.87     207    121.50

                                        2.75

Between    668.41        2    334.20


a F .05°° =3.00

the mean coliform  counts on the Johns-Manville
filters were statistically greater than on Millipore
and Sartorius  filters.  The autoclaved Johns-Man-
ville filters  however,  were noticably more brittle
and  less  flexible than  the ethylene oxide  ster-
ilized filters. We  also  observed  distortion of the
faecal  coliform  colonies on autoclaved filters.

     An  additional  undesirable features  of the
Johns-Manville and  Millipore filters was the inhibi-
tion of growth by grid markings. Colonies growing
near the grid lines developed flat edges, conforming
to the restrictions imposed by the  lines. Colonies
which straddled the lines were split. Interestingly,
grid  line interference  was more  pronounced  with
total  coliform than  with  the  faecal  coliform
colonies.

     We also noted, as did Dutka et  al (2), that the
Sartorius  filters had  irregular hydrophobic  areas
which became evident when  the filters were wet-
ted. These  areas of  reduced permeability  likely
•contributed to the decreased  bacterial recovery we
observed  with Sartorius  filters. This was brought
to  the  attention  of  the Sartorius representative
about six months ago and we have yet to hear a
satisfactory  explanation  of  why this  has  occur-
red.

    We  realize that our culture media, M-Endo
MF broth with agar added and  M-FC  broth base
with agar added, are not recognized as standard
solid media  for membrane  filtration by Standard
Methods. LES Endo agar is the only solid medium
accepted  for  the direct  recovery of  coliforms
from water.  However, we have conducted a com-
parison of coliform  recovery on M-Endo MF broth
with agar and  LES Endo agar by membrane filtra-
tion of  101  natural water samples (unpublished).
Statistical analysis  indicated  that there was  no
significant difference between coliform  recovery
on these two media.

    The conflicting conclusions of our investi-
gation  and  of other  similar studies  comparing
membrane  filters  need  to  be  resolved.  Experi-
mental designs including source  of the test organ-
ism and  statistical evaluations, must be standard-
ized  so  that  some  logical  conclusion regarding
membrane filter performance can be made.
            ACKNOWLEDGEMENTS

     We wish to acknowledge the technical assist-
ance of Mr. B. Ciebin.
                REFERENCES

1.   American Public Health Association. Stand-
     ard  Methods for the  Examination of Water
     and  Wastewater. 13th. ed.  American Public
     Health Association Inc., New York, 1971.
2.   Dutka,   B.J.,  M.J.  Jackson and  J.B.  Bell.
     Comparison  of Autoclave  and  Ethylene
     Oxide-Sterilized  Membrane  Filters  Used  in
     Water Quality Studies. Applied Microbiology,
     28:474-480, 1971.
3.   Harris,  F.J. Coliform  Recoveries on Mem-
     brane Filters. E.P.A.  Newsletter No. 22:  4,
     1974.
4.   Presswood,  W.G. and  L.R.   Brown. Com-
     parison of  Gelman  and Millipore Membrane
     Filters for Enumerating Fecal  Coliform Bac-
     teria.  Applied   Microbiology,  26:332-336,
     1973.
5.   Schaeffer,  D.J., M.C.  Long, and K.G. Janar-
     dan.  Statistical  Analysis of the Recovery  of
     Coliform Organisms  on Gelman and Millipore
     Membrane  Filters.   Applied  Microbiology,
     28:605-607, 1974.
                                               62

-------
      QUESTION AND ANSWER SESSION

Geldreich:  You mentioned that the use of the
           agar preparation of M-Endo and M-FC
           is a non-standard method. I would like
           to  say  that  in the  next edition of
           Standard  Methods we  have become
           so concerned about this problem, that
           I have written a paragraph saying that
           that you can  certainly use it  with
           1.5% agar, and it will  be a standard
           method. In fact, when I go out and
           do  laboratory  evaluations I certainly
           recommend it in my report.

Brodsky:   We  ran into problems, getting  Endo
           base media from  various companies.
           We  were  trying  to  find  alternate
           procedures and this is  how we got
           into it.

Brezenski:  Could   you  please  summarize.  For
           example you  said  for EC positive
           organisms  that  Johns-Manville  was
           equivalent to Millipore. Would you go
           over this? I think this is the crux of
           the  issue, and  I am a little confused
           because someplace where the line goes
           equivalent  from here to here, it's not
           equivalent from here back again.

Brodsky:   I must  admit that I was confused too.
           As  you realize  the study was divided
           into 3  phases.  Perhaps  I should  leave
           phase   1 out completely,  because  I
           really don't want to base a judgement
           on   such a loosely controlled study.
           Let  me do the 2nd and 3rd phases
           which  are  more  complete  and  more
           rigidly controlled. In phase 2 we  selec-
           ted  EC positive cultures, but for the
           sake of argument we will  call  them
           fecal coliform  cultures, and then we
           compared their recovery, that is EC
           positive cultures  on M-Endo media
           and  on M-FC media  in parallel,  using
           the  three same brands of filters. We
           did  5   replicates  for each   filtration.
           We  determined  that  Johns-Manville
           and  Millipore  and  Sartorius  were
           equivalent  when  used  with  M-Endo
           MF  medium  for  these  EC cultures.
           When we used M-FC  medium,  obvi-
           ously   selected for  fecal  coliforms,
           we   found  that  Johns-Manville  was
superior to Sartorius; however, Johns-
Manville  was equivalent to Mi-llipore,
and  Millipore was equivalent to Sar-
torius. Now,  there's some  logic  if you
try and think if A=B,  and B=C, then
A must = C. Think  of it  this way, if
A is 40, and  B  is 35, and C is 30. The
difference between 40 and 35  is not
significant. The difference between 35
and  40 is not significant,  but the dif-
ference between 30 and  40 may  be
significant. It is confusing but that is
the best explanation I can give you as
to  why  this  natural  logic doesn't
apply.

Then in phase 3 we said "fine." Let's
see  what  happens  now  if  we  use
natural cultures or a natural source of
water, rather than using a laboratory
culture in which we have  given it the
best possible condition to grow and
allow them to overcome any possible
inhibition.  We were hoping to find
some  sort   of  parallel.  Perhaps  I
shouldn't   say  that,  because  that
isn't  fair  scientific judgement. What
happened   was  for total coliforms
from a natural polluted sample such
as the Number River, (I think anyone
from the Toronto  area can vouch for
the fact that the  Humber River is a
polluted water source)  we found that
Johns-Manville  was  superior to Sar-
torius for  total   coliforms.  If you
recall  in  the first  part we  said that
they  were  equivalent  for total coli-
forms using  the laboratory cultures.
Johns-Manville  and  Millipore   were
equivalent and  Millipore and Sartorius
were equivalent for total coliforms.
When we looked at fecal coliforms we
didn't find any difference at all be-
tween the  three brands of filters for
recovering  fecal   coliforms from  a
natural polluted source. We performed
two   different  studies;   one   used
laboratory  cultures and   one  used
natural samples. We got two different
results. We have to resolve this prob-
lem.  The  question  that  I'm   asking
is where do  you  get your source of
culture?  Obviously  the  source  of
culture is going to  have a tremendous
influence on conclusi6ns.
                                              63

-------
                     COMPARISON OF MEMBRANE FILTERS IN RECOVERY
                             OF NATURALLY INJURED COLIFORMS
                                               by
                               David G. Stuart, John E. Schillinger and
                                       Gordon A. McFeters

                                    Department of Microbiology
                                     Montana State University
                                     Bozeman, Montana 59715
                 ABSTRACT

     Raw sewage and cultures  of  E. coli  were
exposed  to  natural  stream  conditions  in test
chambers for 24 hours then recovered and enum-
erated with the M-FC test with and without enrich-
ment and with different brand filters.

     ANOVA, F and  t statistics showed that the
Gelman  filters had  only a  slight  advantage  in
recovery over Millipore  filters.  Both brands  re-
covered  significantly  better than the Nuclepore.
Enrichment gave significantly better  recovery with
variation by days. Pure cultures were more sensi-
tive to test and filter variations than were the sew-
age cultures. The major conclusion was that Nucle-
pore membranes should not be used for coliform
analyses in water.
               INTRODUCTION

     Beginning with Dr. William G. Walter's initial
investigation (1964) into the relative bacteriologi-
cal  quality of water produced  by adjacent open
and closed watersheds, numerous studies of water
quality  in  high  mountain  watersheds have been
carried out in our laboratory using Millipore mem-
brane filters exclusively. Presswood  and  Brown's
article in  1973 (6) along with papers and discus-
sions at the 1974  American Society for  Micro-
biology meetings in Chicago, indicating that Milli-
pore  filters might  be yielding  erroneously  low
counts, caused some concern about the data  col-
lected over the last 10 years. With recent publica-
tions (2, 4, 7)  adding conflicting data and  interpre-
tations to the issue, it seemed wise to follow the
advice  of Geldreich  et al  (3)  and compare the
performance of different brands of membranes in
our laboratory.

         MATERIALS AND METHODS

    The preliminary  experiments reported here
were performed with suspensions of  Escherichia
coli C320 MP 25, isolated from water in our labor-
atory  and with  raw  sewage. Procedures followed
Standard Methods (1).

    Aliquots of 24 hour cultures of E. coli were
washed  twice with gelatin  phosphate buffer and
dilutions yielding 1()5  to  106 cells per ml were
placed  in chilled river water, taken immediately to
the stream site, and submersed in the flowing river.
A  1 ml sample was taken from the  chamber, placed
in  9 ml gelatin phosphate buffer, iced and  trans-
ported   back to  the  lab  where  dilutions  were
plated on TSY agar (35 C) to yield a 0 time count.
After 24 hours of exposure to the natural aquatic
environment  to allow  injury to  occur,  a   1 ml
sample of the contents of the chamber was trans-
ported  to the lab in 9 ml of iced gelatin phosphate
buffer.  One ml of this cell suspension was placed
in  9 ml of Trypticase soy  broth + 0.5% glucose +
0.3% yeast extract and incubated at room tempera-
ture for 2  hours before filtering  (enriched). An-
other  1 ml was taken from the buffer suspension,
diluted, and filtered  immediately (non-enriched).
M-Endo  MF  medium (35 C) was used with the
E. coli experiments. (mEndoMF-E. coli).

    The sewage experiments were  performed in an
identical manner except that undiluted raw sewage
was placed in the membrane filter chambers and
                                              64

-------
M-FC fecal coliform medium  was used and incu-
bated at 44.5  C (M-FC-sewage).

     Colonies were counted with  the aid of a bin-
ocular microscope (7X)  and reflected light. Counts
were  statistically analyzed with  regression analy-
sis and analysis of variance utilizing the classifica-
tions: filter, day and enrichment.
         RESULTS AND DISCUSSION

     Results from three M-FC-sewage runs with 10
replicates  with  each  brand  of filter  (Millipore,
Gelman and  Nuclepore) on  non-enriched  and
enriched  samples  along  with   four  identical
mEndoMF-E. coli runs are reported in this paper.

     Counts from the 10 replicate plates for each
filter brand  generally  followed  the Poisson distri-
bution. The mean counts on Gelman filters were
slightly higher than those on Millipore filters. Both
Gelman and  Millipore  filter counts  always  ex-
ceeded counts on Nuclepore filters.

     All  counts were transformed  using  square
roots and  analyzed with a computer program for
ANOVA. Both  F and t  statistics were computed
for various  interactions  of  means by using  the
Day-enriched-filter interaction mean square as the
error term.  Use of  an interaction  mean square as
the error term  is  common  in  randomized block
designs. Here one would  consider a day as a block.

     Day and enrichment differences were found
to be significant at the P  = 0.005 level for both the
M-FC-sewage and the,mEndoMF-E. coli situations.
Enrichment  effect varied  significantly with differ-
ent  days  (i.e.  interaction) for  both  mEndoMF-
E. coli  (0.005) and M-FC sewage  (0.05).  Day by
filter interactions and enrichment by  filter inter-
actions were not significant for M-FC-sewage  but
were  significant (0.005  and  0.05 respectively) for
mEndoMF-E. coli. This would  seem  to indicate a
greater  injury effect  for the washed  E. coli  cells
than  for the raw sewage coliforms. The in-stream
conditions varied during the  course of the  experi-
ments resulting in changing degrees of injury which
are reflected in the above statistics.

     Differences among the 3 filters (all treatments
grouped) were significant at the 0.01  level  for
M-FC-sewage and at the 0.005 level for  mEndoMF-
E. coli. Analysis of all Millipore counts versus all
Gelman counts showed  no  statistical  differences
 in either M-FC-sewage (0.2)  or mEndoMF-E.  coli
 (0.5) trials while  differences  between  Millipore
 versus Nuclepore counts and  Gelman versus Nucle-
 pore counts  were significant. These differences
 were much larger for the mEndoMF-E. coli situa-
 tion (P = 0.001) than for the M-FC-sewage differ-
 ences (P = 0.01 to 0.005) again suggesting a greater
 degree of injury for the E. coli compared to sewage
 coliforms,  and subsequently,  some kind of injury -
 Nuclepore filter interaction. If  enrichment  had
 overcome  this inhibitory  effect  with  Nuclepore
 filters,  one would expect  Nuclepore versus  en-
 richment  t values to  be higher than  those of
 Millipore and  Gelman versus enrichment. This  was
 not the case as shown by mEndoMF-E. coli t values
 of 49  and 54 for Millipore and Gelman versus en-
 richment effect  and a t value of 41 for Nuclepore
 versus enrichment  effect.  Although  the t values
 were smaller  (3.9,  4.7, 3.2)  this interpretation  is
 corroborated  by the  results  of  the  M-FC-sewage
 experiments.

     A  breakdown  of filter comparisons  into non-
 enriched  and  enriched  trials showed  Millipore
 versus Gelman differences to be insignificant for
 M-FC-sewage (0.5, 0.4) but were barely significant
 (0.05, 0.10) for mEndoMF-E. coli results. Millipore
 and  Gelman  were  significantly  different  from
 Nuclepore at the 0.005 level in the mEndoMF-
 E. coli situation  and from 0.05 to 0.005 in the case
 of the M-FC-sewage.


     The  practical  significance of the differences
 between filter counts should  be examined with an
 understanding of day to day changes in the natural
 environment and of errors and variation inherent
 in  membrane  filter techniques. For  example,
 assuming that a  set of replicate counts follows the
 Poisson distribution, the square root of the mean
 will  give  a reasonable  estimate  of  the standard
 deviation   one  can expect   between  individual
 plate counts. Thus, for the mean of 52.6 observed
 for the M-FC-sewage Millipore counts, a standard
 deviation  of ±6.7 would be expected. The actual
 difference  between Millipore  and  Gelman means
 was  observed  to be 7.7 which is about equal to
 the expected variation among  Millipore counts.


     For all practical purposes, these data indicate
that  the bias towards higher counts obtained with
Gelman as compared to Millipore filters is small.
The  difference between Gelman and Nuclepore
means  was 33.9, much larger than the  expected
standard deviation  of 6.7. Thus, the error  when
using Nuclepore  filters would  be  considerable.
                                               65

-------
5.
             LITERATURE CITED

     American Public Health Association. Standard     7.
     Methods for the Examination of Water and
     Wastewater, 13th ed. American Public Health
     Association, Inc., New York, 1971.
     Dutka, B.J.,  M.J.  Jackson,  and  J.B. Bell.
     Comparison of autoclave and  ethylene oxide-
     sterilized  membrane  filters  used in water
     quality studies.  Appl. Microbiology.  28:474-
     480,1974.
                                                  filters for enumerating fecal coliform bacteria.
                                                  Appl. Microbiology. 26:332-336, 1973.
                                                  Schaeffer, DJ.,M.C.  Long and K.G. Janardan.
                                                  Statistical  analysis of the recovery of coliform
                                                  organisms  on  Gelman  and  Millipore mem-
                                                  brane filters.  Appl.  Microbiology.  28:605-
                                                  607,1974.
                                                   QUESTION AND ANSWER SESSION
 Geldreich,  E.E., H.L. Jeter,  and J.A. Winter.     Geldreich:
 Technical  considerations  in  applying  the
 membrane   filter  procedure.   Health  Lab.
 Science, 4:113-125, 1967.
 Hufham, J.B. Evaluating the membrane fecal
 coliform test by using Escherichia coli as the
 indicator organism. Appl.  Microbiology. 27:
 771-776, 1974.
 McFeters, G.A., and  D.G.  Stuart. Survival of
 coliform bacteria in natural waters: Field and     Stuart:
 laboratory   studies  with  membrane filter
 chambers.  Appl.  Microbiology. 24:805-811,
 1972.
 Presswood, W.G., and L.R. Brown. Compari-
son  of Gelman  and  Millipore membrane
Dave  I  think  the  reason  you see so
much difference between the Nucle-
pore and  the  other two  is that the
Nuclepore  is  not  really  the  same
material. As we tried to say this morn-
ing, if that membrane were to be used,
you  would have to redesign  a whole
family of media for it.

I  remember you saying that, and the
reason  that we used it was we  don't
use it for bacterial counts but when we
are using algae in our chambers, we use
a Nuclepore sidewall and it ended  up
being a pretty nice control.
                                               66

-------
                            EFFICIENCY OF COLIFORM RECOVERY
                             USING TWO BRANDS OF MEMBRANE
                                            FILTERS

                                       Frederick L. Harris
                                              and
                                          Carl A. Bailey
                              U.S. Environmental Protection Agency
                                 Surveillance & Analysis Division
                                        25 Funston Road
                                    Kansas City, Kansas 66115
                 ABSTRACT

     The comparative study of  Gelman and Milli-
pore membrane  filters by  Presswood and Brown
(5)  prompted the evaluation of  the two brands of
membrane filters employing routine samples. The
study included  a total of  100 samples from a
variety of non-chlorinated aquatic sources. Gelman
filters averaged 2.5 times greater recovery of fecal
coliforms than did Millipore filters. A comparative
study, with fewer samples, was also made utilizing
total coliform  analyses. Data indicate that total
coliform recovery is  similar with the two brands.
For  verification  as fecal coliforms, some  typical
blue colonies  were  subcultured  from both filter
brands to confirmatory media. Different  lots of
each brand of filters were used.

               INTRODUCTION

     Various brands of membrane filters have been
under close scrutiny by several investigators. A
disparity  in  the ability of  different  brands of
membrane filters to  support the growth of coli-
form bacteria from both natural and stock sources
was found. Using a typical strain of fecal coliform.
Levin et  al.  (4)  observed that Gelman  filters
exhibited a much less adverse effect on the micro-
organism  than  did   Millipore and Oxoid filters.
Presswood and Brown (5),  utilizing  pure  strains,
found  that  Gelman   filters recovered  2.3 times
more fecal coliforms  than  did  Millipore filters.
Comparative analyses of river water for fecal coli-
form bacteria gave results comparable to those for
pure cultures. In the study, total coliform recovery
was  statistically higher  with  Gelman filters than
with Millipore  filters. In a field study employing
Gelman,  Millipore,  and  Sartorius  membrane
filters,  Dutka  et  al. (2)  observed  that Gelman
filters generally produced the highest counts. Huf-
ham (3) found that Gelman  filters demonstrated
higher counts of a  strain of typical fecal coliform
at 44.5 C than did  Millipore filters; however, both
filter brands showed similar results at 35 C. Using
natural  samples with Gelman  and Millipore filters,
Shaeffer et al.  (6) obtained higher total coliform
counts  with  Gelman filters.  The fecal  coliform
counts  were  similar with  the  two  membrane
filter brands.

     Over  the past  few  years,  the  membrane
filter test has become an official method (1) and a
valuable  laboratory  tool.  However,   with  the
mounting data of various investigators indicating
membrane filter brand disparity in microorganism
enumeration, doubts have been  raised  concerning
the present accuracy of the test.

     As  an  in-house  quality  control  measure,
prompted by the results of Presswood  and Brown
(5),  a study was initiated to evaluate Gelman and
Millipore  membranes using routine samples from
non-chlorinated aquatic sources.

         MATERIALS AND METHODS

Sample Sources and Sampling.

     Water samples were obtained from three types
of sources:  (i) aerobic  lagoon  (influent and  ef-
                                              67

-------
fluent); (ii)  river; and (iii) sewage treatment plant
effluent. All samples  from  these  sources were
collected in autoclaved sterilized bottles iced en-
route to the laboratory, and  processed within 8
hours of collection time.

Procedure, Culture Media, and Reagents.

     All procedures, media, and reagents used were
in accordance with  those described in  Standard
Methods for the Examination of  Water and Waste-
water (13th ed.) part 400.

     Samples were taken from the  same dilution
bottle  and  filtered   simultaneously  through each
brand of filter using a Millipore membrane filtering
apparatus with a 3 place-Hydrosol manifold.

Membrane Filters.

     Three lots  each  of two commercial brands of
0.45 Aim porosity membrane filters were used in
the  study:  Millipore  HAWG 047SO  (Millipore
Corp.,  Bedford, Mass.) sterilized  with  ethylene
oxide  by  the manufacturer; and  Gelman  GN-6
(Gelman Instrument  Co.,  Ann Arbor,  Mich.)
sterilized in an  autoclave by the manufacturer.

Confirmation of Colonies.

     Fecal coliform  colonies  from membranes of
both brands were tested to establish the validity of
counts.  Ten blue colonies were picked at random
from  each  of 10 randomly  selected membranes.
The  confirmation  study  was carried out  in two
phases:  (I) subculture of 10 colonies  per mem-
brane to  EC broth  incubated at 44.5 C  for  24
hours and (II) subculture of 10 more colonies per
membrane  to tryptophane broth, MR-VP broth,
and citrate agar.
         RESULTS AND DISCUSSION

     Fecal  coliform colonies  grown on  Millipore
filters appeared larger, smoother, and more mucoid
than on Gelman filters. The fecal coliform colonies
on  Gelman filters, although  generally  higher in
number than Millipore filters, appeared  small and
often dull.  This  observation  was  also  noted  by
Presswood and Brown  (5). When utilizing  the M-FC
test, it  was found that Gelman filters  appeared
blue while  Millipore filters  appeared beige-yellow.
These have also been the findings of other  investi-
gators  (2,  5).  There  has   been  speculation that
this phenomenon  is due to  a difference in.pH and
that this could possibly be  responsible for the
disparity in counts on the two filters (5).

     Table  1  shows that  during 100 test trials,
colony counts on Gelman filters were almost con-
sistently higher than Millipore filters. On 3 trials,
Gelman filters  were  lower  or  equal  to  Millipore
filters in fecal coliform count. In considering the
overall data, Gelman  filters recovered 2.5 times
more fecal coliform   bacteria  when the  same
samples and  identical processing  methods  were
used.
   TABLE 1.   COMPARATIVE STUDY OF GELMAN AND MILLIPORE FILTERS FOR THE
              RECOVERY OF FECAL COLIFORMS.
Colonies Per
Millipore
23
10
10
14
9
17
33
18
12
18
12
7
18
19
26
Membrane
Gelman
58
44
20
44
28
36
59
42
20
42
2-
21
43
38
36
Ratio
G/M
2.52
4.40
2.00
3.14
3.11
2.12
1.79
2.33
1.67
2.33
1.67
3.00
2.39
2.00
1.38
Colonies Per
Millipore
29
23
6
4
7
31
13
32
26
11
10
10
7
44
11
Membrane
Gelman
49
59
35
20
32
36
28
44
36
36
36
30
46
43
22
Ratio
G/M
1.69
2.56
5.83
5.00
4.57
1.16
2.15
1.38
1.38
3.27
3.60
3.00
6.57
0.98
2.00
                                               68

-------
Table 1 cont'd.
Colonies Per Membrane
Millipore Gelman
9
19
11
14
49
27
42
16
23
19
7
19
12
23
29
30
35
39
32
18
26
15
16
8
13
11
9
41
15
16
20
25
23
25
34
Total count
Total count
Mean count
Mean count
No. of times


No. of times


2 of G/M
35
49
26
38
39
44
42
25
42
59
22
30
30
39
46
40
44
49
51
31
46
31
26
45
56
44
22
53
24
39
32
30
31
46
55
Millipore =
Gelman
Millipore
Gelman
Millipore recovered
Higher counts
Lower counts
Gelman recovered:
Higher counts
Lower counts
=
Mean ratio of G/M
Ratio
G/M
3.89
2.58
2.36
2.71
0.80
1.63
1.00
1.56
1.83
3.10
3.14
1.58
2.50
1.70
1.57
1.33
1.26
1.26
1.59
1.72
1.77
2.07
1.62
5.62
4.31
4.00
2.44
1.29
1.60
2.44
1.60
1.20
1.35
1.84
1.62
1896
3821
19
38
:
2
97

97
2
251.51
2.52
Colonies Per
Millipore
22
36
16
4
38
9
12
12
13
31
13
23
20
13
32
16
42
23
16
38
11
10
8
6
8
7
16
7
11
18
18
19
20
5
23












Membrane
Gelman
39
52
39
23
49
33
36
25
44
44
28
59
37
36
58
56
61
44
49
49
55
49
37
35
27
29
42
24
20
39
30
23
36
21
29












Ratio
G/M
1.77
1.44
2.44
5.75
1.29
3.67
3.00
2.08
3.38
1.42
2.15
2.56
1.85
2.77
1.81
3.50
1.45
1.91
3.06
1.29
5.00
4.90
4.62
5.83
3.38
4.14
2.62
3.43
1.82
2.17
1.67
1.21
1.80
4.20
1.26












   69

-------
    Table 2 indicates that total coliform recovery
is similar with the two brands. The contrasting data
of Tables  1  and  2 would  indicate a  possible
elevated temperature-membrane inhibitory effect.
This difference in recovery has been observed by
other  investigators  (2, 3, 5).  However, Presswood
and Brown  (5)  found  a statistically significant
difference at both 35 C and 44.5 C. The compara-
tive total coliform  study is not as comprehensive
as the fecal  coliform study due to: (1) initial tests
indicated the two membrane filter brands gave
similar counts (2)  the  majority of samples pro-
cessed  in our laboratory are for fecal coliform de-
termination.  It is hoped  that a  more in-depth
membrane comparative study with total coliform
procedures will be a part of our quality control
program in the near future.

     Due  to the higher fecal coliform counts on
Gelman membranes,  it was necessary to confirm
typical colonies in order  to eliminate the possibil-
ity  of a high number of false positives on Gelman
membranes. Table 3 indicates that the two mem-
brane filter  brands gave  similar confirmatory re-
sults in both phase I and II.
TABLE 2.  COMPARATIVE STUDY OF GELMAN
           AND MILLIPORE FILTERS FOR THE
           RECOVERY OF TOTAL COLIFORMS.
TABLE 3.  VERIFICATION OF TYPICAL BLUE
           COLONIES ON MILLIPORE AND
           GELMAN MEMBRANES
                                                               No. Colonies*   No. EC
                                                   Filter Brand    Picked        Positive
^uiunies rei ivi

Colonies
Millipore
50
25
25
25
26
44
33
21
29
19
19
42
17
32
19

Total count
Total count
Mean count
Mean count
emuidiie

Membrane
Gelman
54
24
24
27
28
38
24
17
35
24
26
30
24
25
29

Millipore
Gelman
Millipore
Gelman

Ratio
G/M
1.08
0.96
0.96
1.08
1.08
0.86
0.73
0.81
1.21
1.26
1.37
0.71
1.41
0.78
1.53

= 426
= 429
= 28
= 29
I Millipore 100 92

Gelman 100 90
No. Colonies* IMViC No. of
Filter Brand Picked Pattern Cultures

II Millipore 100 + + - - 97
1
- + - - 2

Gelman 100 	 + ~ 95
1
- - + + 4

* 10 typical blue colonies were picked at random
from 10 randomly selected membranes.


In order for a microbiologist to report accur-
ate data, he must have efficient tools. Hopefully,
there will be a standardization and a centralized
routine evaluation of membrane filter brands.
The varying efficiencies of membrane filter brands
cannot be allowed to continue.
No. of times Millipore recovered:


Higher counts
Lower counts
No. of times Gelman recovered:


2 of G/M
Mean Ratio

Higher counts
Lower counts

of G/M

7
= 8

= 8
= 7
= 15.83
1.06


SUMMARY
(1) A comparative study has been conducted with
two brands of membrane filters.

(2) The data indicated that Gelman filters re-
covered 2.5 times more fecal coliforms than
Millipore filters.
                                             70

-------
(3)   Recovery  of total  coliform  bacteria was
     similar  with Gelman  and  Millipore filters.

(4)   There is a need to improve the quality control
     of  membrane filters for use  in water micro-
     biology.
           ACKNOWLEDGEMENTS

     The authors wish to express their apprecia-
tion to H. Bunczewski of the Bio-Analysis Group,
U.S.  Environmental  Protection  Agency,  Region
VII Laboratory, for media-reagent preparation and
filtration of  various  samples. A  special  note of
thanks  to S. Finley of the same laboratory for
her clerical assistance.
                REFERENCES

1.   American Public Health Association. Standard
     Methods for the Examination  of Water and
     Wastewater, 13th ed. American Public Health
     Association Inc., New York, 1971.
2.   Dutka, B.J.,  M.J.  Jackson, and  J.B.  Bell.
     Comparison  of  Autoclave  and   Ethylene
     Oxide-Sterilized Membrane Filters  Used in
     Water Quality  Studies. Appl.  Microbiology.
     28:474-480, 1974.
3.   Hufham, J.B. Evaluating the Membrane Fecal
     Coliform Test by using Escherichia coli as the
     Indicator Organism.  Appl.  Microbiology. 27:
     771-776, 1974.
4.   Levin, G.W., V.L.  Strauss, and W.C. Hess.
     Rapid Coliform Organism Determination with
     C14. J.  Water Pollution  Control.  Fed.  33:
     1021-1037, 1961.
5.   Presswood, W.C., and  L.R. Brown.  Compari-
     son   of  Gelman and  Millipore  Membrane
     Filters for Enumerating Fecal Coliform  Bac-
     teria. Appl. Microbiology. 26:332-336, 1973.
6.   Schaeffer, D.J., M.C. Long, and K.G. Janardan.
     Statistical Analysis of  the  Recovery of  Coli-
     form Organisms  on  Gelman  and  Millipore
     Membrane Filters. Appl.  Microbiology.  28:
     605-607, 1974.
      QUESTION AND ANSWER SESSION

Sladek:     The only comment I have is with  re-
           spect to your statement on good lots
           and bad lots. I  don't believe we came
           to exactly that conclusion. One of the
           lots that Mr. Harris studied was about
           5 years old and we had manufactured
           it at a time before we had  begun to
           quality   control  those   filters  with
           respect to the fecal coliform test.

Harris:     Well how do you explain that the two
           good  lots  gave  lower recovery com-
           pared to the Gelman lots?

Sladek:     I think  without being able to inspect
           the  numbers  and  the  experimental
           conditions,  it  is not  really worth
           discussing.

Dazio:     I would like to ask you or anybody
           else in the audience, if you could tell
           me what basic difference there is be-
           tween   different  brands,  let's  say
           Gelman  and Millipore,  which  could
           account  for the differences for the
           recovery  of  fecal coliforms. Does
           anyone   know?   I  think  we  should
           being asking  some  basic  questions,
           and  analyse and perhaps chemically
           determine  what  differences  exist in
           the composition of membranes which
           could  account   for  differences  in
           recovery.
Harris:      Well,  there  is one possibility. The
            Gelman  membrane has phosphate in
            it whereas Millipore does  not. Milli-
            pore  filters  are  supposedly  almost
            completely  inert.  So  it could be  a
            possibility that the phosphate in Gel-
            man membranes  acts as  a  nutrient
            source for the fecal coliform organism
            and therefore  tends to  enhance  the
            growth of the fecal coliform organism
            on the membrane.

Question    Somebody,  this afternoon, said that
from the    the phosphate buffer  was very dela-
Floor:      terious  to   fecal   coliform  growth.

Grasso:     I want to ask the gentleman from the
            University of  Florida,  "Will you  be
            here tomorrow?"  I  think that  Dr.
            Litsky put me  in my place  and said to
            hold my information  until I present
            my paper. So, without letting  it  out
            of  the  bag  today,  I  think  I  could
            probably give  you some encouraging
            results on what is  the effect of  differ-
            ent brands of filters tomorrow  in  our
            paper. Thank you.
                                               71

-------
Harris:     We have some comments that ethy-
           lene-oxide  leaves a toxic residue  but
           according  to  the  presentation   by
           Millipore  Corporation  this morning
           they can't find it.                       Harris:

Brodsky:   Not  to put  down the  individual
           companies  producing membrane  fil-
           ters, but I am curious to know where
           you got your membrane filters? Were
           they, pardon me, a donation from the
           company  or were  they  purchased
           from stock?
                                                  Lane:
                                                  Harris:


                                                  Presswood:
Harris:     They  were  purchased from  stock.

Brodsky:   From a stock, both brands?

Harris:     Right, both brands.

Brodsky:   O.K.  Because  I  think  that  is  one
           aspect  of   the   investigation  that
           really should be settled.

Harris:     We did not go directly to Gelman to
           get the filters or directly to Millipore,
           they were from the clearing house.

Rusnell:    I just  wanted to know if  the samples
           were chlorinated or unchlorinated.

Harris:     We  took  this  into  consideration be-
           fore starting our study,  and  we de-
           cided  that it wouldn't be wise to use
           chlorinated sources.
Mack:       I was wondering if anybody had taken
            a look at the coliform organism under    Question
            the electron microscope, because we    from the
            are talking  about filter  sizes  and  all    Floor:
            the scientific work that goes with this
            and nobody has come up with the fact    Furman:
            that  many  of  these  organisms are
            terribly  large  and  some  are quite
            small. Some of them have capsule-like
            material  and  others don't, and this
            would have a  great deal to do with
            whether  or  not they  are  retained.
Lane:       You said that the Gelman membranes
            were autoclave-sterilized and Millipore
            were ethylene-oxide  sterilized.  How
            can you compare one membrane that
            is  sterilized  by  one  method with a
membrane  that  is  sterilized  by an
entirely different method? Shouldn't
they both be sterilized the same way?

According to the manufacturers they
are supposed  to give  equal results.
There are not supposed to be inhibi-
tory effects  from ethylene oxide or
autoclave  sterilization.  Millipore says
ethylene oxide is nontoxic so we used
their filters and  compared  to auto-
el aved filters.

I can't visualize that heating a mem-
brane to the temperature of 121° C
for 20  minutes or  a half an hour is
not going to do something to the pore
size of  the  membrane  whereas ethy-
lene-oxide might be nontoxic.

But the fact remains one gets higher
counts than the other.

Ours  was  a  comparison  of   filters
similar  to  Mr. Harris'  and we also
used autoclave and ethylene  oxide
membranes. The  reason  we did this
was that this is the way the manufac-
turer  sells them.  I  think  Millipore
does sell autoclaved filters but as Mr.
Harris   said  the  ethylene-oxide  is
suppose to  give  comparable  results.
Also some of the Millipore filters be-
come distorted  if you  do autoclave
them. You can see  by the grid lines,
that the filters are distorted.
                                                              What about Nuclepore?

                                                              The  Nuclepore  membranes do  not
                                                              grow colonies with media that  you
                                                              use, so  I would like to publicly tell
                                                              you not to use Nuclepore for that pur-
                                                              pose and with these media. We have
                                                              never  made the  claim that the poly-
                                                              carbonate   would  work   for  water
                                                              analyses. If you have some research
                                                              needs for polycarbonate and you can
                                                              use some wrinkles, add surfactants to
                                                              the media to improve growth. But for
                                                              the routine use we do not recommend
                                                              Nuclepore  for  microbiological  pur-
                                                              poses. Thank you.
                                               72

-------
                      COMPARISON OF MEMBRANE FILTER BRANDS FOR
                             RECOVERY OF THE COLIFORM GROUP

                                    A. P. Dufourand V.J. Cabelli

                                  Environmental Protection Agency
                              National Marine Water Quality Laboratory
                                         Narragansett, R.I.
                  ABSTRACT

     Recoveries  of pure cultures  of E. coli, K.
pneumoniae and E. cloacae and fecal  coliforms
from natural waters were compared using the M-FC
test and different lots  and  brands  of membrane
filters.   MPN  recoveries  were used  as  reference
values for measuring accuracy.

     Except for  Nuclepore,  the  brand of mem-
brane filters did not significantly affect  recoveries
of pure cultures or natural source fecal coliforms.
Variability  increased with natural  water samples.
The  variability  (precision) of results from lot to
lot  within  a  brand and from culture to culture
within  an MF  brand precluded any generalization
about  acceptance of  one  brand  over  another.

                INTRODUCTION

     Our interest  in the recovery efficiencies of
membrane  filters originated  a few years ago when
we began an epidemiological/microbiological study
of the relationship between pollution  levels and the
incidence of disease at marine bathing beaches. It
resulted  from  the necessity  to evaluate standard
membrane  filter  (MF)  methods  in marine waters
prior to their possible use in  the study. The evalua-
tion  of the M-FC test  revealed  that it measured
much lower fecal  coliform densities than the EC
most probable number (MPN) test. This in turn led
to a  further investigation  of various membrane
filter  brands  in  order  to determine if  the filters
themselves  affected the accuracy and precision of
the assay method. For the  purpose  of this  pre-
sentation, I want to define those two terms as we
use them.

     The accuracy of a method for an organism or
a group of  organisms, which may be defined as the
ability to detect the "true density" of bacteria in a
given volume of a water sample, ideally should be
determined using natural samples. However, since
"true density"  is, in  fact, unknown  in  natural
samples,  the  accuracy  can be  determined only
relative  to   some   "standard"   or   "reference"
method. Secondly, when a group of organisms is
being enumerated i.e., coliforms or fecal coliforms,
the  relative  accuracy may vary  from location to
location depending  on the particular distribution
of  the component  biotypes  in  the particular
sample.

     The precision of a method,  which  may be
defined as the degree of dispersion of obtained
values around a mean  estimate, is equally as
important as  its accuracy or recovery efficiency.
This is so because of the practice of determining
bacterial  density values with a single filter which
is  a common  practice.  Thus,  a  lack of precision
caused by the filter  would affect the ability of a
method to  detect differences  between bacterial
densities of two water samples.

     With these  two  parameters in mind, the
following experimental methodology was used to
determine which type  of filter  would be most
efficient  for  our purposes. I would  like to state
that all of these were done with surface waters in
mind,  particularly bathing beach surface  waters,
and  I would not like to expand this to other types
of water samples.
         MATERIALS AND METHODS

     Media  Dehydrated  M-FC   broth,  M-Endo
broth,  nutrient  agar, Trypticase soy  broth,  EC
broth,  lactose broth  and  brilliant  green lactose
bile broth were obtained from Difco and  prepared
                                               73

-------
for use according to  the directions of the manu-
facturer.
M-Endo  broth,  M-FC  broth  and  nutrient  agar
media.
     Test Organisms  Strains of E. coli, Klebsiella
pneumoniae and  Enterobacter cloacae were main-
tained on nutrient  agar  slants.  The Klebsiella
strains were able to produce gas in  EC broth  in-
cubated at 44.5  C.

     Filters  Membrane filters from  the following
manufacturers  were  obtained  from  commercial
sources: Millipore (M), Gelman (G),  Sartorius (S),
Schleicher and Schuell (s/s) and Nuclepore (N).
All of the membranes were packaged  sterile with
the exception  of  the Nuclepore and Schleicher and
Schuell  membranes,  which were   sterilized  by
autoclaving for  15 minutes at 121   C. All mem-
branes, with  the  exception of Nuclepore,  were
gridded and  had an average pore  size of  0.45
microns.  Nuclepore  membranes were ungridded
and had a 0.4 micron pore size.
     Control Procedure Spread  plates used  to
determine the density of the test cell suspensions,
were prepared by pipetting 0.2 ml of the test cell
suspension onto each  of five plates. The suspen-
sions were spread over the surface of the nutrient
agar medium with a sterile glass rod and allowed to
dry, after which the plates were incubated in an
inverted position at 35  C.

     The MPN procedures were carried out as des-
cribed  in  Standard  Methods  for the Examination
of Water and Wastewater (1).

     Statistical   Analysis   Tukey's   Studentized
Range   Procedure  for  comparing several  means
was used for the statistical analysis of the data (2).
The dispersion  about the mean was described as
the coefficient of variation.
     Natural Samples Natural samples were obtain-
ed from three  locations in the  Rhode  Island area.
They were: (1) Wickford Harbor, a salt water cove
which receives mainly septic tank  overflows, and
this  would be  human waste of which the major
component was E. coli. (2) the Saugatucket River,
which receives mainly  industrial effluent  from a
textile  finishing  plant  and  these  effluents con-
tained  Klebsiella species. (3) the  Pawcatuck Es-
tuary   whose  bacterial  pollution  comes  from
domestic  sources, as well  as industrial effluents.
The   samples  usually  arrived  at  the laboratory
within  three hours  of  collection  and they were
immediately assayed.

     Test Cell Suspensions  Eighteen to  twenty
hours before each experiment, a loopful of the test
strain was transferred  to  Trypticase  soy  broth.
After  18  hours incubation  at  35   C,  the  culture
was  diluted  in sterile, phosphate-buffered saline
(pH  7.2) to a density of between 20 and 60 organ-
isms per ml.

     Membrane Filtration  One ml  of the test cell
suspension, or  an appropriate volume of a  natural
sample, was passed through each filter after being
mixed with 20 ml of buffer. The  filter was then
rolled onto  a  broth saturated pad  or  the  agar
surface of a  randomly chosen plate.  The M-FC
plates were  incubated  in  "whirl-pak" bags in  a
44.5  C water bath.  The M-Endo plates and mem-
brane nutrient agar  plates were  incubated in an
inverted position in  a 35  C incubator. All mem-
brane filter  brands  were tested in triplicate on
         RESULTS AND DISCUSSION

     The accuracy of the M-FC procedure for fecal
coliforms as it is affected by  the  brand  of mem-
brane filter used, was determined by comparing the
densities in Klebsiella and E. coli test suspensions
obtained by the  MF  procedure with those from
nutrient agar spread plates. It was assumed that the
density, as determined from nutrient agar spread
plates, would  provide the  best  estimate  of  the
number of  bacteria actually  present in a given test
cell  suspension.  The  effect  of five  membrane
filter brands on the accuracy of the M-FC proce-
dure with four fecal coliform strains is shown in
Table  1. The percent recoveries for the  E. coli
strains ranged  from 16 to 75%,  and for the Kleb-
siella strains  they  ranged  from  14 to 94%.  Al-
though  the filter brands are ranked according to
their mean recovery values, most of the observed
differences were  not  statistically  significant  as
shown by  the  underscoring  of  mean  recovery
values of each organism. The underscoring indi-
cates that any two means, not underlined by the
same line,  are significantly  different at  the 95%
confidence level.  Statistically significant  differ-
ences appeared  more  frequently  with the Kleb-
siella strains than with the  E. coli  strains. Of the
brands tested, the Gelman product appeared  to
provide the most, and Nuclepore the least accurate
density  estimates  with  the four  fecal  coliform
strains tested.  However, none of the filter brands
was consistently accurate. This appeared to depend
more on the  basic method relative to the strain
being tested rather than the filter brand.
                                                74

-------
 TABLE 1.   RECOVERY  OF  E.  COLI  AND  K. PIMEUMONIAE ON THE  M-FC MEDIUM WITH
             VARIOUS MEMBRANE FILTER BRANDS
 E. coli      no. 3

             Filter type
             Mean (2)
             % Recovery
             ST. (3)

 E. coli      no. 5

             Filter type
             Mean
             % Recovery
 S.T.        ST.

 K. pneumoniae no. 450

             Filter type
             Mean
             % Recovery
             S.T.

 K. pneumoniae no. 444

             Filter type
             Mean
             % Recovery
             ST.
(1)
              N
              14
              16
              N
              22
              18
M
34
39
s/s
38
43
S
41
47
G
46
52
N
50
38

s/s
57
44


S
71
55


M
77
59

G
98
75
M
21
14
S
29
19
s/s
31
20
s/s
66
53
S
86
68
                                                  N
                                                  47
                                                  30
M
99
80
                                     G
                                     75
                                     48
 G
118
 94
 1.          Gelman (G), Millipore (M), Nuclepore (N), Sartorius (S) and Schleicher and Schuell (s/s)

 2.   Relative to bacterial density on nutrient agar spread plate incubated at 35C.

 3.   S.T. = Statistical Test: Tukey's Studentized Range Procedure; all  means not underscored by the
      same line differ significantly at the P = 0.05 level.
     The effect of membrane filter brand on the
precision of the M-FC test is shown in Table 2. The
filters are ranked  according to the magnitude  of
dispersion around  the mean.  The precision of the
Sartorius filters  was consistently better  (a  lower
coefficient of variation) than the other brands,
and  the  Nuclepore and  S  &  S  filters were con-
sistently  poor. Unlike the accuracy, the precision
appears to be filter  brand  dependent rather than
strain dependent.

     The recovery of fecal  coliforms from natural
samples is shown in Table 3. The trend, with regard
to filter  efficiency, appears to be similar to that
found  with  pure  culture  suspensions.  However,
                           there were fewer statistically significant differences
                           between filter brands with  natural sample suspen-
                           sions.  These differences were  probably due to the
                           greater filter variability of all  brands (see Table 4).
                           This variability would, in  fact, cause the statistical
                           test used to be less sensitive for detecting recovery
                           differences between filter  brands.

                               Table 4  indicates that the  precision of the
                           M-FC  test was much  lower with all brands of filter
                           when  compared with the  pure culture  data. It can
                           also be noted that the rank position  of the filter
                           brands was not similar  to those obtained with
                           pure cultures. This  is probably  indicative of the
                           heterogeneous nature of the fecal coliform group.
                                               75

-------
   TABLE 2.  PRECISION OF MEMBRANE FILTER BRANDS ON M-FC MEDIUM
E.


E.


K.


K.


coli no. 3
Filter type (D
C.V. (2)
coli no. 5
Filter type
C.V.
pneumoniae no. 450
Filter type
C.V.
pneumoniae no. 444
Filter type
C.V.

N
43

s/s
18

N
19

N
45

s/s
24

N
12

M
14

s/s
18

G
17

G
12

s/s
10

M
14

M
12

M
6

G
7

G
3

S
5

S
3

S
4

S
2
   1.   See footnote 1, table 1.

   2.   C.V. = Coefficient of Variation; (standard deviation/arithmetic mean) x 100
TABLES.  FECAL  COLIFORM RECOVERIES FROM NATURAL SAMPLES WITH VARIOUS
          MEMBRANE FILTER BRANDS USING THE M-FC TEST.
Wickford
Harbor



Saugatucket




Pawcatuck
Estuary





Filter type (D
Mean (2)
% Recovery
ST. (3)
R.
Filter type
Mean
% Recovery
9 T
O. 1 ,

Filter type
Mean
% Recovery


N
8.3
12


N
4
33



N
1
7



s/s
12
17


s/s
5
38



s/s
4.7
32



S
17.7
25


M
6
46



M
5.2
36



M
20.3
29


G
8
62



S
12
82



G
21.3
30


S
12
92



G
16
109

1.   See footnotes 1 and 3, table 1.

2.   Relative to E.G. MPN estimate

                                      76

-------
    TABLE 4.   PRECISION OF MEMBRANE FILTER BRANDS ON M-FC MEDIUM
Wickford
Harbor
Filter type M)
C.V. (2)
Saugatucket R.
Filter type
C.V.


s/s
35.8

N
78


G
35.7

M
45


S
25

G
44


N
18.7

s/s
20


M
7.5

S
17
Pawcatuck
Estuary
Filter type
C.V.


N
100


M
40


S
25


s/s
22


G
19
    1.    See footnote 1, table 1.

    2.    See footnote 2, table 2.
     The effect the filter brand has  on the re-
covery of total  coliforms by  the M-Endo proce-
dure was also examined. Table 5 shows that, with
pure cultures of  E.  coli,  Klebsiella, and  Entero-
bacter cloacae,  recovery  between  brands  was
neither  appreciable  nor  statistically  significant.
The one exception was the Nuclepore brand, which
was  consistently  poor in its  ability to  recovery
coliforms.  Relative  to spread plates on  nutrient
agar  (NA),  all of the filter brands, with the excep-
tion  of Nuclepore, were reasonably accurate when
using Klebsiella or Enterobacter as the test strain.
The  recoveries with  E. coli no. 3 were rather poor.
In order to determine if this might be due to inhib-
itors  present in  M-Endo  medium,  the mean re-
coveries on M-Endo medium  were compared to
those of  membranes  also  placed  on  nutrient
agar. As can be seen  in Table 6, the relatively poor
recoveries of E. coli no. 3 on  M-Endo were also
obtained  on NA, suggesting  a  general filter effect
or a nutrient deficiency for this particular organism.

     The  relative accuracy  of the various filter
brands, when examined using natural samples was
essentially  the same  as noted above. The recovery
relative to the MPN procedure was excellent in two
of the three samples examined. However, the fact
that  the  MPN procedure was used to estimate the
"true" coliform  density does not allow substantive
conclusions to  be drawn about these data in this
regard. See Table 7.
     The  variability,  from  lot  to lot, of  some
filter brands would preclude broad generalizations
about the acceptability of one brand over  another
except for those  brands actually tested under the
same conditions.  This point is illustrated in Table
8. In two out of the three brands tested,  there
were statistically different mean recovery values
between lots. Thus, it would appear that with some
membrane filter  brands, the recovery  efficiencies
can not be predicted from lot to lot. And you will
notice that this is a different E. coli strain than we
used in the other three experiments. What happen-
ed is that we had lost this strain, but last week we
found a  lyophilized culture of  it.  So, we ran it
through the same test procedure and,  if you will
put  on  the last slide, you will  see the strain to
strain variabilities causing differences with the
Gelman filters.
                  SUMMARY

    The following conclusions can be made:

    1.   The  use of  pure  cultures  to detect dif-
         ferences in  accuracy  in the M-FC  pro-
         cedure between membrane filter brands
         is  more  sensitive  than  using  natural
         samples. However,  there  is a  strain to
         strain variation in the  way in which pure
         cultures  react to  each  filter brand and
                                               77

-------
TABLE 5.  RECOVERY OF E. COLI, K. PNEUMONIAE AND E. CLOACAEON M-ENDO MEDIUM
          WITH VARIOUS MEMBRANE FILTER BRANDS
E. coli     no. 3

          Filter type (D
          Mean (D
          % Recovery
          ST. (D

K. pneumoniae no. 444

          Filter type
          Mean
          % Recovery
          ST.

K. pneumoniae no. 450

          Filter type
          Mean
          % Recovery
          ST.

E. cloacae  no. 491
                    N
                   32
                   36
                    N
                   68
                   51
G
58
66
s/s
59
67
M
61
69
S
68
77
  G
107
 86
  M
109
 88
 s/s
112
 90
  S
120
105
N
8
44

s/s
14
78

S
15
83

G
15
83
M
20
111
Filter type
Mean
% Recovery
N s/s S
11 23 24
58 121 126

G
243
128
M
26
137
1.   See footnotes 1, 2 and 3, table 1.
 TABLE 6.  COMPARISON OF E.COLI RECOVERIES ON M-ENDO AND NUTRIENT AGAR WITH
          VARIOUS MEMBRANE FILTER BRANDS
 E. coli
                                          % Mean Recovery Value
                                                             (1)
no. 3
Filter type (1'
mEndo
Nutrient Agar
N
36
52
G
66
70
s/s
67
68
M
69
69
S
77
75
 1.   See footnotes 1 and 2, table 1.
                                       78

-------
TABLE?.  TOTAL COLIFORM  RECOVERIES FROM NATURAL SAMPLES WITH VARIOUS
           MEMBRANE FILTER BRANDS ON M-ENDO MEDIUM
Wickford Harbor
            Filter type
            Mean
            % Recovery
            ST. (D
Saugatucket R.
            Filter type
            Mean
            % Recovery
            S.T.
Pawcatuck Estuary
 N
 g
26
s/s
33
94
G
37
106
S
38
109
M
39
111
N
27
15

G
56
32

S
67
38
s/s
71
41
M
83
47
Filter type
Mean
% Recovery
N
8
22
s/s
31
86
G
33
92
S
36
100
M
37
103
1.    See footnotes 1 and 3, table 1.

2.    Relative to completed total coliform MPN estimate.
       this may influence the choice of a "best"
       membrane filter. Furthermore, pure cul-
       ture  may  not truly  mimic the physio-
       logical  state of  organisms  in  natural
       samples.

   2.   The  precision of the M-FC test  is in-
       fluenced by  the filter brand, and this is a
       rather  consistent  characteristic  from
       strain to strain with pure cultures.

   3.   Differences  in precision due to  mem-
       brane filter brand have a tendency to be
       lost when natural samples are used as the
       test  inoculum in the M-FC test. This  is
       assumed to  be caused  by the hetero-
       geneity  of the fecal coliform group and
       the physiological state of the organisms.

   4.   The   relative accuracy  of  the  various
       brands  of membrane filters was essen-
       tially the same whether  pure cultures or
       natural  samples were used. However, in
       the latter case the differences were mask-
       ed due  to the decreased precision when
       natural samples were used.
                  5.    In  general,  acceptable  total coliform
                       recoveries were obtained  by the M-Endo
                       procedure  with  all  the filter  brands
                       (except Nuclepore).

                  6.    If pure  cultures are  used to compare
                       various filter brands or  lots within  a
                       filter brand, the choice of test organism
                       is exceedingly  important, since,  with
                       some  strains,   an  overall   membrane
                       filter effect may mask true differences.
                       If natural  samples are  used, a  large
                       enough  number  of  filters   must  be
                       examined  to  compensate  for the de-
                       creased  precision  observed  with  this
                       type of test suspension.

                  7.    Three factors that  must be  considered
                       in  evaluating   membrane  filters  are
                       accuracy,  precision,  and  lot to  lot
                       variability.

                             REFERENCES

                 Standard  Methods  for  the Examination of
                 Water and Wastewater, 13th Edition, Ameri-
                 can Public Health Assoc., New York, 1971.
                                            79

-------
 TABLES.  RECOVERY  OF  E.  COLI  ON  M-FC MEDIUM WITH DIFFERENT LOTS OF  MEM-
            BRANE FILTER BRANDS

 Millipore
 (E. coli no. 104)
            Filter lot #
            Mean
            % Recovery (D
            S.T.  (D
 Millipore
 (E. coli no. 104)
            Filter lot #
            Mean
            % Recovery
            S.T.
 Gelman
 (E. coli no. 5)
            Filter lot*
            Mean
            % Recovery
            S.T.
 Sartorius
 (E. coli no.  104)
1295
12.3
10.4

NN
25.7
21.7

9396
26.3
22.3

6112
38.3
33

8752
47.7
40
3283
  6.3
  6.3
80714
70
73
                                                      9733
                                                       89.7
                                                       76
5000
 16.7
 16.7
9768
 17.7
 17.7
                                 9733
                                   62
                                   62
          80706
          75
          78
         80730
         76
         79
Filter lot #
Mean
% Recovery
S.T.
10
65
34

30
27
14

 1.   See footnotes 2 and 3, table 1.
2.    Lecture Notes in Applied Statistics, Joseph L.
     Ciminera, Villanova Univ. Press, 1956.

      QUESTION AND ANSWER SESSION

Geldreich.  What do you think of the possibility
           of using pure cultures in some of the
           evaluations  of  these products? We
           can't say, as you are saying, to use a
           natural sample,  although I  always like
           to  use them in  preference to pure
           cultures. What  about the possibility
           of using a mixture of pure cultures in
           a lyophilized condition in which you
           can  standardize and then use them  as
           a way to come up with a test group
           which  would  allow  us to look for
           sensitivity, as  well as selectivity  in
                          suppressing  some  organisms.  You
                          think this is a  possible way out of the
                          problem?

               Dufour.    I don't really know. Dr. Hufham out-
                          lined this rather well in his presenta-
                          tion  this  morning. Any one culture
                          has a certain part of the distribution,
                          even  if we called it pure culture, that
                          is killed by the elevated temperature. I
                          don't  know whether putting in dif-
                          ferent types of organisms and  using
                          this as a test suspension will overcome
                          that shortcoming.

               Geldreich.  One  of the problems  that  we have
                          here  is that if you use an E. coli of a
                          certain type, you are only checking it
                                              80

-------
           for one  thing,  and that  is sensitivity
           to let's say E. coli. But, if  you are
           trying  to check it for suppression of
           other organisms, you would  be  miss-
           ing that point.

Dufour.    Yes. This is why  we used  samples
           from  what we thought were three
           different   environmental   situations.
           But  obviously, we  couldn't show
           anything  because  of  the great  dis-
           persion  we  found  between filters.

Geldreich.  We noted, and  you may have too, for
           instance  that Gelman membranes may
           at times  have more background count
           of other organisms growing on them.
           This  may occur  because of some
           nutrient  material present there.

Dufour.    Again, this was not one  of the prob-
           lems that we-ran into because of the
           way we  chose our natural  sample.

Presswood. I noticed that  if you  are using  pure
           cultures, the source where you isolate
           the  culture  makes a  difference.  I
           have isolated some bacteria from raw
           sewage before  it was chlorinated and
           those bacteria did  not respond as well
           to the membrane  filter  technique as
           bacteria  isolated  from  river water.
           Even after passing them through 2 or
           3  passes  of EC  medium,  or some
           other nutrient broth, they still did not
           respond  as well  to  the membrane
           filter.  Especially  at 44.5C as E. coli,
           or what we call fecal coliform bacteria
           isolated  from  river  water.  I  don't
           know  why this  is, but it  happens.

Dufour.    It was mentioned today, that  all of
           our strains were not isolated using the
           M-FC  procedure. They were isolated
           using a  non-inhibitory medium  and
           found to  be fecal coliforms using EC
           broth. So, at least we didn't have that
           shortcoming.
Bordner.    It just occurs to me that one of the
            plans we  have  in mind at MDQARL,
            Methods  Development and Quality
            Assurance in Cincinnati, is  to accept
            bids for a contract that will develop
            pure   cultures  from  environmental
            samples and  fecal samples. These can
            then be used  as reference samples.  It
            occurs to me that it might  be appro-
            priate  to  look  at this type of culture
            which  we hope  to  have lyophilized
            and made available to our laboratories
            and others  later  on  for membrane
            filter and  media evaluation.
                                              81

-------
  A COMPARISON OF MEMBRANE FILTERS, CULTURE MEDIA, INCUBATION TEMPERATURES,
    POLLUTED WATER AND ESCHERICHIA COLI STRAINS IN THE FECAL COLIFORM TEST

                                         Paul J. Glantz
                                Department of Veterinary Science
                                The Pennsylvania State University
                                   University Park, Pa. 16802
                   ABSTRACT

     Twenty  lots of  membrane filters  from  3
manufacturers produced variable results when the
fecal coliform test was used with  polluted water
samples and Escherichia coli (E. coli). The growth
of nineteen E. coli cultures varied with incubation
temperatures  of  35, 43, and  44.5 C, with mem-
brane filter lots, and with culture  media (Trypti-
case soy agar,  M-FC,  and  violet  red bile agar).
Growth was  better  on violet red  bile agar than
on  M-FC medium, which was not always due to
the 44.5C  incubation  temperature. Some of the
E. coli strains tested did not grow as well in poured
agar plates as they did on membrane filters. Evalua-
tion of the results indicated that the efficacy of
the membrane  filter for the  fecal coliform  test
could be affected by, (1) the strain of E. coli, (2)
the method utilized for growth (pour plate, pad
and  broth, pad and agar), and  (3)  the tempera-
ture inside the plate.  Therefore,  it  is apparent
that test methods must be standardized for these
factors.
               INTRODUCTION

     The membrane filter method used for mea-
suring  fecal coliform  bacteria in  polluted water
has been critically evaluated by several investiga-
tors. Poor  recovery of fecal coliform  or  E.  coli
strains has  been attributed to the variation of the
quality of the components in the M-FC agar (10),
to the differences in  the  brands and  lots of the
membrane  filters  (1,5,8,11),  to the variation in
the incubation temperatures (3), and to the type
of water tested (1,7,9).


     This study was undertaken to determine the
efficiency  of  three brands of membrane filters to
recover  fecal  coliforms and E. coli isolated from
polluted  water.  Three  incubation temperatures
(35,  43,  and  44.5 C) were studied to determine
the effect of temperature on the growth of the
bacteria.  A  non-inhibitory  medium,  Trypticase
soy  agar  (TSA, BBL),  was  used to determine
the actual counts  of the E. coli cultures being
compared for growth  on M-FC  and violet red
bile  agar  (VRB, BBL).  Additional studies  were
initiated  to determine  the cause of the erratic
behavior  of the membrane filters and the E. coli
bacteria.
         MATERIALS AND METHODS

    The manufacturer's  name  and  lot number
of the membrane filters (with the prefix used as a
test number in this study) were as follows:
Test  Millipore
No.   Lot No.
Test  Gelman   Test  Sartorius
No.   Lot No.   No.   Lot No.
M-1
M-3
M-5
M-6
M-7
M-8
M-11
M-12
M-13
95434-3
93448-10
11013-4
95434-2
12732-6
12667-3
06602-8
06618-3
34798-4
G-3
G-6
G-7
80730
80822
80901
S-1
S-2
S-3
S-4
S-5
S-6
S-7
S-8
713649
753 649
773 284
993 447
993711
993 186
013870
773 568
    All of the above except M-11  and M-12 were
pre-sterilized  by manufacturer. These two  lots
were used as supplied (not sterile).

    M-FC  broth  BBL lot  204625)  with 0.01%
rosolic acid (Difco lot 488535) added was used for
                                              82

-------
tests  with pads, and  1.5% agar added  for pour
plates. These two media were dissolved by heating
to a  boil (as was the  VRB). The TSA was auto-
claved at 121 C for 15 minutes. The agar media
/vere cooled to 45 C before pouring plates.


     Sterile demineralized  water (pH 6.5) was used
to prepare dilutions of the E.  coli culture and as
a dispersal solution (20 to 30 ml) during filtration
of all samples.  Thirty  to 40 ml of this water was
filtered after every 5 replicates as controls. A 1 ,000
Ai1 Eppendorf,  or a 5 to  10 ml  Oxford pipet with
sterile tips, was used  to  transfer 1  to  10 ml  of
diluted E. coli  culture, or  creek  water, for filtra-
tion  and  agar pour plates. Membranes were placed
on pads in sterile 48 x 8.5 mm Millipore or 12 x 50
mm  Falcon plates. Sterile  15  x 100 mm plastic
petri  dishes were used for pour plates. The upper
half  of the  Gelman magnetic filter funnel was
placed under  a germicidal lamp (G.E.  G8T5)
between  filtrations. Plates were  sealed inside Whirl-
pak (Nasco) or Ziploc  (Dow Chemical Co.) plastic
bags prior to immersion in  the water baths. After
24  hours  incubation,  typical   blue M-FC  or  red
(VRB) colonies were counted at 7X magnification
with  daylight fluorescent illumination. The 15  x
100  mm  agar pour plates were  counted using a
Quebec colony  counter.  Plates  with  20 to 100
colonies for the 48 x 8.5 mm plates and 30 to 300
for the  100 x  15  mm  plates,  were considered
countable.

     The  nineteen  E.  coli cultures isolated from
creek water were incubated at 44.5 C  on  M-FC
medium.  Biochemical  reactions  of  the E. coli
cultures  tested  produced acid-slope  and acid-gas
butt  on  triple   sugar iron  agar (TSI,  BBL). The
Minitek (Bioquest) test method gave the  following
results:  positive  for indole, methyl  red, lactose,
arabinose,  lysine  decarboxylase;  negative  for
Voges-Proskauer,  citrate,  phenlalanine, inositol,
malonate,
     Variations occured with ornithine decarboxy-
lase  and  rhamnose  respectively  as  follows: +, +,
Nos. 2A, 3A, 9, 16, 16A,  16X, 16Z, 24;  -, +, Nos.
1A,  5; -, -,  Nos.  1,  2, 2Z,  3,  3X, 3Z; 4,  +, -,
Nos. 2B, 2X. Culture  #4A gave results typical of
Klebsiella pneumoniae.
     Stock cultures were maintained at room tem-
perature  on TSA slants sealed with waxed corks.
Growth cultures were transferred from stock plants
to VRB  agar plates, incubated  at 35 C overnight,
and a single colony suspended  in sterile  deminer-
alized water (pH 6.5). One ml of the suspension
was  transferred  to  membrane filters, or to pour
plates for subsequent tests.
     The temperatures of the  Hotpack Incubator
#5528 (35  C), the GCA-Precision  Scientific Co.
Coliform Incubator Bath (44.5 C), and the Blue-M
non-circulating waterbath  (43 C) were continually
monitored  during  the study. The waterbath tem-
peratures  and  the temperatures  of the corres-
ponding immersed  plates did not vary.

     The different lots of Gelman, Millipore and
Sartorius membrane filters were tested with  11
E.  coli cultures.   The  membrane  filters were
placed either on corresponding pads saturated with
M-FC medium or  on TSA plates and incubated at
44.5  C  (GCA-Coliform  Bath).  The  same test
cultures were also plated with TSA  agar and  in-
cubated at 35 C. One ml  of a suitable dilution of
the culture  was used for filtration and for poured
plate counts. All  tests were  done  in  five repli-
cates.

     The results based on the average counts, are
expressed  as percent (%)  recovery of E. coli. The
pour plate counts  obtained with TSA at 35 C were
considered  representative  of the actual number of
E. coli in the 1 ml  volume tested. Membrane filters
were placed on solid TSA agar for comparison with
membrane filters placed on pads. A dilution of one
strain  of  E. coli  was tested with  all  membrane
filters listed  for that strain at one time.
                   RESULTS

Comparison of  E. coli Recovery by M-FC  Brand
and Lot

     In Table  1  the results are presented in 3 cate-
gories, A, B, and C. Category A is the percent re-
covery of the E. coli counts on membrane filters
                                                 83

-------
















d
in
5
Q
Z
<
O
in
CO
<
<
CO
1—
T.
H
3»
>
Q
LU
DC
<
a.
S
O
o
0
LL
s
z
o
Ij
o
U
ai
1 1
0
>
DC
LU
>
O
o
LU
DC

h-
2

O
DC
LU
a.

LU
i
.j
CO
<
I—
1
in
*t U3
Tf CO
°< <
CO CO
1- h-




LO
5 u>
m co
LL <
£ £



" 3
»rMCM«— ^I^CM^— CMr^^^^fr^T— incocM^'Oi^i'OJin^'
cocoincocoo5'-ooooo5»-'-ooo30oooo)o>ooinoor»oop^
«— «— «— T— «— r— T—






O)OOOO)^-'-r>oocNinr>.oO'-CMcocMcDin«*oo^co^-o
inr^-coco^-rv-O)^'CDtN-inooooooocDcooocDtocDr^-cor^r>-i>>




incM.cor«srvr^r^r^cor>a>ooo5





«<<colnCOrxoocoCDtt>oo'7{Or^c?'-
S(3coC5coIEC3cn2ci)coSC3coC3coSC3coC3coSc3coCDco



co'-o^f'— cococococooocM^tinr^^i-O'-LnincM^t
ococMOTt— T-or"»O)cooor^coinrxcor-.oocoo>O)CO
T— t— T— r— r—




o 05 »— cMOOOcoOor^ococoincNr^rMinoocMCOco
ooooooom CM «- inr^incDinincDinincoin
*™~





co^ri^«r>.CMOCooocMO'*Tj-O5»-oo'tr~oincM^r
oooouscoin CM CM r*r-o>cocor»o>O)inr-.cQ







««« «<«


*
0) co CO CO
topooooo 5'7coT_coco^-T7Cor-L9toroCOioT7{?'7
i^r /« » ,. L ~y "^ /n > «>. ^^ /H **» ^5 ^n * *\ ^S ^^ */> ^^ f }\ ^> (^ r f\
^tDcOCOZSUCO^UCO^VJCOSC-'COwCO^tJCO
































^
CD
OJ
4-»
c
i—
5
E
3
c
•1-1
^
52
*
i^
CD
CO
JO
E

o
X)

L
CO
05
CO
>
o
C/)
CD
(A
CO
CJ
'•(-I
Q.
>
H
C
o
~n
w
0>
CJ
CO
Q.
L_
d)
^r
^-
0
"8
L_
CO
a
E
o
CJ
E
3
T3
CU
E
o
LL
^
^
-C
i
•a
CO
Q.
C
0
x_
O)
r^:
H—
a>
c
£
n
E
CD
E
c
o
1
0}
g
y
0)
^
(1)
UJ
L.
3
•(-»
3
CJ
"5
U
LU
"o d
c ^q
OJ ^1-
s 5
CD 4.!
a. co

<


<
CO
1—
c
o
1 J
c
"2
o
o
5!
CO
a.
CO
O5
CO
•D
£
3
O
a.
(->
2
"D
£
(D
a
E
o
CJ
O

in
5
•U
CO
E
_3
T3
CD
E
o
LL
^
2:
jz
I
"D
CO
a
c
0
x_
CD
—
^^
CD
C
CO
L~
jO
E
CD
E
c
0
•D
2?
CD
O
0
CD
l_
n\
£
3J
+-<
D
O
1
LU
'o
C "
05 in
fc? co
05 +-•
a. co

00









d
in
CO
*j
CO
CO
O
4-1
c
23
O
U
CD
+J
JO
a
L_
CO
05
CO
•a
£
^
o
a
o
+j
~a
CD
1 _
CO
a.
0
CJ
O
in
*t'
^~
+-<
CO
<
&
c
o
53

*4—
0)
CD
L-
-Q
E
CD
E
c
0
T5
2?
CD
O
0
CD
^.

£
D
•M
3
CJ
1
LU
*+••
O
<-<
05
£f
05
a.

O
84

-------
(M-FC pad)  compared with the average counts of
filters  on  TSA  at 44.5 C. The percent recoveries
ranged from  44 to 133%, with cultures 2A, 3A,
and 24 providing good  recoveries while  cultures
1A and 3 were poor.

    When the  1A  culture was filtered  and the
membrane filters were placed on pads with M-FC
medium at 44.5 C, growth  occurred on Gelman
lot numbers  3 and 6, but not  on Millipore lot
numbers 3 and 13, and  Sartorius lot numbers  1
and 4. Growth was produced on all 6 lot numbers
when  the  filters were placed on TSA medium and
incubated  at 44.5  C.  Since this  1A culture was
unusual,  it will be  omitted  in  the remainder of
results for Table  1. However, further tests were
conducted later on this culture ( Refer to Special
test-A).

    The second category-B shows the percent re-
recovery of  E. coli counts on membrane filters
M-FC  at  44.5  C  compared  with  poured  TSA
plate counts at  35 C. The percent recovery ranged
from  47 to  108%, with culture numbers 1 and  9
providing best results.


    The third category C presents the percent re-
covery of E. coli  on  membrane  filters (TSA) at
44.5 C compared with the poured agar plate count
on TSA incubated at  35 C. The percent recovery
ranged from 55 to 131%, with culture numbers 1,
9  and 3Z providing best results. These results are
applicable only for  the lot  numbers of membrane
filters tested.
     With respect to efficiency of each filter brand
and lot number, the variations apparently were due
to the culture  (1A and  3), or to the filter used
(S-3 with culture 2A). Since one specific culture
dilution was tested  at one time with the different
filters, the efficiency of the latter can be compared
for that particular culture.  Due to  lack of supply,
not all filter lot numbers could be tested with all
cultures  and some of  the results  may  require
further tests.
     Gelman lot number 3, tested with 11 cultures,
had a percent recovery  of 71  to  114% (Category
A), 65 to  109%  (Category  B), and 64 to 131%
(Category C). Compared with  the other lot num-
bers,  G-3  provided  the  best  overall  percent
recovery  for  the  11 cultures tested. Lot number
G-6, tested with 6 cultures, did not provide results
that were as good  as G-3.
     Lot M-13, tested with 5 cultures had a per-
cent recovery  of 55 to  83% (Category A), 51 to
74% (Category B)  and  72 to 95% (Category C).
Lot  M-6  had  a percent recovery  of 83 to 95%
(Category A),  59 to 80% (Category B), and 62 to
96% (Category C). Lot numbers M-3, M-5, M-7,
M-8, and M-12, each were used with  only one test
culture, with M-5  and M-8 producing comparable
results for all categories. Lot M-12  had an unusual
result with culture 24, a high (133%) recovery for
Category  A compared  to  59%  for Category B.
     Lot numbers S-1  and S-3 were used with 4
cultures. Lot S-1 varied in percent recovery, having
64 to 99% (Category A), 53 to 70% (Category B),
and 55 to  84% (Category C). For the same cate-
gories respectively, lot S-3 had recoveries of 44 to
87%, 47 to 82%, and 70 to 107%.


     Lot numbers S-4, S-5, S-6, S-7, and S-8 each
were tested with 2 cultures.  Lot S-4 provided best
results with culture #3A, and S-7 was good with
culture #24,  but not with culture #9.  Lot S-5 was
good with  culture 2A but not quite as good with
culture #9.
     There were some filter lot numbers that gave
nearly  identical results with  the same culture,
such as S-5 and G-3 with culture #9, S-6 and G-6
with culture  #16, and M-6  and S-3  with culture
#1.

     In  the test with culture  #1 where  no  filter
was used,  the recovery on poured agar plate counts
for Categories A and B was lower than  that ob-
tained with the filters. Further  tests (Special test —
A) indicated  a  much lower count for poured agar
plates with this culture.
     The results in Table I are an indication of the
erratic behavior of membrane filters with  respect
to recovery of E. coli  test  strains.  Filtering the
culture  dilution  and placing  the membrane filter
on TSA medium at 44.5 C sometimes gave  a lower
percent recovery (57 to  69%, G-3 and S-4, culture
3A, Category C) than when the filter was placed on
M-FC at 44.5 C (102 to 104%, Category A). How-
ever, the percent recovery (Category  B) of culture
3A  was similar when filters were placed on M-FC
(pad) at 44.5 C (60 to 70%)  or on  TSA at 44.5 C
(57  to  69%)  and  compared with  poured  TSA
plate count at 35 C.
                                               85

-------
    The opposite effect was obtained when  cul-
ture #9 was used with filter lot number S-5,  and
filter G-3 was used with cultures #9 and 4.
                            *
Comparison of E. coli Recovery on Pour Plates at
35, 43, and 44.5 C

    The comparison  of  the percent recovery of
nine E. coli cultures (average count of 5 replicates),
obtained by the pour plate method on TSA  and
M-FC media incubated at 35 (air), 43 (Blue-M),
and 44.5 C (GCA) is listed in Table 2.  Results ob-
tained  with  culture 1A  are  not  included as it
did not grow on M-FC or TSA at 44.5 C, and was
markedly inhibited  on M-FC at 43 and 35 C (30
colonies), but had good growth on TSA at 43 and
35 C (126 colonies).
    The percent recovery on M-FC agar as com-
pared  with TSA was best  for culture numbers 9,
16, and 24 at 35 to 44.5 C. The lowest percent
recovery  occurred when the counts of the other
6 cultures (1,2, 2A, 3A, 3Z, and 4) on M-FC were
 TABLE 2   PERCENT RECOVERY OF E. COLI CULTURES ON TSA AND M-FC MEDIA (POURED
            PLATE COUNTS9) INCUBATED AT 35, 43, AND 44.5 C.

            % Recovery of E. coli on media at temperature listed.
Media
Temp



TSA
44.5







TSA
43







TSA
35




E.G.
No.
1
2
2A
3Z
3A
4
9
16
24
1
2
2A
3Z
3A
4
9
16
24
1
2
2A
3Z
3A
4
9
16
24
TSA
44.5
100
100
100
100
100
100
100
100
100
100
103
108
116
90
99
107
117
118
90
109
100
116
88
113
114
109
129
TSA
43
100b
97
93
86
111
101
93
85
85
100
100
100
100
100
100
100
100
100
90
105
93
100
98
114
106
93
110
TSA
35
111
92
100
86
113
89
88
92
77
111
95
108
100
102
88
94
108
91
100
100
100
100
100
100
100
100
100
Media E.G.
Temp No.
1
2
2A
M-FC 3Z
44.5 3A
4
9
16
24
1
2
2A
M-FC 3Z
43 3A
4
9
16
24
1
2
2A
M-FC 3Z
35 3A
4
9
16
24
TSA
44.5
76C
75
76
70
62
51
98
104
112
69
65
74
66
49
49
93
107
100
63
59
90
52
51
53
96
107
112
TSA
43
76
72
70
61
69
51
91
89
95
69
63
68
57
54
50
86
91
85
63
57
83
45
57
54
89
91
95
TSA
35
84
69
76
61
70
45
86
96
86
77
60
74
57
55
44
82
98
77
70
55
90
45
58
47
84
98
86
  a  Average of 5 replicates for each medium at each temperature.
  b  Average count on TSA at 44.5 C ^ average count on TSA at 43 C.
  c  Average count on M-FC at 44.5 C + average count on TSA at 44.5 C.
                                             86

-------
#1 -
#2 -
#2A-
#3A-
#3Z -
83 to 121%
7 9 to 126%
82 to 122%
79 to 127%
74 to 135%
#4
#9
#16
#24

compared  with  TSA  at all 3 temperatures of in-
cubation. The  percent recovery, when counts on
TSA were compared with TSA at the 3 tempera-
tures, dropped below 85% in only one instance
(77%, culture #24). Although not listed in Table 2,
the counts obtained  for  each culture on M-FC
agar were cross-compared  for  the  3 incubation
temperatures, and correlation in percent recovery
was obtained as follows:
                               93 to 108%
                               95 to 105%
                               98 to 102%
                               89 to 112%
From the results  of the poured  plate counts, it
appears that  E.  coli cultures #9, #16 and #24
would  be a good choice for a standardized test.
Comparison  of  E. coli  Recovery  on TSA, M-FC,
and VRB at 35,43, and 44.5 C

     The percent  recovery of 9  E. coli  cultures
as poured plate counts  on TSA, M-FC, and VRB
media  incubated at 35 (air), 43 (Blue-M), and
44.5 C (GCA) is listed in Table 3.  Cultures 1A and
4A (Klebsiella)  which did not grow at 44.5 C on
TSA and M-FC  in previous tests, produced similar
results on VRB media  and will be excluded. As
noted in  prior tests (Table 2), the percent recovery
of all nine cultures on TSA at the three tempera-
tures showed close correlation (83  to 121%).

     The percent recovery of M-FC compared with
that of TSA, was lower at all three temperatures.
Culture  numbers 2, 2A, 3A, and 9 gave best
results (61 to 98%), culture numbers 1 and 4 next
best (54 to 80%), while cultures # 24 (21  to 40%),
#3 and #16X  (4  to 10%) were very low. The re-
sults obtained  with VRB agar were  better than
with M-FC for  all nine cultures.  Recovery of cul-
tures #2 and #3A on VRB were nearly  identical
with TSA at all three temperatures. Culture num-
bers 1, 2A,  and 9 also had a good recovery (72 -
107%), with culture numbers 4 and 24 somewhat
lower  (24 to 73%)  and numbers 3 and  16X the
poorest (0-29%).

Comparision of E. coli  Recovery on TSA, M-FC,
and VRB at 35 and 44.5 C

     Twelve E.  coli cultures were tested using
poured plate counts on TSA,  M-FC and VRB
media  incubated at 44.5 (GCA)  and 35 C (air)
(Table 4).  The percent recovery, based on the
counts obtained on  TSA  at  44.5 C, was  good
(± 10%)  for  10 of the 12 cultures.  Recovery of
cultures 3X and 24B were low (73 and 77%)  on
TSA at 35 C.  When  the counts  on  M-FC  were
compared with those on TSA, the percent recovery
varied  from 31 to 88%, with cultures #16X and
#2Z lowest at  31  to 47%.  Seven  cultures  (num-
bers 2B, 2Z,  16X, 16Z, 24,  24B,  24X)  had a
similar (± 5%) recovery on M-FC at 35 and 44.5 C
when compared with TSA at 44.5 C, and at 35 C.
Culture 3A was better on M-FC at 44.5 C than at
35 C (82%  compared  with 64%). Culture #24 had
74% recovery on M-FC at 44.5 and 35 C vs. TSA
at 35 C, compared to 69% on TSA at 44.5 C.

    The percent recovery, varying  from  50 to
105%,  on VRB  was higher than on M-FC at both
temperatures.  Culture  #16X was again low (32 to
40%)  in recovery. Six cultures, numbers 2B, 3A,
3X, 3Z,  16Z, and  24 had  recoveries between  71
to 105%.

    Variations  noted  with M-FC  media were ap-
parent with VRB. Culture numbers 2B and 24 had
higher  percent recovery (85 to 96%) when VRB at
44.5 C was compared with TSA at  44.5 and 35 C,
than when VRB at 35 C (71 -80%) was compared
in similar manner. The reverse  (VRB at 35 C  vs.
TSA at 44.5 and 35 C) occurred with cultures 2X,
3A, 3Z, and 16Z where recovery was 82 to 105%
on VRB at 35C as compared to 69 - 86% on 44.5C.

    Culture  numbers  2Z,  3X,   16A and  24B
varied  in another manner. On VRB incubated at
44.5 and 35 C and compared with  TSA at 44.5 C
the recovery was 75  - 100%, while  the TSA at
35 C, the recovery was 63 - 77%.

Comparision of Fecal Coliform  Recovery on M-FC
and VRB Media at 43 and 44.5 C

     Variations  in  the average colony counts (5
replicates)  and  percent recovery  of  fecal  coli-
form  bacteria were apparent for different  brands
and lot numbers of membrane filters in two dif-
ferent  tests (Table 5). Sartorius lot 8 and Gelman
lot  7  provided the best  results when 20  ml of
filtered creek water were incubated on 2  lots of
M-FC and VRB media at 44.5 C (GCA). Millipore
lots 6  and  11 gave the poorest results. While the
colony counts and percent recovery of the other
six  lots were comparable in  most cases after incu-
bation at 44.5 C, they were one-fourth to one-half
that obtained  on  VRB  agar  at 43  C (Blue-M).
                                             87

-------

tn
Q
z

CO
in
CO
h-
Q
LU
H
OQ
O
Z

CO
to
z

o
o
LU
|-
_J
a.
Q
LU
tr
D
O
Q.
<
Q
LU
2
m
QC

Q
Z
^
co"
LL
^^
*»
-
12
Z
o
CO
LU
DC
D
t—
O

O
O
LU
LL
O

DC
LU

0
CO
LU
DC
h-
Z
LU
CO
DC
LU
CL

CO
LU
	 1
m
1-
























-o
CD
+-I

0)
3
to
CD
a
£
+-*
CO
CO
"§


c
o
1
LU
"o
a>

cd
DC
°
























CO CO
1—



to «*
1-


< in
tO Tfr'
i ^~


<-> 0
LU Z



co cL
TJ P
.£




o o o>
^" rL


<






CO 00 00





CO 00 00





00 CO O)





<








00 tO CN




oo en en





o o o
o o o





<








0



°



o



CO





m
CO





CO





CO






CO



co
LL
s


CO




en





o
o












CO
h-
o in oo CN in



o in i-> CN CD



o co en CM in
*~'~


< X
«- CN




s
en co co CN





C3> LO CD CN





LO LO o oo ^r
O) co r~ CN




< X

«- CN


LO
5


O CN CN CN O




O CO 00 O O





o o o o o
o o o o o





< X

«- CN




5
^~
CO CO 00 O
r**» 05 r*1"" c^



PI g °° J;



00 O 00 «-
*™


<





m
DC
r» r~- rv





CD r*»- r*>





O «t CO  CD •<»•
«- CN



"


o o o en oo




88888





CN CO LO LO CN
o r- r- en en





< X

«- CN




CO
^
SScnSS^



O 00 LO O5 CN ^~
,


O> O) O CN O IT)
^— r—


-CNCNCOCO^





m
DC in
SLO^COCO





in r^ r~ oo in





t CO CO C5 ^- CO
co co r^ co CD




< <
«- CN CN CO CO t



CO
LL in



888888




^c in o in oo co
o o o oo en o>





CO CO 00 O5 O 00
«- «- O 00 O O





< <






3*>
L_ CO
5



CO



(35



m






CO





CO





en
^





en







o
o




in
en





00
o














SS



CO CN



CN in



X
«- CN





"S





^ °





r^ r»
CO




X
CD t
T- CN






§O
0




CO «-
O CN





to <-
O) «-





X

<- CN














































8
X
o
CO
•q-
4-J
CO
to
1-
c
o
c
o
o
•- £
~D ci)
CD >
C CO
JI •'•
(J LO
S 5
O -^
4— (D
S3 <
" to
CO l_
c_) ^^
™ c~
Q. 0
i_ •*-*

O
"o o
CD CD
O) O)
CO CO
CD CD
CO CO
co -Q
88

-------
TABLE 4   PERCENT RECOVERY OF E. COLI CULTURES ON TSA, M-FC, AND VRB MEDIA
           (POURED PLATE COUNTS)9 INCUBATED AT 35 AND 44.5 C.

                 % Recovery of E. coli on media at temperature listed
Media
Temp.




TSA
44.5











TSA
35





E.G.
No.
2B
2X
22
3A
3X
3Z
16A
16X
162
24
24B
24X
2B
2X
22
3A
3X
32
16A
16X
162
24
24B
24X
TSA
44.5
100
100
100
100
100
100
100
100
100
100
100
100
93
90
125
100
130
103
106
91
111
93
133
114
TSA
35
108b
100
90
100
77
97
94
110
90
108
73
88
100
100
100
100
100
100
100
100
100
100
100
100
Media TSA
Temp. 44.5
62
47
39
82
M-FC 79
44.5 60
61
36
76
69
72
56
66
55
47
64
88
M-FC 48
35 36
32
70
69
76
59
TSA
35
68
53
31
82
60
58
58
40
68
74
53
49
72
61
38
64
67
47
34
35
63
74
55
52
Media TSA
Temp. 44.5
88
69
86
80
VRB 100
44.5 85
75
59
86
85
86
77
74
82
81
88
97
VRB 95
35 75
50
105
71
93
83
TSA
35
96
76
69
80
77
83
70
65
78
92
63
68
80
91
64
88
74
92
70
55
95
77
68
73
a average of 5 replicates for each medium at each temperature.
° average count on TSA at 44.5 C -=- average count on TSA at 35 C x 100.
    Transparent colonies  were  frequently  ob-
served on the S-8 and M-6 filters placed on M-FC
medium lot A (H6DBXP),  but  not on  lot  B
(910666).  The transparent  colonies were lactose
negative, dextrose acid  on  triple  sugar iron agar
(BBL) and  IMViC reactions were—+—.

    The results for Test 1  in Table 5 were ob-
tained from 20 ml of creek water  which normally
produced  25   or more  colonies.  While  the  low
counts obtained with M-FC and  VRB at 44.5  C
may appear  insignificant,  they  do  indicate  a
decided drop  from the  count obtained  on VRB
agar at 43 C.

Comparison  of  Fecal  Coliform  Recovery  with
Different  Membrane  Filter  Lots on M-FC  and
VRB Media
     In Table  6 two different volumes (10 and
20 ml) of one sample of creek water were tested
with 5 lots of membrane filters on M-FC and VRB
media incubated  at 44.5 C (GCA). The fecal coli-
form counts per  100 ml water were higher (80  -
155) on  VRB  agar than on M-FC medium (60  -
125). The most consistent results on M-FC medium
were produced by Gelman  lot 7,  Sartorius lot 5,
and  Millipore lot 7  for the two  volumes  tested.
However,  the  difference in count per 100 ml,
between  manufacturer and  lot number,  are ap-
parent (M-8, 60-125 vs. G-7, 110 - 120).

     The  highest  count on VRB agar was obtained
with G-7 filters, with S-5 next highest and the
other three filters about equal. The M-8 filter was
the  only  one that  recovered  100%   more fecal
coliform  on M-FC medium than on VRB agar. The
                                            89

-------
TABLES   PERCENT RECOVERY OF FECAL
           COLIFORM FROM 20 ML. CREEK
           WATER USING MEMBRANE
           FILTERS ON TWO LOTS OF M-FC
           MEDIA AT 44.5 C AND ON VRB AT
           43a AND 44.5 C.

       Media, incubation, lot no., count °
Test
No.

1


2

Filter,
Lot No.
M-6
G-7
S-7
S-8
M-11
M-12
G-7
S-8
M-FC
Lot A
1b
8
4
6
15
27
30
19
44.5
LotB
3
9
5
6
24
32
30
20
VRB
43
20
32
23
25
50
58
56
44
Agar
44.5
5
8
7
5
23
30
28
23
a 43 - actual temperature 41.8 to 43.6 C in plate.
" Average colony count, 5 replicates.

percent recovery  of the other four lots varied from
60 to 94%.

Comparison  of  Fecal  Coliform  Recovery with
Different Membrane Filter  Lots at 43 and 44.5C

     Results of six daily tests using 20 ml volumes
of creek water with different lots of filters on
M-FC and VRB media incubated at 43 and 44.5 C
are summarized in Table 7.  For  the percent re-
covery of fecal coliform,  the  average count  (5
replicates)  M-FC at 44.5 C, (GCA) was compared
with  the averages obtained  on  M-FC at  43  C
(Blue-M)  and VRB  at  43 and 44.5  C for each
individual filter lot number.

     Overall, the  lowest percent  recoveries were
obtained  when  M-FC  at 44.5  C  was compared
with VRB at 43 C. At times the percent recovery
on M-FC at 43 C was identical with that obtained
with VRB at 43 C (M-8, G-6, S-8). In the majority
of tests, the percent recovery was as good or better
on VRB as on M-FC at 44.5 C.

     Gelman lot  7 was  most consistent,  and the
best tested in four out of five tests.
Comparison  of  Fecal  Coliform  Recovery  on
Rinsed and Unrinsed Membrane Filters

     In an effort to determine whether residues on
the membrane filters were affecting fecal coliform
counts, four different lots were rinsed  in sterile
phosphate buffer, pH 7.2,  just prior to use.  A
separate beaker, containing 100 ml of  buffer, was
used for rinsing  no more than five filters. The only
difficulty encountered was the  reduction in flow
when the M-13 filters were transferred  from the
buffer rinse to the filtration unit.

     As  indicated  in Table 8,  filter  rinsing im-
proved the percent recovery of fecal coliform on
M-FC media with all except the M-6  filters. The
 TABLE 6   FECAL COLIFORM COUNTS8 PER 100 ML. FROM 10 AND 20 ML. CREEK WATER,
             FILTERED AND INCUBATED ON M-FC AND VRB MEDIA AT 44.5 C.
    Average of 5 replicates.
                                          M-FC
                               VRB
Filter,
Lot
M-7
M-8
S-5
S-6
G-7

count/100 ml.
count/100 ml.
count/100 ml.
count/100 ml.
count/100 ml.
20ml.
44.5
75
125
75
75
120
10ml.
44.5
70
60
80
60
110
20ml.
44.5
90
100
110
125
155
10ml.
44.5
120
110
130
80
150
                                             90

-------
 TABLE  7  PERCENT RECOVERY OF FECAL COLIFORM FROM 20 ML. CREEK WATER USING
            MEMBRANE FILTERS ON M-FC AND VRB MEDIA INCUBATED AT 43a AND 44.5 C.
                   % Recovery
                   Recovery
Test
No.



1






2






3


Filter
Lot No.
M-8
G-6

S-6
S-7
S-8

M-1

M-3

G-7
S-2

M-5

M-6

G-7
S-3
M-FC
43
20
50

25
150
67

100

29

42
58

77

72

90
82
VRB
44.5
33
100

50
150
100

150

83

160
117

93

98

96
97
VRB
43
20
50

20
100
67

60

45

40
47

82

76

84
78
Test Filter
No. Lot No.
M-6
S-8
4
G-7
S-7

M-11
M-12
5
G-7

S-8

M-7
M-8
6
G-7

S-5

M-FC
43
NDC
ND

ND
ND

63
84

100

95

78
89

98

104

VRB
44.5
20
120

100
57

65
90

107

83

58
64

92

85

VRB
43
5
24

25
17

30
47

54

43

41
49

68

71

 3 43 - Actual temperature 41.8 to 43.6 C in plate.
 ° Average count 5 replicates (M-FC at 44.5 C) ^ average count 5 replicates, (media and temp, listed)
   x 100. For example, in Test 1, M-8, M-FC at 43 C, the 20% means that M-FC at 44.5 C recovered
   20% of M-FC at 43 C.
 c Not done.
non-rinsed filters recovered 73 - 81% of the buffer
rinsed filters.

     On VRB agar,  the non-rinsed filters of M-6
and M-13 recovered 80% of the counts obtained on
the rinsed filter, while with the G-7 and S-6 filters
the percent recovery was  104% and 110% respec-
tively.  The M-13 filters gave the only consistent
results (75 - 80%) on M-FC and VRB media.
Special Test - A

E. coli Grown on Surface or Embedded in Medium

    Culture #1A,  originally isolated as  a deep
blue colony on  M-FC medium incubated at 43 C,
produced few colonies on TSA medium and none
on VRB and M-FC agars at 44.5 (GCA) as poured
plate counts. This #1A culture grew best on TSA
plates with marked inhibition noticeable on M-FC
and VRB when incubated at 43 and 35 C.
     In a repeat test, culture #1A was plated using
M-FC and TSA agars (5 replicates each). One ml of
diluted culture was used for  15 x 100 mm plates
and 0.1  ml of the same dilution and a more con-
centrated suspension was  used with  two sets of
12 x 50 mm plates. No growth occurred on any of
the poured  plates after overnight incubation at
44.5 C in the (GCA) water  bath.
    In contrast, the #1A culture on membrane
filters and  pads with M-FC  medium that were
incubated at 44.5 C had growth on Gelman lot
numbers 3 and  6, but not on Millipore lot num-
bers 3 and  13 and Sartorius lot numbers 1 and 4
(Table  1). Growth  was  produced on all six lot
                                             91

-------
TABLE 8   FECAL COLIFORM RECOVERY FROM 20 ML. CREEK WATER USING BUFFER RINSED
           OR NOT RINSED MEMBRANE FILTERS ON M-FC AND VRB MEDIA INCUBATED AT
           44.5 C.
M-6
                              M-FC-Pad
                                                       VRB - Agar
Filter,
Lot No.
No
Rinse
Rinse
No
Rinse
Rinse
           Mean count a
                %b
14         12
     117
16         20
      80
M-13       Mean count         12         16
                %                  75
                                                       20         25
                                                             80
G-7        Mean count         26         32

                O/                  O *I
                /O                  O I
                                                       27         26
                                                             104
S-6        Mean count         16         22
                %                  73
                                                       23         21
                                                             110
a Average of 5 replicates

" % = Mean counts, no rinse -=- rinse, x 100
numbers when  the  filters were placed on TSA
medium and incubated at 44.5 C.
     The average colony count of 0.1 ml of con-
centrated suspension of E. coli #2, filtered, and in-
cubated on TSA agar at 35 C was 115, while that
on  an  M-FC  (pad) was  101. The average colony
count of the 1 ml  filtered and placed on a pad
containing M-FC broth and incubated at 35 C was
179. Due  to  excess  moisture  and  the  volume
(0.1 ml) streaked  on the 12 x 50 mm plates, the
bacterial  growth  ran  together and counts were
too erratic to be used.
     Culture  #2 was  tested  by filtering  (M-13)
0.1 ml and 1 ml of two different concentrations
and placing the filters on 12 x 50 and  15 x 100
mm  plates of M-FC  and  TSA. Then the same
volumes (0.1  and 1.0 ml) were respectively spread
on M-FC and TSA agar plates (12 x  50 and 15 x
100 mm). Five replicates of each category were in-
cubated in a 44.5 C (GCA) waterbath.
                                                    The agar plates that were streaked or received
                                                the filters  (1 ml)  had average  colony  counts
                                                (44.5 C) as follows (E. coli #2):
                                                E. coli #2

                                                1 ml (filter)

                                                1 ml (filter)

                                                1 ml (streaked)

                                                1 ml (streaked)
                                    12x50

                                     119

                                     169

                                     omitted

                                     omitted
                     15 x 100

                       134

                       204

                        73

                        71
Media

M-FC

TSA

M-FC

TSA
                                                     The average counts (71 - 73)  of the surface
                                                streaked (1 ml) plates therefore were much lower
                                                than when the filters were placed on agar or pads.

                                                     E.  coli  #2, in two concentrations, was then
                                                tested by the pour plate method using M-FC and
                                                TSA agars in  12 x"50  (0.1 ml) and 15 x  100
                                                (1 ml)  plates  incubated at  44.5 C (GCA bath).
                                                The  colony count of  #2 culture in  the M-FC  agar
                                                with both concentrations and plate sizes was about
                                             92

-------
one-half that which occurred with the TSA agar.
Repeating  the  experiment with  the  #2  culture
gave the same  results  with  poured plate  counts.

     One ml of E. coli  culture #1 was then tested,
as mentioned above for #2,  using filters on agar
(12 x 50 mm plates), and agar poured plates (15 x
100  mm). The results (average colony counts) were
as follows:

Filter         M-FC   TSA     Temperature

M-6            20      24      44.5 C (GCA)

G-5            27      32      44.5 C (GCA)

S-8            20      30      44.5 C (GCA)

S-3            20      23      44.5 C (GCA)

Poured         14      27      44.5 C (GCA)

Poured         10      25      35 C Incub.
     When 0.1  ml of concentrated #1  suspension
was poured in 12 x 50 plates, the average count on
M-FC was 12 and on TSA, 24.  Using 15 x 100
plates, the poured plate counts were 5 on M-FC
and 22 on TSA. Incubation was 44.5  C (GCA)
in both tests.

     It should  be noted that where two  concen-
trations of a culture were  used,  the counts are
comparable only with one  or  the other  concen-
tration.

     There was,  therefore,  a difference  between
placing  the culture  on  the  surface of the agar
medium via a filter or by spreading the liquid and
incorporating the culture in the medium as in the
poured  plate method with  E. coli #1, #1A, and
#2.

Special Test - B

     Tests were  conducted  to  determine:   1)
whether  the  bacteria are trapped  on  the filter sur-
face but  not all will grow,  2) some are  trapped
deep in  the  membrane  and  would grow  through
the other side  (bottom of filter), and 3)  some  of
the bacteria actually escape through the membrane
into the filtrate. It was  assumed  that all  bacteria
trapped  on  the surface of a membrane  filter
would be transferred by contact to the surface  of
an agar plate  . The contact time used was one, two,
and four hours, or the filter was left in place.
     Two cultures,  #2 and 3A, were diluted and
1 ml (5  replicates for  each) filtered through Gel-
man, Millipore,  and Sartorius  membrane filters.
For  comparison, 1 ml was also plated with TSA for
pour plate counts and incubated at 44.5C and 35C.

     One group  of filters was placed upside down
and  a second group right side up on corresponding
pads saturated  with M-FC medium. In a similar
manner the  same brand and lot numbers of filters
were placed on  M-FC agar (Test A) or VRB agar
plates (Test  B) and  incubated at 35 C and 44.5 C.
One  hour later, some  of  the upside down filters
were removed from the plates, with others left in
place.  This  procedure  was again  repeated  after
four hours incubation, and all plates returned to
their respective  incubation  overnight  (Table 9).

     No  bacterial colony  growth  was  evident on
either side of the filters placed upside down on the
pads or from the filtrates. However, the colonies
present on  filters upside down on  the agar plates
could be counted through the agar.  The colony
counts on plates (and percent recovery) with filter.
removed after  one hour  incubation  were lower
than those removed four hours later. The latter
colony counts were higher than where filters were
left  on the plates, (Test A, Table 9), with the ex-
ception of culture 3A on Millipore lot  #13. How-
ever, the recovery of E. coli by the filters on the
pads and agar plates did not equal the poured plate
counts on TSA  medium for culture 2  and 3A on
M-FC at 35 C and 44.5 C.

     In test  B  (Table 9), when the same filters and
cultures were used, there was a marked increase in
present  recovery of culture #2 with filters placed
on VRB  agar at 35 and 44.5 C. A  slight increase
occurred when  filters  were  removed  after  two
hours.  However,  results of E. coli  #3A on  VRB
agar  were about the same as with M-FC medium.
                 DISCUSSION

     To determine efficiency of membrane filters
for the  fecal  coliform test one must test polluted
water samples, and  confirm  that the colonies are
fecal coliform.  To  detect deficiences in the test
procedure,  E. coli strains are used, but these vary
in different reports  (1, 5, 6, 8). However, all test
methods and  equipment  must  be  accurate  and
rigidly adhered to. The water baths and incubators
used must maintain  a temperature of 44.5 C inside
the plate, but actual reports  are scarce. The GCA -
Precision Scientific  Co. water bath was the only
                                               93

-------
TABLE 9
          COLONY COUNTS AND PERCENT RECOVERY ON MEMBRANE FILTERS PLACED
          ON PADS AND AGAR MEDIA AND REMOVED OR LEFT IN PLACE.
                                  M-FC Medium or agar plate and incubation
Test A
Culture






2
2
2
3A
3A
3A
Filter3
Lot No.
G-6 count
%d
S-4
%
M-13
%
G-6
%
S-4
%
M-13
%
Pad
44 35
43
60
35
49
49
68
74
71
57
55
61
59
55
79
32
46
51
73
77
74
67
60
73
65
1 hr.
44 35
21
29
17
24
27
38
23
22
54
52
65
63
31
44
33
47
40
57
24
21
44
39.
38
34
4hr.
44 35
47 57
65 81
36 35
50 50
51 47
71 67
75 86
72 77
43 63
41 56
63 76
61 68
Left on
44 35
18 23
25 33
30 22
42 31
31 32
43 46
50 72
48 64
54 68
52 61
51 65
49 58
TSAC
44 35
72 70
100 100


104 112
100 100


Filter Removed
Test B
Culture






a
b
2
2
2
3A
3A
3A
Filter8
Lot No.
G-6 count
%
S-4
%
M-13
%
G-6
%
S-4
%
M-13
%
Gelman, Sartorius, Millipore
Pili-orc romr\\/orl aftor 1 hr A.
VRB
44 35
40
80
44
88
47
94
24
77
18
58
18
58
(average
hrc nr If
48
92
54
104
49
94
23
82
20
71
14
50
of 5
aft nr
2hr
44 35
56
112
58
116
46
92
18
58
16
52
15
48
replicates)
i nlato
58
112
55
106
60
115
18
64
18
64
15
54
•
4hr
44 35
ND
ND
ND
ND
ND
ND

Left on
44 35
ND
ND
ND
ND
ND
ND

TSAC
44 35
50 52
100 100


31 28
100 100



c Poured Trypticase soy agar plate count.
" Percent recovery, count at 44 •=- TSA count at 44.5 C.

ND = not done.
                       35 -j- TSA count at 35 C.
                                        94

-------
one used  in our tests that kept a steady 44.5 C
temperature inside the  plate and the surrounding
water.

    Three incubation temperatures were utilized
to determine differences  in percent recovery. The
number of bacteria that grew on TSA medium at
35 C  should be an  indication of the number of
E. coli present  in the dilution  being tested. From
this base, comparisons with counts obtained at 43
and 44.5 C were made, not only with TSA medium
but also with M-FC and VRB. The 43 C incubation
temperature (41.8 to 43.6 C inside plate) was used,
as we had  noted these temperatures in improperly
working water  baths and  air  incubators  set at
44.5 C.

    In prior reports, the exact number of bacteria
cells being tested was based on poured plate counts
(8), or on counts obtained on membrane filters
with pads, using a non-inhibitory medium (5). That
either or both of  these methods could effect the
results obtained with E. coli cultures was apparent
in our study.

     The results we obtained indicated that eleven
of the E.  coli test strains varied not only in their
ability to  grow at 35, 43, and 44.5 C, but also on
or  in the  culture  media used. Culture 1A would
grow on TSA agar, but not on M-FC agar at 44.5 C.
Theoretically,  a   temperature  tolerant  culture
should grow as well at 35 C as at 44.5 C on a non-
inhibitory  (TSA) medium, but this did not always
occur (Table 1, cultures 2 and 9). At times, the
percent recovery of membrane filters on M-FC vs.
TSA at 44.5 C was  better than  that obtained on
TSA at 44.5 (C)  and 35 C (B), (Table 1, cultures
2A and 3A). This is similar to  the results obtained
by  Presswood and  Brown  (8) with three E. coli
cultures. However, while they used M-FC medium
for both  the membrane filter and poured  plate
counts,  there  is  agreement that  the  membrane
filter  was  responsible  for  low  percent recovery
more frequently when M-FC medium was used.

     When the E.  coli  strains were tested  by the
pour  plate method,  the percent recovery of nine
cultures on TSA  medium was very close for the
three  temperatures of incubation (Table 2). In con-
trast,  only three cultures (9, 16, 24) grew as well
on  M-FC   as on  TSA  at the  three temperatures
of incubation. The addition of VRB medium for
comparison with TSA and M-FC media by the pour
plate  method indicated  that  VRB had a better
percent recovery than M-FC (Tables 3, 4). Culture
#9  again  produced good  results, but in contrast
culture #24 showed inhibition on M-FC and VRB
agar. This would indicate variations of temperature
tolerance in some of the test cultures.

     A  comparison  of  VRB  agar  with  M-FC
medium in respect to recovery of fecal coliforms
from polluted water indicated that VRB was better
than M-FC when the  incubation temperature was
44.5 °C (Tables 5, 6, 7),  and quite  superior at
43 C (Table 5).

     The  possibility  that  the membrane  filter-
culture media interaction was responsible for low
recovery  was obvious when the membrane filter
counts were compared with poured plate counts.
An  attempt  to  transfer bacteria on the membrane
filter  to  M-FC  medium did not improve percent
recovery  as  well  as on VRB  medium  (Table 9).
Placing  three lot  numbers  of filters  containing
E. coli #2 on  VRB agar at 35 and 44.5 C  pro-
duced  much better  results  than on  M-FC with
pads (Table 9, Test B). An improvement in percent
recovery  was also noted when the same culture
and membrane  filters were  placed face down on
the VRB agar, and then removed two  hours later.
However, when E. coli #3A was tested in the same
manner,  there  was  little difference  in  percent
recovery  on  VRB and  M-FC  media.  The results
obtained  on M-FC with  E. coli #2 and  #3A (Table
9, Test A) are similar  to those obtained in Table 1
(Category B).  Therefore, we must assume  that
VRB was better than M-FC,  at least for culture #2
in this test.  Whether placing the membrane filter
face down on the VRB agar  and stimulating better
growth by direct contact can be compared with the
two  layer technique  (7, 9) requires further study.
     Temperature alone did not affect the growth
of the nineteen E. coli  strains tested  in an equal
manner since all grew well on TSA medium. The
E. coli strains were affected more by their ability
to form less colonies on M-FC agar than on TSA
and VRB  media. Differences in strains or isolates,
with  some showing  no  growth  inhibition while
others were markedly inhibited, were apparent. It
therefore  appears  that  the  M-FC  medium  did
support good growth of  some, but not all, of the
E. coli strains. This applied to pour plates as well
as membrane filter tests. Since one lot  medium
was used for most all tests, different lot numbers
were not responsible.

     A  comparison of  different brands  and lot
numbers of membrane filters indicated variations
in percent recovery of E. coli cultures were due to
                                              95

-------
the filter, the culture, or the test method. Culture
#1  as poured plate count recovered 52 to 58% of
the E. coli  on M-FC when compared  to TSA at
44.5 and 35 C (Table 1), and similar results were
obtained with poured plate counts (Special Test
A). In  each of these two tests,  the  membrane
filter  counts were higher (67 - 109%)  and mem-
brane filters M-6, S-8, and  S-3 produced identical
results in respect to  percent recovery (67 - 87%)
at 44.5 C.  Cultures  #1A  and 2  also produced
lower counts  when  tested  by the poured  plate
method  and  compared with jnembrane  filter
counts.  The M-FC agar was more  inhibitory  (low-
ered counts) than  the TSA for these cultures via
the membrane filter or poured plate method. With
culture  #2,  this inhibitory activity of  M-FC was
less marked  when streaked  on the solidified M-FC
and TSA plates, even though the surface streaked
count was much lower than  the poured plate count
(Special Test A).

    Therefore, in  evaluating the efficiency of the
membrane filters in respect  to percent recovery of
E. coli  bacteria, the  ability of the test strain to
grow  in or on the medium should be considered.
In Table 1, when the membrane filters were placed
on M-FC medium (pads) or on TSA agar, Category
A probably  provides a fair  comparison  of the dif-
ferent  lot  numbers.  That  the percent recovery
might  be  due  to  the method, i.e.  Category C,
where  the  membrane filter was  placed on TSA
(44.5  C) and compared with poured plate count
on  TSA (35 C) is evident with cultures #1  and
#9 vs. #2 and  #3A.  All four cultures  (in fact all
cultures tested),  were  almost equal  in  percent
recovery when compared at 44.5 C and 35 C on a
poured plate basis  (Tables 2, 3, 4). In this respect,
there  was  little difference reported  in  bacteria
counts  obtained  when  membrane  filters  were
tested  with  plate count broth and total coliform
broth at 35  C, but a  decided drop occurred in the
counts when plate count  broth and M-FC broth
were used at 44.5 C (5).


    Why  the  membrane filters are  so erratic  is
still  not clear.  Rinsing the  filters prior  to use
helped  somewhat  \jn  limited  tests  (Table 8). The
bacteria trapped on  the membrane  filter surface
required more than one hour contact time with a
solid  culture medium to grow. It is possible the
bacteria on  the filters may  be  in clusters, or some-
how  injured and  do  not  grow as readily  as in
poured  agar plates.   However,  we did find  two
strains of E. coli that grew better on  the surface
of the medium than  in the  poured agar. If this was
due to melted  agar  being too hot, all samples
done at that time  would have been affected, but
this did not occur.

    While  it appears that VRB medium might be
as effective or better  than M-FC  medium,  this
should be  evaluated by  others. Klein and Fung
(6)  reported  that the VRB poured  plate method
was as good as the MPN and membrane filter
method for the fecal coliform tests. However, an
air  incubator at 44.5 C  and only two  replicates
per sample were used in their tests.

                  SUMMARY

    The conclusions to be drawn from this study
emphasize  that a test method  should be standard-
ized with respect to the E. coli strain, the culture
medium used to determine the actual number of
bacteria  in  the dilution  being tested,  the  best
method to  be utilized  (poured  plate,  pad  and
broth, or solidified agar plate) and control of tem-
perature of incubation inside the medium.  Studies
to further  standardize  the culture medium to be
used at 44.5 C have been initiated  (9) and hope-
fully this will provide more uniform blue colony
types.
            ACKNOWLEDGEMENT

     The  technical  assistance  of Greg  Bossart,
Mike Davis, Loretto Yanaitis and Charlotte Smith
is greatly acknowledged.
                REFERENCES

1.    Dutka,  B.J.,  M.L. Jackson, and J.B.  Bell.
     Comparison of autoclave and ethylene oxide
     sterilized  membrane  filters used  in  water
     quality   studies.   Appl.  Microbiology.  28:
     474-480, 1974.
2.    Evaluation of Water Laboratories, U.S. Dept.
     H.E.W.  Public Health  Service Public No. 999-
     EE-1,p44, 1966.
3.    Fishbein, M. The aerogenic response of E. coli
     and strains  of Aerobacter in EC broth and
     selected sugar -broths at  elevated  tempera-
     tures. Appl. Microbiology. 10:79-85, 1962.
4.    Hartman,  P.A.,  P.S.  Hartman, and  W.W.
     Lanz. Violet red bile 2 agar for stressed coli-
     forms. Appl.  Microbiology. 29:537-539, 1975
5.    Hufham, J.B.  Evaluating the membrane fecal
     coliform test by using E. coli as the indicator
     organism.  Appl.  Microbiology.  27:771-776,
     1974.
                                               96

-------
6.   Klein, H. and D.Y.C. Fung. Identification and
     qualification of fecal coliform using violet red
     bile agar at elevated temperature. ASM Alle-
     gheny Branch Meeting, Pittsburgh, Pa., 1974.
7.   Nash, H.D. and  E.E. Geldreich. Applicability
     of  a  modified membrane fecal coliform med-
     ium  and technique.  ASM abstracts,  75th
     Meeting, N.Y., N.Y., 1975.
8.   Presswood, W.C. and L.R. Brown. Compari-
     son of Gelman and Millipore membrane filters
     for enumerating fecal coliform bacteria. Appl.
     Microbiology. 26:332-336, 1973.
9.    Rose,  R.E., E.E. Geldreich, and  W. Litsky.
     Improved membrane  filter method  for fecal
     coliform  analysis. Appl.  Microbiology. 29:
     532-536, 1975.
10.  Shahidi, S.A. and M.H. Backer. Effect of fecal
     coliform organism medium. ASM Abstracts,
     73rd Meeting, 9.,  1973.
11.  Sladek, K.J.  Optimum membrane structures
     for growth of coliform organisms.  Ft. Lauder-
     dale, Fla., This Symposium, 1975.
12.  Standard  Methods for the  Examination of
     Water   and  Wastewater.  APHA,  AWWA,
     WPCF, 13 ed. 1971.
                                              97

-------
                    RECOVERY CHARACTERISTICS OF BACTERIA INJURED
                          IN THE NATURAL AQUATIC ENVIRONMENT
                                               by
                               Gary K. Bissonnette, James J. Jezeski,
                              Gordon A. McFeters and David G. Stuart
                                   Department of Microbiology
                                     Montana State University
                                        Bozeman, Montana
                  ABSTRACT

     The recovery and  enumeration of indicator
organisms from  natural waters were found to be
adversely  influenced  by  several  phases  of the
system. These  dictated the  degree  of sub-lethal
injury  observed  in water-borne bacteria that ren-
dered them  incapable of growth under conditions
that are routinely  used for their identification.
Therefore,  the  recovery  of indicator organisms
following  water-induced  injury   was  examined
using various isolation procedures. A high degree of
variation was found  and the membrane filtration
technique was  the   least  efficient.  The  injury
inflicted upon   E.  coli and S.  faecalis  by the
aqueous environment  could be reversed by a short
enrichment  treatment.  This  procedure  was  es-
pecially applicable to  membrane filtration  pro-
cedures.
                INTRODUCTION

     In evaluating the problem of detecting par-
ticular  microorganisms  from  various   sources,
proper  consideration  must be given  to the in-
fluence of environmental factors upon detection
methods.  Data  are available indicating that after
exposure  to freezing, heating, or freeze-drying,
some  microorganisms are   either physiologically
debilitated or   injured  to   such  an  extent  that
significant problems  arise  upon  attempts at de-
tection  and  enumeration.  Such  stress-injured
microorganisms    become  more  sensitive to in-
hibitory agents  in specific selective media and are
unable to grow and produce colonies.

     Most sanitary indicator organisms and enteric
water-borne pathogens are  bacteria whose natural
habitat  is  the  intestine  of  man  and  warm-
blooded animals. Once  these  microorganisms are
deposited  into water they are in an environment
that is not favorable to the maintenance of viabil-
ity  for most heterotrophic  bacteria.  Therefore,
proper  interpretation of  sanitary  water quality
data  relies  partly  on  a basic understanding of
survival characteristics of bacteria in water. In the
majority of  reported survival studies, only two
sub-populations of the total have been considered:
those cells which  can withstand  the aquatic en-
vironment  as reflected  by  their detection and
enumeration  when  using  standard  laboratory
procedures  and  conversely,  those cells  which
cannot  persist  in  the unfavorable environment,
resulting in death  and non-detection.  There is a
paucity  of  available literature  concerning  the
possibility  that a substantial fraction  of  the total
population of cells in water may be injured to the
extent that they fail to  grown on selective media.
     This report presents research directed toward
determining whether aquatic environments  pro-
voke stress upon indicator bacteria such that these
cells  become   physiologically  debilitated  and
cannot be detected by direct selective procedures.
The report also  describes  methods to  recover
these injured cells.

     In the  first study, membrane chambers were
filled with washed suspensions of  a typical  EC+
strain of  Escherichia coli, type I,  and  immersed
at different stream sites, and sampled daily. The
organisms were tested  for  their ability to form
colonies on Trypticase soy agar supplemented with
yeast extract (TSY agar), to  indicate the maximum
number of recoverable organisms, and on desoxy-
cholate agar (DLA), to yield the number of bac-
                                               98

-------
teria that were able to form colonies in the pres-
ence of  this selective  medium.  From these  data,
the percentage survival and the percentage death
were obtained  in addition to the fraction of the
total viable population  that  was injured to the
extent  that they  were  unable  to grow on the
selective  medium. This portion of the total viable
population that became debilitated in water repre-
sents the cells that would not be enumerated  using
the current standard procedures.

     The  results of the first study indicated that
the number of organisms in these segments of the
total  bacterial  population  varied   considerably
among the  stream  sites that  had different water
quality  characteristics. Also,  at  the  sites where
the greatest percentages  of death  was  observed,
there was a greater  proportion  of  the bacteria
that were injured. The extent of this non-lethal in-
jury varied  from 10%  to 96% of the total viable
population  after four days,  depending on the
physical and chemical  characteristics of the water.
In addition, it was observed that the proportion of
the  survivors  that  reflected  non-lethal  injury
increased  as the length of exposure to the aquatic
environments increased. This observation was also
found  when comparable experiments were con-
ducted using Streptococcus faecalis.


     Because the  isolation of  various  indicator
bacteria  is the primary objective in  assessing the
microbiological quality of water, the use of selec-
tive media  is  required to suppress the growth of
other organisms  that  could  interfere  with the
detection and enumeration of the desired micro-
organisms. Inhibitory  agents in  these media may
exert unexpected  inhibition on  cells subject to
stress.  Thus, the combination of  environmental
stress with the subsequent utilization of selective
media  may  result  in  the diminished  recovery of
injured  cells.  This  problem  is  further compli-
cated by the varying degrees of injury observed as
a  function  of  time  and  water characteristics.
Therefore, improved enumeration methods should
be  developed  to more efficiently recover injured
bacteria.

     In the  second  study, experiments were done
to  establish the  relative efficiency of various
selective  media  in  recovering E.  coli that  were
progressively debilitated in natural water. Samples
were enumerated daily for four days by different
methods   using several media. These studies  re-
vealed  the superior recovery efficiency  of liquid
media  (MPN)  over broth  plating  and membrane
filtration  methods:  MPN  with TSY  = MPN  with
lactose  broth  > MPN  with  brilliant green lactose
bile  broth >  plating with  desoxycholate lactose
agar >  membrane  filtration  with M-Endo MF
medium  > membrane filtration with M-FC  med-
ium. It  should be emphasized that the membrane
filter procedures were less  efficient in recovering
injured  bacteria found in natural waters than the
other selective procedures  by a  statistically sig-
nificant  margin. Similar  results were found when
comparable  experiments were  done using S. fae-
calis and the appropriate media. These findings in-
dicate that,  while conventional methods might be
adequate to enumerate bacteria from water exert-
ing minimal  environmental stress or cells deposited
into the water shortly before it was sampled, care-
ful  consideration should  be given  to the  develop-
ment of new  methods to  recover  debilitated
bacteria  from  stressful water environments. This
is particularly  true, since injured cells were found
to be a  large fraction of the total population when
exposed  to  certain  aquatic environments for as
little as two days. Therefore, it is important that
this  damaged  population be recovered  to more
correctly evaluate the bacterial quality  of many
waters.


     In   the third study, experiments were con-
ducted  to determine if  cells injured by environ-
mental stress in natural waters  had the capacity to
repair themselves when exposed to  a suitable en-
vironment. An EC+  strain of E. coli was exposed
to water in membrane chambers for two days, the
organisms removed, inoculated into liquid  TSY
medium  and enumerated every hour by plating on
TSY and DLA agar  for  six hours. This procedure
was used to compare the growth  kinetics of bac-
teria exposed  to water with a  control suspension
of  the  same  organism that was  not subject to
environmental  stress. As in the  previous experi-
ments, the difference between the counts on the
TSY  and DLA  media  reflected  the debilitated
population. Control cultures that were not exposed
to  the  water  contain virtually no cells  in  this
weakened  physiological   state and  the  bacteria
exhibited normal  growth kinetics.  However, the
cell  suspension that had  been  exposed  to the
stresses  of the aquatic environment for a period
of  two  days  contained  a  substantial  proportion
(95%) of injured cells, and the lag period  was three
times longer than in the control  suspensions.  As
the injured cells were  exposed to  the TSY broth,
the injured population progressively repaired itself,
so that after three hours, they were capable of pro-
ducing  colonies  on  both  TSY  and  DLA  agar.
Further experiments, where bacteria were in the
water for longer times, demonstrated increases in
                                               99

-------
the lag period and recovery time. The same kinds
of  observations  and  relationships were found
when similar experiments were  conducted using
S. faecalis.  These studies indicate that appreciable
repair from  environmental injury may be attained
by  exposing the  cells  to a  rich, non-selective
medium for a short  period prior to the use of a
selective  medium. This procedure  affords  the
more complete  enumeration of indicator bacteria
from the aquatic environment on selective media.
Enrichment  techniques appear  to  be especially
applicable to membrane filtration methods that
have a low efficiency of recovery for injured cells
and are easily adapted  to enrichment procedures.

     Additional  studies  were  conducted to deter-
mine if a two  hour  enrichment step, before ex-
posure to the selective medium, would enhance the
recovery of total and fecal coliform bacteria from
a mixed  natural population of bacteria, as  they
were in contact with the aquatic environment for
various times.  In these experiments,  suspensions
of raw sewage were placed in  membrane chambers
and then in a stream, followed by daily sampling
using the membrane  filter procedure. Then dupli-
cate filters were incubated on  M-Endo agar at 37 C
and M-FC agar at 44.5 C, after a two hour enrich-
ment step on TSY agar at 37 C. These experiments
indicated that the enrichment procedure improved
the recovery, in comparison to  control samples,
where  the  enrichment was omitted. Throughout
the entire three day period that the bacteria from
the  sewage were  in  contact  with  the aquatic
environment, the most efficient method for re-
covering  total  and fecal coliform  bacteria was  a
two  hour enrichment on TSY  agar  followed by
transfer to the selective medium.  The two hour
enrichment period provides a non-toxic, nutrient
rich  environment for  the gradual adjustment and
repair that  is needed by these organisms to success-
fully grow on the selective medium. The adoption
of enrichment  techniques would  help overcome
some of  the limitations of the  membrane  filtra-
tion  procedure, arising from the exclusive use of
selective  media, by providing the necessary bridge
for bacterial adaptation between the environment
encountered in natural waters and the selective
media in the laboratory.

           ACKNOWLEDGEMENTS

     This project was supported by funds from the
U.S. Department of the Interior  authorized under
the Water Resources Research Act of 1964, Public
Law 88-379, and administered through  the Mon-
tana University Joint  Water Resources  Research
Center (grants  OWRR  B-035  Mont,  and B-040
Mont.).
                                               100

-------
             A LAYERED MEMBRANE FILTER MEDIUM FOR IMPROVED RECOVERY
                                OF STRESSED FECAL COLI FORMS

                         Robert E. Rose, Edwin E. Geldreich and Warren Litsky
                  ABSTRACT

     A  two layered agar method employing tem-
perature acclimation  and lactose enrichment with
diffusion transfer into M-FC agar is  proposed for
improved recovery of stressed fecal  coliforms on
the  membrane  filter. The  procedure  was field
tested  in  three  laboratories  using samples of raw
and  chlorinated wastewater, reservoir,  river and
marine  waters.  Verification  of  1013  fecal coli-
form colonies  isolated from 61  water samples
averaged 92 percent using this proposed procedure.
Comparisons with the Standard Methods M-FC
procedure  revealed the two-layered agar method,
provided an overall increased sensitivity for fecal
coliform  detection   in  chlorinated  secondary
effluents, marine waters and any natural waters
that contained  pollutants with heavy metal ions.

                INTRODUCTION

     Recent reports of reduced  recovery of fecal
coliforms from chlorinated sewage effluents, when
the membrane filter  procedure is used (1-3), have
caused much concern both to regulatory agencies
and  thos  laboratories  involved  in  monitoring.
Until this  problem is resolved, Federal require-
ments for  the bacterial quality  assessment of ef-
fluents  under the  National  Pollution  Discharge
Elimination System has specified that fecal coli-
form densities must be determined by the multiple
tube procedure. As a result of this decision, many
small laboratories are unable to meet this analysis
protocol  because  of a  limited  bacteriological
testing capability based solely on  the  membrane
filter concept.

     Review  of  the  attenuated fecal  coliform
recovery problem suggested that  chlorine inactiva-
tion  of  some coliform cells  might  be  reversed
provided  enrichment  (4,  5)  and  temperature
acclimation  (5,  6)  were  possible  without com-
promising specificity   of the test. All enrichment
procedures,  previously developed  for  the  mem-
brane filter technique, required a manual transfer
of the membrane filter cultures from one medium
to another (4-9). Recognizing media manipulations
is time consuming in the  laboratory so a new ap-
proach  incorporating  a  two  layer enrichment-
differential  growth  medium was explored.  The
method was evaluated on a variety of waters which
might contain attenuated fecal coliforms.

           MATERIALS & METHODS

     Preparation of a two  layer medium  (Table 1)
was  accomplished  by  dispensing approximately
5 ml of M-FC agar into each culture dish (50 x
12mm), permitting the  agar to  solidify, then add-
ing 2 ml of normal  strength lactose broth  in 1.5
percent agar over the M-FC agar. Since the ingred-
ients of the two agar layers will eventually diffuse
TABLE 1.   FORMULATION OF THE TWO-
            LAYER MEDIUM

Differential Medium (Botton Layer)        3.7 gm

    M-FC Medium                       1.5 gm

    Agar                             100  ml

    Distilled Water

Resuscitation Medium (Top Layer) *       0.3 gm

    Beef Extract                        0.5 gm

    Peptone                            0.5 gm

    Lactose                             1.5gm

    Agar                             100  ml

    Distilled Water

* Resuscitation medium equals 1x lactose broth
 plus 1.5% agar.
                                              101

-------
into each  other,  it  is suggested  that  the  base
M-FC agar  be prepared in advance and the lactose
agar overlay added within one hour prior to use.

     After  the membrane filter was placed on the
two-layer medium, the  plates were incubated at
35 C for 2  hours after which the temperature was
increased to 44.5 C for 22-24 hours to attain the
necessary selectivity. All blue colonies were count-
ed with the aid of a binocular  scope employing
10-15x  magnification  and  a  flourescent  light
source. Verification of fecal coliforms isolated on
the test medium  was performed by subculturing
each blue colony  into  either phenol  red lactose
broth or lauryl tryptose broth for 24 to 48 hours
at 35 C. Tubes showing  gas production within this
period were subcultured to E.G. broth and incuba-
ted in a water  bath for 24 hours at 44.5 C ± 0.2 C.

     Samples were collected  from diverse waters
that included  estuarine waters  of Massachusetts,
raw sewage and chlorinated sewage effluents from
the Billerica Massachusetts sewage treatment plant,
polluted stretches of the Merrimack, Fort and Mill
rivers, and  sampling at varying  depths  of  a  raw
water impoundment near Walton,  Ky. The reser-
voir water  samples were collected during a period
of prolonged dry weather and following a signifi-
cant stormwater  runoff into  the  impoundment.
All bacteriological examinations of these waters
were performed at one of  three different labora-
tories located  near the sampling sites:  Millipore
Corporation Laboratory at  Bedford,  Mass.; Uni-
versity of  Massachusetts Department of Environ-
mental Sciences Research Laboratory at Amherst,
Mass.;  and the US EPA Water  Supply Research
Laboratory, Cincinnati,  Ohio. Three or  five repli-
cate portions were prepared for cultivation on both
the  two-layer  experimental   medium  and  the
M-FC agar direct method  as  recommended  in
Standard Methods (10).

         RESULTS AND DISCUSSION

     A  total  of  sixty-one  water samples  were
analyzed  in   the evaluation  of  the  two-layer
medium procedure. The choice of samples used in
this evaluation  were oriented to  those waters that
might have attenuated fecal coliform  populations
resulting from exposure to chlorination  of sewage
effluents,  contact with the marine environment,
antagonistic action of  metal  ions  in chemically
polluted fresh  waters and to  natural  forces of
self-purification  induced  during  storage  of  im-
pounded natural  waters. Data presented in Tables
2 to 5  are based on average colony counts per
membrane filter test rather than as counts per 100
ml  so  as to avoid distortion  from factoring dilu-
tions to the base 100 ml level.

     The fecal coliform colony counts on the two-
layer method were greater than those detected by
the  companion  direct  M-FC procedure  when
chlorinated  sewage  effluents  were  examined
(Table  2). The average colony count ratio obtained
by  the  experimental procedure and the standard
M-FC method was  18.2.  Inspection  of the com-
parative data for raw sewage revealed only three of
18 samples had a  ratio higher  than the lowest ratio
calculated  in  the chlorinated  sewage effluents.
These  preliminary  findings  suggest  attenuated
fecal coliforms are more  numerous in chlorinated
TABLE 2.  COMPARISON OF THE 2-LAYER
           AGAR VS DIRECT M-FC
           PROCEDURES COLIFORM
           DENSITIES FROM RAW AND
           CHLORINATED SEWAGE
                                      Ratio

                 M-FC  2-LayerAgar 2-LayerAgar
     Source     Count     Count    Direct M-FC

Raw Sewage 57
23
32
16
16
10
4
1
2
46
5
15
24
15
25
52
12

103
26
72
20
26
31
11
10
14
91
12
23
33
22
42
98
32
1.8
1.1
2.3
1.3
1.6
3.1
2.8
10.0
7.0
2.0
2.4
1.5
1.4
1.5
1.7
1.9
2.7
5.7
 Chlorinated Sewage6
                  1
                 26
                  5
                          40
228
 26
127
 19
Avg.  2.9

     38.0
     26.0
      4.9
      3.8

Avg. 18.2
                                               102

-------
sewage than might be expected in  raw sewage and
thus support the observations of Lin (1).

     Attenuated fecal coliforms also appear to be
present  in the estuarine samples collected from a
coastal  site  in  Massachusetts (Table 3).  Stevens
et al. (9) were of the opinion that the initial shock
at 44.5  C adversely affected reproduction of meta-
bolically injured cells present in the marine  water
environment. Here again, the two layered medium
recovered from  3.8 to 7 times more fecal coliform
colonies than the direct  M-FC procedure.
TABLES.  COMPARISON OF THE 2-LAYER
           AGAR VS DIRECT M-FC
           PROCEDURES FOR  FECAL
           COLIFORM DENSITIES FROM
           MARINE WATERS
                                    Ratio
M-FC
Source Count

Marine Waters 3
21
3
21
3
30
20

2- Layer Agar
Count

12
79
12
79
16
210
92
Avg.
2-Layer Agar
Direct M-FC
4.0
3.8
4.0
3.8
5.3
7.0
4.6

, 4.6
     Attenuated  fecal coliform  occurrences  in
fresh waters are more varied as related to the in-
tensity of heavy metal ions found in a particular
stretch of river (Table 4). The toxic effect of heavy
metal ions in river water has been  reported to be a
factor in  coliform  recovery  from  transported
samples (11-13). Adsorption of metal ions from a
water sample may also occur on the membrane fil-
ter,  producing  a  concentrated toxic  effect  (11).
Thus, the more frequent occurrence of attenuated
coliforms observed   in the Merrimack River,  as
contrasted  to data  obtained  on  Fort  River,  is
assumed  to be a reflection of the  more numerous
industrial  waste  discharges  to  the  Merrimack
River.

     Results  obtained from a study of  a water
reservoir   supply  (Table  5)   indicate  the   least
amount of difference between the proposed two
 TABLE 4.   COMPARISON OF THE 2-LAYER
            AGAR VS DIRECT M-FC
            PROCEDURES  FOR FECAL
            COLIFORM DENSITIES FROM
            RIVER WATERS

                                      Ratio

             M-FC   2-Layer Agar   2-Layer Agar
     Source  Count     Count     Direct M-FC
Merrimack
River







Fort River





Mill River



4
76
4
76
23
15
8

11
4
7
9
24

6
22


25
183
25
183
53
38
56

21
11
10
16
29

12
60

6.3
2.4
6.3
2.4
2.3
2.5
7.0

1.9
2.8
1.4
1.8
1.2

2.0
2.7

Avg. 3.1
 layered medium  and the direct  M-FC  medium.
 This  water contained  few fecal  coliforms  per
 100 ml during  the day  weather and no industrial
 waste discharge in the drainage basin. Fecal pollu-
 tion that enters from stormwater  runoff is from
 cows grazing on  the  hills surrounding the  reser-
 voir. For these reasons, those few attentuated fecal
 coliforms present, represent debilitated cells com-
 mon to natural die-off.

     Verification  of  1013  typical blue colonies
(Table  6)  from  the two-layer agar procedure con-
firmed  our contention that the reported increased
fecal coliform  recoveries attributed to this pro-
posed  procedure  were  valid. Of  1013  colonies
picked  from all samples  tested, 930 produced  gas
at the  elevated  temperature, for an average verifi-
cation rate of 92%.
                CONCLUSIONS

     The results  indicate  that the proposed two-
 layer agar membrane filter procedure  allows for
                                             103

-------
TABLES.  COMPARISON OF THE 2-LAYER
           AGARVS DIRECT M-FC
           PROCEDURES FOR FECAL
           COLIFORM DENSITIES FROM
           RESERVOIR WATER
                                     Ratio
             M-FC
     Source  Count
2-LayerAgar  2-LayerAgar
   Count     Direct M-FC
Water Supply
Reservoir
(Dry Period)








Water Supply
Reservoir
(Storm Water
Runoff)








11
17
44
10
21
94
6
8
36


91
32
86
100
140
80
95
53
130


10
30
44
13
31
89
10
10
43


96
44
95
106
170
100
125
55
160

0.9
1.8
1.0
1.3
1.5
0.9
1.7
1.3
1.2

Avg. 1.3
1.1
1.4
1.1
1.1
1.2
1.3
1.3
1.1
1.2

Avg. 1.2
TABLE 6.  VERIFICATION OF BLUE COLONIES
           FROM TWO-LAYER CULTURES
                       Colonies as
             Number of   Verified
     Source    Colonies    Fecal     Percent
            Subcultured  Coliforms  Verification
Raw Sewage
Chlorinated
Effluent
River Water
Marine Water
Reservoir
538

70
145
80
180
477

69
132
79
173
88.7

98.6
91.0
98.8
96.1
repair and subsequent reproduction of those fecal
coliform which have been debilitated by exposure
to chlorine, industrial waste or marine waters. The
decision to use the slightly more involved two-lay-
ered medium procedure in preference to the direct
M-FC method should be based on a demonstration
of increased verified recovery of fecal coliforms
from  samples routinely examined.  With  a  major
interest in the fecal coliform test being related to
the bacterial  quality assessment of effluents, the
proposed  technique should be  considered  as  an
alternative  Standard   Methods  fecal  coliform
membrane  filter  test,  specifically  intended for
those waters  known to have significant  levels of
attenuated fecal coliforms.

           ACKNOWLEDGEMENTS

     The authors wish to thank Mr. David Lentine,
Mrs.  Barbara Green,  Dr.  H.D.  Nash,  Mr. D.F.
Spino  and Mrs.  M.  Rutland  for their technical
assistance.

                REFERENCES

1.   Lin,  S.  Evaluation  of  Coliform Tests  for
     Chlorinated Secondary Effluents, Jour. Water
     Poll.  Contr.  Fed. 45, 498, 1973.
2.   Green,  R.A., R.H. Bordner, and  P.V. Scar-
     pino. Applicability of the Membrane  Filter
     and  Most  Probable  Number  Coliform Pro-
     cedures  to   Chlorinated Wastewaters. Abs.
     Ann. Meeting, Amer. Soc. Microbiol.  p.  34
     1974.
3.   McKee, J.E.,  R.T. McLaughlin, and P. Les-
     gourgues. Application  of  Molecular   Filter
     Techniques  to the Bacterial Assay of Sewage.
     III.  Effects  of Physical and Chemical  Disin-
     fection. Sew. and Ind. Wastes. 30, 245, 1958.
4.   McCarthy,  J.A.,  J.E.  Delaney,  and R.J.
     Grasso. Measuring Coliforms in Water, Water
     and Sew. Works 108, 238, 1961.
5.   Taylor, E.W., N.P. Burman, and C.W.  Oliver.
     Membrane  Filtration  Technique Applied to
     the  Routine Bacteriological Examination of
     Water.  Jour.  Inst.  Waters Engrs.  9, 248
     1955.
6.   Burman, N.P., E.W. Oliver, and J.K. Stevens.
     Membrane Filtration Techniques for the Iso-
     lation  from  Water,  of  Coli-aerogenes,
     Escherichia  coli, Faecal  Streptococci, Clos-
     tridium   perfringens,  Actinomycetes  and
     Microfungi.  127-135 in Isolation Methods  for
     Microbiologists, Technical  Series  No.  3  ed.
     Shapton, D.A. and  G.W.  Gould, Academic
     Press, p. 178, 1969.
                                             104

-------
7.   Clark, H.F.,  E.E. Geldreich,  H.L. Jeter, and
     P.W.  Kabler. The  Membrane Filter in Sani-
     tary  Bacteriology.  Pub. Health  Repts. 66,
     951, 1951.
8.   Goetz,  A. Application of Molecular  Filter
     Techniques to the  Bacterial Assay of Sewage.
     I. Proposed Technique. Sew. and Ind.  Wastes
     25,1136, 1953.
9.   Stevens, A.P., R.J. Grasso, and J.E.  Delaney.
     Measurement  of  Fecal Coliform  in  Estua-
     rine Water.   Presented at the 7th  National
     Shellfish Sanitation Workshop. New Orleans,
     La., June, 1974.
10.  American Public Health Association. Standard
     Methods for  the  Examination of Water and
     Wastewater.  13th ed. American Public  Health
     Association, Inc., New York, 1971.
11.  Shipe,  E.L.,  and A.  Fields.  Chelation as a
     Method  for  Maintaining the Coliform Index
     in  Water Samples. Pub.  Health Rpts. 71,
     974, 1956.
12.  Coles,   H.G.   Ethylenediamine  Tetra-acetic
     Acid and Sodium Thiosulphate as Protective
     Agents  for   Coliform  Organisms  in  Water
     Samples Stored for One Day at Atmospheric
     Temperature. Proc. Soc. Water Treat.  Exam.
     13,350, 1964.
13.  Hoather,  R.C.  The Effects of  Thiosulphate
     and of  Phosphate on the Bactericidal Action
     of  Copper and  Zinc  in Samples of  Water.
     Appl Bact. 20, 180, 1957.
     QUESTIONS AND ANSWER SESSION

Lin.        Are there  any  shortcomings  when
            you combine two media in one?

Geldreich:   Yes. By  putting a top layer as an en-
            richment  medium  in  that  position,
            within two  hours  if diffuses into the
            M-FC. They  diffuse together and be-
            come the same medium. At the begin-
            ning,  the membrane is separated from
            the more specific  medium, the M-FC
            medium, by  this layer of an enriched
            material.  We  used  the   procedure
            many years ago when we were trying
            to  work out an overlay concept for
            pour  plates. We  were  at one  time
            hoping that we could come up with a
            coliform pour  plate approach  which
            would be  cheap  and wouldn't  use
            membranes and other materials. That
            is the concept.  Now you  cannot lay
            the top layer on more than an hour
           in advance, because it will diffuse into
           the other medium and you will have
           the fused media and you won't have
           the advantage of an automatic transfer
           concept.

Dufour:    I  would like to know what percentage
           of  debilitated  organisms  you  are
           recovering. In  other words, what were
           you using as the best estimate, or why
           didn't you use the MPN for  instance,
           as the best estimate?

Geldreich:  We are using MPN. The material is so
           preliminary  in nature  that  I  didn't
           have time to  add  the MPN results.
           That is being done for the ASM meet-
           ing.

Dufour:    What  bothers  me is the fact  that you
           may be getting seven times  more re-
           covery but if your regular M-FC count
           is down  by two logs it really doesn't
           make much difference.

Geldreich:  Well, we will be able to give you more
           information related to a base of an
           MPN in May in the ASM.

McFeters:  Is your verification data based strictly
           upon  bacteria that were isolated that
           appeared  as fecal coliforms  from the
           medium?

Geldreich:  I would like to have done what some
           of  the  rest of you  have done but  I
           didn't have time. That is to take some
           of these organisms out of the environ-
           ment, which you  know you are having
           trouble with,  or  pure cultures which
           you  know don't  work  too  well,
           put them through the system and then
           look at it. We didn't have this time as
           of now.

McFeters:  I just wondered about the background
           of non-fecal coliforms.

Geldreich:  We had very little background on all
           samples  except those which were  in
           the reservoir.  The reservoir  samples
           occasionally   gave  us an   increased
           background of other bacteria which
           were  white colonies.  That is one of
           the difficulties when you  begin play-
           ing around with changing temperature
                                              105

-------
           and media enrichment. You are going
           to  increase the risk of getting some
           bacteria that you don't want as back-
           ground.

Ginsberg:   The instructions for performing the
           M-FC  test  state,  rather  adequately,
           that the  plates  must  be incubated
           within 10 to 20 minutes after filtra-
           tion. Now,  you  are suggesting that
           you incubate at the lower temperature
           for  two  hours. How do you  explain
           this contradiction?

Geldreich:  I  explain  this  for  the  very reason
           that we  are trying  to get an enrich-
           ment started at  35 C, not at room
           temperature. If you  lower the temper-
           ature further to room temperature, we
           are  going to  have  more and more
           problems  with background organisms.
           I would like to avoid completely, any
           use  of a  lower temperature,  but  we
           need a temperature acclimation, and I
           think what you are going to see may
           be  in the next paper. I've talked to
           Jack Delaney  and  Grasso and  they
           have come to the conclusion,  in their
           preliminary  work,   that   one  of the
           problem  is a shock of these organisms
           when coming  out of a cold environ-
           ment and suddenly put  at a jarring
           44.5 C   temperature.  Maybe  they
           need a period of acclimation.
Brezenski:   I  am a little disturbed.  If I read you
            correctly, you  said this  is a proposed
            method to substitute  the  routine
            membrane filter procedure for chlor-
            inated  effluents.  In  Figure  2,  you
            showed  four   chlorinated  effluent
            samples and I think that three out of
            the four were below the valid statisti-
            cal range. I'm happy to see something
            like this but I am afraid that I don't
            see enough data to make a  proposal
            at this time, based on this amount of
            data. Are the sewage effluents chlori-
            nated?

Geldreich:   As I told you when I was reading the
            paper,  we  recognize this as  prelimi-
            nary data  and  I myself would  not
            begin  to vote for it to be put in the
            Standards  Methods  or EPA  Methods
            until I  see not only more data from
            our  laboratories   but  from   other
            laboratories that  have  checked it out
            in  a field test. This is  only the begin-
            ning, and if these preliminary  results
            still prove to be promising as others
            check it out, then I would entertain
            that idea; but if it doesn't, let's forget
            it.

Brezenski:   During  these  sessions, we have been
            using such terms as, attenuated cells,
            injured  cells, damaged  cells and a few
            others that  I  don't remember at this
            point. I  don't know if anybody has
            done anything to describe what these
            terms  really  mean. I  get  confused
            between what is  injured and what is
            dead. Would  somebody  define  these
            terms for me so that I  can get clear in
            my mind what we are talking about.
            So  far, I  have seen no data  to  show
            any physiological  problems with the
            cells. There is no enzyme work.  No
            one has shown me a normal physio-
            logical  reaction taking place within
            the cell. I would  like to see what has
            happened to  the  cell  as  a result of
            the aquatic  environment.  If it was in
            salt water, is  some enzymatic system
            blocked and  therefore when  it comes
            into contact  with  a certain  type of
            medium does it  need  a specific sub-
            strate as a booster? I just fail to  see
            the  damage  assessment   because   I
            haven't seen the proof.

Geldreich:   Well, I agree with you. We are assum-
            ing  many things but  we have  never
            proven any of it.

Hufham:    We have seen  the damage; it results in
            different types of effects on the cell.
            You can get lack  of separations with
            chromosomes duplication,  in  which
            case  you get filaments  up to  500
            microns  long.  Unfortunately,  you
            cannot  see these as colonies because
            the width of the cell does not change.
            This is a stress factor. There have been
            several studies on this in  which they
            are trying to form filaments in cells
            and they use heat shock to do it. This
            is one of the ways of getting cultures
            synchronized  by using  heat shock and
            then bringing  them back to the appro-
            priate incubation temperature.
                                               106

-------
Brezenski:   You can see these filamentous forms    Bordner:
            which Zobel also showed with great
            hydrostatic pressures. But, as  it re-
            lates to these studies  in terms of your
            recovery data, and relating back to
            the  reasons why  the recoveries were
            low, we just say  they  were stressed
            cells. Zobel showed  that when cells
            are subjected to a certain depth at a
            certain  hydrostatic pressure, the cells
            formed filamentous branches. We have
            shown a difference in count assumed
            on the differences of a  stressed cell.
            But nobody has shown the differences
            in the stressed cell.
Geldreich:   There are various forms of doing this.
            Zobel used pressure,  various people
            used  temperature or antibiotics and
            we used  oxygen. By  increasing the    Geldreich:
            percentage of oxygen  in the culture
            you can get the same phenomenon. If
            you return those to the normal condi-    Bordner:
            tions, the cell  immediately starts to
            divide and you  get a colony formed.
            But we have not done it on solid agar.    Geldreich:
            We can see it  in  the liquid media.
            Theoretically,   if these  injured  cells
            are not kept too long and you return
            the plate  to 35 C, you  should  get a
            colony formed.  I think  in  the pre-    Bordner:
            enrichment medium you are getting
            the micro-colony. Some of these are
            capable of growing at higher tempera-
            tures.
Williams:   Getting  back to basics, one of the
           reasons we chlorinate  is to  attenuate
           the  cells. We like to  kill them and I
           guess my point is  that evidently we
           are making an effort to recover more
           and  more of these attenuated cells
           and  at what point do these attenuated
           cells no  longer have a sanitary signifi-
           cance.
Geldreich:  Well, there is some evidence. I think if
           you remember this morning, Ted, one
           of  the  speakers commented  about
           Salmonella  cultures  which had been    Geldreich:
           severely  attenuated through a freeze-
           drying   condition  and   they   were
           proven  to be very  pathogenic. What
           we really want is a kill.
Ed,  you  know  that we  have been
investigating chlorinated effluents for
sometime.  Originally, in some work
that a graduate  student did with us,
we  were  shocked  at  the very low
levels that  were  recovered by the MF
as compared to the MPN. Most of the
results did not fall within, or approach
the lower range of the very broad 95%
confidence level  of the MPN. The MF
was that much lower than the confi-
dence levels  of  the MPN. Also, we
have been  looking  at the M-FC agar
for  two or more hours at 35 C before
placing  that same  membrane  on the
same agar at 44.5 C. It looked promis-
ing  since we  knew the interest in the
overlay technique.
Are you talking about an overlay or a
two layer?

A  single layer of M-FC is what we
were reviewing.

These  membranes are sitting on top
of the  lactose  agar which is overlayer-
ed  on  M-FC agar.  I  thought you
meant agar over the membrane.

I am talking about the comparison of
two-step M-FC and the overlay  of the
lactose  agar over the M-FC agar.  In
the twenty-two samples of chlorinated
effluents  and  twenty  with  stream
samples, we see very little difference.
The  recoveries  were  comparable.  I
would   like to  call  it  temperature
acclimation with a  ratio of one to
one  in  both  cases. There were three
samples of  chlorinated effluents where
there was  a background of tiny pin-
point blue  colonies. We found them in
both  the  lactose overlay  and  the
non-overlay.  It  seemed  to me that
there is  some possibility of avoiding
the overlay. I am not sure at this point
but  I think that it  would be  worth
getting more data.

I  am  intrigued as much as many  of
you here with this concept  of a Milli-
pore membrane  that we have  heard
about today and  wonder if that will
help  too. But  I think one  of the im-
                                               107

-------
portant things that we must try and
get across  here  is  that these  are
methods that we are only discussing
today. They are preliminary and we
must have many laboratories field test
these  things before they  go into any
published procedures.  I  don't want
any of you  to  have the  illusion that
because these three labs  did it this is
the final work. This is just the begin-
ning and I hope that we evaluate these
ideas  in  many  different  geographical
areas.   Many times  in the  past  we
personally developed some media that
worked great in the Ohio area, but out
in  the  West or in New England it did
not work  as well,  because bacterial
flora are different.  So whenever we
evaluate these procedures it is of great
importance  that  we  have  as  many
people involved in  it and  as  many
different kinds of samples. We should
make our minds up from all of these
numbers whether that  method goes
into the Standard Methods or an EPA
manual. Thank you very much.
                                   108

-------
                 MEASUREMENT OF FECAL COLIFORMS IN ESTUARINE WATER

                                               by

                      Alanson P. Stevens, Rosario J. Grasso and John E. Delaney
                 ABSTRACT

     Recoveries  of  fecal coliform bacteria from
estuarine waters were compared using the standard
MF procedure and the standard MPN EC Count.
The  experimental  two-step  procedure  recovered
85%  of the MPN count as compared to 24% by the
standard  MF procedure.  Ninety-three percent of
the colonies in  the experimental procedure did
verify.

     In later work, significant variation in recover-
ies using different batches and brands of membrane
filters suggested  possible  influences on the earlier
work. Further tests are being conducted.

               INTRODUCTION

     The fecal coliform concentration  in sea water
can be measured, at the present time,  by the E C
MPN procedure,  and by a membrane filter method
using M-FC broth as the culture medium.  However,
the inherent disadvantages associated with these
procedures have prevented either one from achiev-
ing absolute acceptance by marine microbiologists.
The  multitube method requires a period of three
days to complete, and yields a concentration esti-
mate that embodies neither precision nor  accuracy.
In addition, the MPN procedure is time consuming,
requires considerable  incubator  space,  and  large
amounts  of sterilized media  and glassware. These
requirements restrict the number of samples that
can  be  processed.  The  direct-count, membrane
filter (MF) procedure, on the other hand, although
more rapid,  possessing  greater built-in  precision
and requiring a minimum of preparatory  labor and
laboratory glassware, recovers only  a fraction of
the  fecal  coliform population  in  a sea  water
sample.  A study in which  over 200 sea water
samples were analyzed by both the E C  MPN  and
standard  MF procedures indicated  that,  on  the
average, only 10%  of the fecal coliforms enumer-
ated  by the multitube technique were  measured
by the current MF procedure.
     Our  studies  have indicated  that the  low
recovery exhibited by the membrane filter pro-
cedure  is  intrinsically  associated  with  the  im-
mediate exposure  of the fecal coliforms to the
elevated selective temperature 44.5 C. The stress
imposed  by  this temperature on  the individual
fecal coliform cells, during the initial 15 minutes of
exposure,  has been  shown  to be  lethal  to  the
majority of  the fecal  strains filtered from  sea
water.  Since  the poor  MF recovery of fecal coli-
form in sea water is unquestionably related to the
immediate  exposure of these organisms  to  the
elevated  temperature,   our   initial  attempts  at
developing  a more accurate  MF   test procedure
focused on methods of acclimatizing these species
before exposing  them to the selective temperature.

Enrichment on Minimal Medium at 25 C Methods

     It was recognized at the outset, that the med-
ium  and the temperature employed during  the
acclimatization period would  have  to control the
number of bacterial generations,  so  that subse-
quent  exposure  to the  selective stage would  not
foster  colony overgrowth. A number of media
formulations  were  tested for  their  ability  to pro-
vide   optimum   resuscitation  conditions,  while
simultaneously controlling growth  for the fecal
coliforms  in  sea  water  samples.  Initial  studies
employed  a one  per  cent  (1%) solution of M-FC
broth  as the minimal  medium and full-strength
M-FC agar as the  growth-indicator medium. This
media  combination in the two-step, two-day pro-
cedure  markedly  increased   the   fecal coliform
recoveries  from   sea water, when compared to the
results from the  standard  method  one-day, direct
count  procedure.  However,  extensive testing  of
this media combination failed to yield the desired
recovery efficiency to  warrant its  acceptance as
a standard test system.
     Recent work has focused on formulating and
testing a  minimal medium  composed  mainly of
                                               109

-------
simple carbohydrates. It was theorized  that these
substrates would  be easily attacked and metabol-
ized by  the  fecal  coliforms  and would foster
bacterial  cell  repair and resuscitation for those that
had  been attenuated by exposure to sea water.
This minimal  medium, coupled with  full-strength
M-FC broth  as the  selective-growth  medium,  has
yielded very  satisfactory recovery levels for fecal
coliform  in sea water.

     The  formulation  of  the  minimal  medium,
labeled   L.E.S.   Minimal  Holding  Agar,  is  as
follows:

       L.E.S. MINIMAL  HOLDING AGAR
     Tryptose

     Dextrose

     Lactose

     Oxgall

     Sodium Chloride

     Agar
 0.5 grams/I

 0.5 grams/I

 0.5 grams/I

0.25 grams/I

 0.4 grams/I

15.0 grams/I
L.E.S. Two-step, Two-day Procedure

     The minimal holding agar is allowed to warm
to  room  temperature  before  "prepared"  mem-
brane filters are  placed  on the agar surface. All
standard precautions are exercised at this stage to
guard against improper placement of the filter on
the agar surface.  The dishes are  tightly sealed and
placed into a watertight, plastic bag, making sure
that all plates are facing upright. The  sealed bags
are inverted and  incubated in a  water bath or air-
jacketed incubator  regulated at 25 C  for 18 ± 2
hours. Subsequent to  the enrichment  period, the
membranes are transferred  to  absorbent pads in
tight  sealing petri dishes that have been  saturated
with  Standard Methods  M-FC broth.  Approxi-
mately  1.7 to 1.8 ml of broth are required to
saturate an absorbent pad.  Excess media should be
discarded  to  prevent  excessive and  "running"
bacterial growth. The  dishes are tightly  sealed,
placed into water-tight plastic bags and incubated
in an inverted position in  a  water bath at 44.5 ±
0.2 C for 24 hours.

Counting of Fecal Coliforms

     The fecal coliform colonies are counted with
the aid of a stereomicroscope and a light source
                  above, that is approximately perpendicular to the
                  plane of the membrane being counted. The  fecal
                  coliform colonies are recognized  by their  blue
                  coloration  and the  crystallized deposits on the
                  surface. Both of these identifying characteristics
                  must be employed  in counting the fecal coliform
                  colonies. The  "quartz"  surface  appearance has
                  been  found to be  a complimentary  and distinc-
                  tive   characteristic   of  fecal  coliform  colonies.
                  Certain non-fecal coliforms capable of  growing
                  under the  test conditions will form  blue-colored
                  colonies but will  lack the  distinguishing crystalline
                  characteristic on their surface.
                   RESULTS

Fecal Coliform Recovery — MF vs MPN Procedures

     Twenty-five  samples  of sea water containing
varying concentrations  of  fecal  coliform and  salt
content were examined  to  determine the recovery
efficiency of the L.E.S.  two-step, two-day pro-
cedure. From each sample, five 10 tube E C MPN
analyses were  made and  ten  membrane  filters
(Millipore HAWG 047SO)  were prepared and pro-
cessed  by  both  the Standard  Methods  MF pro-
cedure  and  the  experimental  MF technique. The
results, obtained from each of the MF procedures,
were  logarithmically  averaged  and proportioned
against the average logarithmic MPN result of each
sample. Table I presents the percent of recoveries
obtained  by the two  MF procedures  for each
sample analyzed, and  employ  the   logarithmic
average E C  MPN result as the true estimate. These
data indicate that the two-step, two-day procedure
yields  a significant recovery increase  over that
attainable by  the Standard  Methods  MF proce-
dure.  The Standard Methods  MF procedure pro-
duced  an average recovery of only 24% while the
L.E.S.  two-day technique yielded an 86% recovery
of  the fecal coliform  concentration  in  the  sea
waters examined. It is worthwhile noting that the
lowest  recovery  obtained by the experimental
method (64%) is  higher than the highest recovery
obtained by the Standard Methods MF procedure
(46%). Table I also  presents data on the selectivity
incorporated into the two-step,  two-day MF pro-
cedure. Over 93% of the  colonies exhibiting  the
two identifying characteristics of fecal  coliform
verified as bona fide fecal coliform strains. Overall,
the percent verification in the 22 samples ranged
from 87 to 100.

     Figure  I presents a graphical analysis of  the
recovery data obtained by  the E C  MPN procedure
                                               110

-------
TABLE 1.   FECAL COLIFORM RECOVERIES
            BY MEMBRANE FILTER
            PROCEDURE
      Percent Recoveries (*)
   2.0
   1.0
Standard Methods
Procedure (1)
17
39
34
16
11
12
6
9
34
5
46
40
27
22
31
37
22
30
16
15
21
16
32
29
24
L.E.S.
Procedure (1)
126
114
96
64
89
97
75
76
77
80
71
71
89
72
79
92
97
78
87
66
82
84
106
117
90
/o
Verification of
L.E.S. Colonies
92
95
88
93
96
95
100
90
95
91
87
100
95
91
92
92
—
96
94
96
91
87
100
90
100
(*) E C  MPN ten tube procedure used as Control.
(1) % Recoveries based on 5 MPNs and 10 MFs
   per sample.


and  the two membrane filter methods. The ratios
(Table I) were arrayed in order of magnitude and
every other value plotted against the appropriate
percentile on log-probability  paper. The  resulting
lines of best fit clearly indicate  that a significant
difference exists between the recovery attainable
by the Standard Methods MF procedure and by the
experimental MF method. Based on  the 50 per-
centile (median) ratios, the average recovery of the
Standard Methods MF procedure was 22%, while
the experimental MF procedure recovered an aver-
age of 84%  of the fecal coliform concentration in
the sea  water samples. This analysis indicates that
the experimental method, on  the average, recovers
2.9 times more fecal coliforms from sea water than
                                                     0.1
CURRENT L.E.S.
 PROCEDURE
                                                            STANDARD
                                                            METHODS
                                                             PROCEDURE
               1      10      50    80   95  99

        PERCENT EQUAL TO OR LESS THAN

                   Figure 1.

the Standard Methods MF procedure. A further
analysis of the plots  in  Figure I  clearly demon-
strates the improvement in fecal coliform recovery
as a result of the two-step, two-day procedure. The
plots  indicate that the Standard Methods MF pro-
cedure recovered  less  than  19% of the fecal coli-
form  in 40% of the samples and less than 40% in
90% of the samples, while the experimental pro-
cedure recovered  at least 60% in  all samples and
less than 72% in only 20% of the samples analyzed.

                  SUMMARY

    The direct count procedure for fecal coliform
in estuarine waters, developed under this research
project  1, appears to  possess  the  necessary accur-
acy,  precision  and selectivity to  warrant its ac-
ceptance by marine microbiologists as a more than
adequate  replacement for the  E C  MPN  pro-
cedure.  The  recovery  efficiencies of the L.E.S.
experimental  MF  procedure were  calculated  from
MPN values that  had  not been corrected for their
inherent positive  bias. If this bias  factor had been
used in the computation of the recovery ratios, the
median  recovery  percentage (84%) of the experi-
mental procedure  would consequently have  been
adjusted to  a  more favorable level, namely  92%.
This truer recovery capability of  the  L.E.S. two-
step, two-day  MF procedure, when coupled with
the MF advantages of precision, time of analysis,
less preparatory  labor and  incubator  space, are
strong arguments  for accepting and employing the
experimental  MF technique as a  standard proce-
dure for measuring fecal coliform concentrations
in sea water.

    This research project was supported, in part,
by a Research Grant from the Massachusetts Water
Resources  Commission.   We  acknowledge  this
support and their active interest in  this project.
                                              111

-------
Recent Investigations
                       Gelman
  Millipore
     Recent studies  have indicated that different
brands of membrane filters  have  an effect on the
recovery levels of fecal coliform organisms (1, 2).
Since all  the data generated  in  developing the
L.E.S. two-step, two-day procedure for fecal coli-
forms was compiled  using the Millipore Type HA
filter, we deemed it necessary to compare the fecal
coliform recoveries attainable by using  two brands
of membrane filters (Gelman GN-6, lot No. 80706
and  Millipore HAWG 047SO, lot No. 934487)  in
our newly developed method.

     This work was not  motivated by the desire to
prove one membrane  filter  brand superior over
another, but rather by a strong curiosity as to why
different reputable brands of filters should produce
statistically significant recoveries of fecal coliform.
The  results  of our  comparative studies  showed
that Millipore  filters consistently gave lower re-
coveries than the Gelman filters  in the Standard
Methods MF  procedure for fecal  coliform. The
Gelman  filters,  conversely, gave poor  recovery
compared to  the Millipore filters when  used  in the
two-step,  two-day L.E.S. procedure. In addition,
the Gelman filters produced non-typical fecal coli-
form colonies  subsequent  to  incubation.  These
colonies were  small  with  irregular shapes, and
varied  in  color from light brown to  the typical
fecal coliform blue.  It should also be  noted, that
even with the higher recoveries of fecal coliform
by the  Gelman filters  in the  Standard  Methods
fecal coliform procedure, only  30% of the actual
concentrations of  fecal  coliform in the  samples,
as judged by the E C  MPN method,were recovered.

     From these results, it seemed plausible, that
any  substances  present  in  the membrane  filters
capable of affecting development of fecal coliform
colonies during the  incubation  period would  be
soluble  in the media,  producing pH  and  other
physico-chemical  changes.  In order to determine
the  water soluble chemicals in  these two filter
brands,  six filters of each were immersed in 100 ml
of distilled water for 18 hours at room tempera-
ture, the  filters were removed  and the following
results were obtained upon analysis of the water.

     It  was thus possible to project that  the dif-
ferences  in  quantities   of soluble components
                        GN-6
                     Lot #80706

pH                     3.23

Ammonia-Nitrogen      13.0 mg/l
Ortho-Phosphate (as P)   43 mg/l
Total dissolved solids     185 mg/l
   HAWG
Lot #93448-7

    6.20

    0.24 mg/l
    0.06 mg/l
      6 mg/l
between  brands  of  filters  could somehow be
responsible for the differences that were obtained
in fecal coliform  recoveries. Neither the Standard
Methods M-FC broth, nor the two media used in
the L.E.S.  two-step, two-day procedure have an
inorganic  buffer system  in  their  formulation.

     Low recoveries  exhibited  by  the  Gelman
filters  with  the  new L.E.S.  procedure could be
attributed to  the increased hydrogen-ion  concen-
tration that these filters imposed on the holding-
minimal  agar,  while the  higher  recoveries by
Gelman filters on standard M-FC broth at 44.5 C
could be attributed  to  the leaching of beneficial
inorganic nitrogen and phosphate into the medium,
as well  as  the suppressing of the pH change by
buffering action of the organics in the formulation.
Based on this hypotheses, we added a monopotas-
sium phosphate and dipotassium phosphate buffer
system to the  L.E.S. minimal  holding agar and also
to the M-FC broth. Results obtained from experi-
ments  employing this buffered  media, produced
encouraging results with  both  brands  of filters.
At present, additional tests are being conducted on
estuarine waters using the new buffered media with
the two-step, two-day L.E.S. procedure in order to
substantiate that either brand  of  filters can be
employed in this technique.


                 REFERENCES

1.   Presswood, W.B., and L.R. Brown. Compari-
     son  of  Gelman   and   Millipore   membrane
     filters for enumerating fecal coliform bacteria.
     Appl. Microbiol, 26:336, 1973.
2.   Hufham,   J.H.,  Evaluating  the   membrane
     fecal  coliform  test  by  using   Escherichia
     coli as the  indicator organism. Appl. Micro-
     biol, 27:771, 1974.
                                               112

-------
                  EVALUATION OF METHODS FOR DETECTING COLIFORMS
                                             AND
                    FECAL STREPTOCOCCI IN CHLORINATED SECONDARY
                                     SEWAGE EFFLUENTS

                                              by

                                          Shundar Lin
                                      Water Quality Section
                                    Illinois State Water Survey
                                          P.O. Box 717
                                      Peoria, Illinois61601
                 ABSTRACT

     Total  coliforms  (TC),  fecal coliforms  (FC)
and  fecal streptococci (FS) recoveries in chlori-
nated secondary sewage effluents were investigated
using the membrane filter (MF) and multiple-tube
(Most probable number, MPN) methods. The LES
two-step MF method was found to be comparable
to the MPN procedure for determining TC. The TC
detection was  1.5 times greater when using the
LES two-step technique than that obtained by the
M-Endo  one-step MF  procedure.  Fecal  coliform
recovery by the M-FC MF  procedure was lower
than the recovery obtained using the MPN  method.
The  use of each of  azide-detrose broth,  brain-
heart infusion  broth,  and peptone  yeast-extract
Casitone  with   the  M-Enterococcus agar  MF2
(2-day incubation) procedure was not satisfactory
for the  recovery of FS. The M-Enterococcus agar
procedure with  bile broth enrichment (MF2>  or
prolonged incubation for 3 days (MFs) signifi-
cantly increased FS  recovery and were comparable
to the MPN method. The results cited should  be
useful  in assessing the efficiency  of  disinfection
practices for waste  treatment  plants  employing
effluent chlorination.

               INTRODUCTION

     The year-round  disinfection  of wastewater
treatment plant effluents has become mandatory
in Illinois and  in several  other  states. The most
common method  of  disinfection  at treatment
plants is chlorination.  Its effectiveness has gener-
ally  been  measured  by  residual  chlorine. The
Illinois  Pollution Control  Board  (1)  requires  a
limitation on fecal coliforms (FC) densities inde-
pendent of residual chlorine thus requiring deter-
minations for FC densities in chlorinated effluents.
The  Board's  rules stipulate that fecal  coliforms
densities  in a  waste  effluent shall  not  exceed
400/100 ml.

     Total coliforms  (TC)  have been used for mea-
suring  the disinfective efficiencies of water and
wastewater treatment units.  The TC  index  is still
valid and  reliable for the water industry. In Euro-
pean countries fecal  streptococci  (FS)  are  com-
monly  looked for in the sanitary analysis of water
supplies (2).  In  the United States, they are used
currently  in conjunction with  FC for  determining
the sanitary quality of water.  Although FS deter-
minations are not required by  most regulatory
agencies  the usefulness of the procedure  should
not be overlooked.

     The requirement for  bacteria enumeration in
treated effluents necessitates the development of
adequate and economical procedures for determin-
ing bacteria densities in chlorinated effluents. The
series of investigations described in this report were
undertaken with these objections in mind.
     Indicator Organisms. The purpose of the rou-
tine bacteriological examination of water samples
is usually to estimate the hazard due to fecal pollu-
tion and the probability of the presence of patho-
genic organisms. The  isolation of pathogens from
water and  sewage is  expensive  and laborious.  It
is  not  a  routine  practice.  Normally  occurring
bacteria in  the intestines of warm-blooded animals
have  been  used as indicators of fecal  pollution.
Total coliforms, fecal coliforms, and fecal strep-
                                              113

-------
tococci  have  all  been used as pollution  indicators
at various times (3, 4). Other bacterial  indicators
have been  proposed.  These  include Clostridium,
Pseudomonas  and  Aerobacter.  Presently  their
value has been considered questionable or irrele-
vant (5).

     Correlations between coliforms and  patho-
genic bacteria  have  been cited frequently i.e.,
coliforms vs Salmonella (6, 7, 8, 9). Less known is
the relationship,  if any  exist,  between coliforms
and  viruses.  A  coliform  index is  not  a  reliable
index for viruses (10,  11). In spite of the lack of
documented  relationships there is  little evidence
that enteroviral  or  other microbial  diseases are
transmitted  frequently  by  the drinking  water
route in the absence of coliforms (5).


     Until  more  definitive studies  are completed
on the relationship  of  pathogens  and indicator
organisms, the use of TC for water supplies and FC
and  FS for sewage and stream quality, as indicators
or enteric pollution, are valid.


     Bacteria Enumeration. The basic methods for
the assay of pollution indicators (TC, FC, and FS)
in waters are outlined  in Standard Methods (4).
These include the multiple-tube, or  most probable
number  (MPN)  technique   and  the  membrane
filter (MF)  procedure.  Standard  Methods (4),
however, states that "Experience indicates that the
MF  procedure  is applicable to the examination
of saline waters but not chlorinated wastewaters".
Because the MF technique is not comparable to the
MPN  procedure  and  is  less time  consuming, it
seems unfortunate that the  MF technique cannot
be used as a control procedure  by the waste plant
operator who uses chlorination.


     McKee et  al. (12)  reported on the  lack of
correlation  between  MPN   and  MF  techniques
while assaying  chlorinated settled wastewater for
total  coliforms.  Because monochloramine is the
predominant   bactericidal  agent  in chlorinated
wastes,  they  advanced the hypotheses that partial
reversibility  is  responsible   for  the discrepancy
between MPN  and  MF  results;  that is, the  MF
technique  produces  considerably fewer colonies
than the numbers that develop by the MPN method.
Presumably,  when inactivated  cells  are deposited
on a membrane with limited nutrient availability,
the cells cannot rid themselves of monochloramine
and therefore cannot grow. However, when inacti-
vated cells are  put in an aqueous medium rich in
organic matter,  such  as  lactose broth, the mono-
chloramine may diffuse outwardly from the cells,
permitting them to recover, grow, and produce gas.

     In  the  McKee et al. (12)  investigations, the
culture  media used  for the MF  technique was the
same as that previously described (13). Dehydrated
scheduled nutrient  (DSN)  pads were used. They
contained two elements with an upper leaf impreg-
nated with an Endo-type inhibitory nutrient. The
results obtained  using DSN  pads with the MF tech-
nique were comparable to those obtained from the
confirmed MPN  procedures on  raw settled  waste-
water.   McCarthy  et al.  (14), though  working
initially with  water, were  not  satisfied with the
one-step,  M-Endo  broth  MF  techniques.   Their
work suggested  that  enrichment  plus  an agar
substrate  (E&A) was  superior  to the  one-step
technique based on a higher degree  of coliform
recovery.  Examinations  of  natural  waters and
wastewater  demonstrated  that  the  E&A  results
were comparable to  standard  MPN  data.   From
their work an agar-based  medium  (LES M-Endo
agar) was developed. Its use with the MF technique
is  basically  a  two-step enrichment  procedure.

     The need has  developed  not only for deter-
mining  total  coliform  but also for  enumerating
fecal coliform  densities.  Geldreich  et  al. (15)
recommended the  use  of  an  M-FC  medium  at
incubation temperature of 44.5 ± 0.5  C as part  of
the  MF  technique for the  direct county of fecal
coliform. It has been reported (16, 17, 18) that the
determinations for fecal coliform rather than total
coliform are  a more realistic measurement  of the
public  health significance of microbial discharges
in wastewater plants. Illinois requirements specify
maximum  permissible limits  for  fecal coliform
concentrations  in  treated effluents. This will
require  fecal  coliform enumeration  in chlorinated
effluents.

     Several  investigators  (19,  20, 21)  reported
that Escherichia coli injured during a physical  or
chemical  treatment failed to  form  colonies  on
membrane filters incubated on  M-FC medium,  to
grow and produce gas in lactose broth or to grow
on selected media. Braswell and  Hoadley  (21) sug-
gested that standard methods for enumeration  of
total and fecal coliforms in water and wastewater
should  not  be  applied  to  chlorinated effluents.


     Even for unchlorinated samples Hufham (22)
claimed that a large relative error in the results of
MF  method  was found  to be  dependent  on the
brand of MF used,  the medium, and the tempera-
ture of  incubation. A  study  by  Presswood and
                                               114

-------
Brown  (23)  showed  FC counts  incubated  on
Gelman filters at 44.5 C averaged 2.3 times greater
than those on Millipore filters. Hufham  (22) sug-
gested that the MF method for FC recovery should
not be accepted.

     Lattanzi and Mood (24) used the Winter and
Sandholzer method for the detection of entero-
cocci. Later Litsky et al. (25) suggested the use of
glucose azide broth as a presumptive medium and
ethyl  violet  azide (EVA) broth as a confirmatory
medium  for  enterococci  detection  with  MPN
procedures.

    Slanetz and Bartley (26) proposed the use of
M-Enterococcus agar for the isolation of FS by the
MF method. Kenner  et al. (27) introduced the KF
streptococcus  agar.  Rose  and  Litsky  (28)  found
they could increase the recovery of FS from river
water by  more than 2-fold when using  peptone
yeast-extract  Casitone  (PYC) compared  to  M-
Enterococcus agar.  Recently Pavlova et  al. (29)
suggested  that fluorescent  antibody techniques
may be useful, for FS detection, in determining the
presence and source of fecal pollution  in water.


Objectives

     During the course of the study two separate
investigations  were performed. One  dealt princi-
pally with TC and FC; the other with  FS. The pur-
poses of the study were:

     1.   To  determine  whether or  not the  MF
          technique for TC, FC and FS detections
          in  chlorinated  secondary  effluents  is
          comparable to the MPN method.

     2.   To determine whether or not the LES
          two-step enrichment  MF technique for
          TC  detection,  in chlorinated secondary
          effluents, is  comparable to recommend-
          ed MPN methodology.

     3.   To  develop  improvements in the  MF
          method for the  detection of FS in chlor-
          inated secondary sewage effluents.

          MATERIALS AND METHODS

     Grab samples of final settling tank  effluents
from  three wastewater treatment plants serving the
cities of Peoria,  Morton, and Washington in Illinois
were used in the study. A minimum of five effluent
samples  from  each plant  were examined. The
Peoria plant employs the high-rate activated sludge
process treating a combination  of domestic and
industrial  wastewaters. Contact stabilization com-
parable to the  standard-rate activated sludge pro-
cess is used at Morton. This plant treats principally
domestic  wastewater. Washington  is served by a
standard-rate trickling filter plant, treating domes-
tic wastewater also.

     One  liter portions of each effluent were dosed
with  calcium  hypochlorite  (HTH,  70%  available
chlorine)  up through  6 mg/1  of  chlorine.  The
samples were stirred  gently but intermittently, and
after varying periods of contact (up to 30 min.)
they were dechlorinated with an excess of sodium
thiosulfate.   The  dechlorinated  samples  were
assayed immediately for bacterial  densities using
parallel MPN and MF methods.

     The  MPN  procedures  were  performed  by
inoculating a series of four decimal dilutions per
sample, using five tubes for each dilution.  Lauryl
tryptose  (LT)  broth was  used for the presump-
tive  tests  in  TC and FC determinations. The TC
test was confirmed using brilliant green bile (BGB)
medium, and was completed with gram-strain. For
FC  confirmation EC medium at  44.5 ±  0.5  C
(water bath)  was used.  In the MPN procedure for
FS  tests, axide-dextrose (AD)  broth was used for
the  presumptive  test;  while  ethyl  violet  azide
broth was used for confirmation.

     In the MF  procedures for TC, FC, FS, three
duplications for each sample were filtered through
an 0.45 //m membrane filter (Millipore)  for each
bacterial  test. For TC tests,  the two-step enrich-
ment for   LES M-Endo agar  (14) was  followed.
Occasionally, parallel tests with  the standard one-
step M-Endo procedure were  performed. For TC
verification purposes, representative colonies (3 to
6  sheen  colonies  per filter)  were  subcultured
through LT broth  into  BGB broth (30). The pro-
duction of gas  on  BGB broth was deemed verifi-
cation.

     When using MF  procedures for FC detections,
the  recommendations  of  Geldreich  et  al. (16)
were  followed.  Several  colonies  (3  to  5 blue
colonies per  filter)  grown on the M-FC medium
were verified by inoculating in phenol red lactose
broth for  a 24 to  48 hour  period at 35  C and
noting gas production. All positive tubes were con-
firmed at 44.5 C (water bath) in EC broth.

     In the determination of FS densities by the
MF  technique, the  standard  one-step M-Entero-
coccus agar (4) was  used. According to Sies (31)
                                               115

-------
M-Enterococcus agar is superior to the KF strep-
tococcus  agar   for   sewage   effluents  because
some  of the non-streptococci  species in sewage
samples grow red and  pink colonies on KF Strep-
tococcus agar. The FS counts on the membrane
filters were  generally  made after  2, 3, 4, and 7
days incubation. Parallel  tests  with the  two-step
enrichment were also  performed. The enrichment
media used include AD broth, brain-heart infusion
(BHI)  broth, bile broth medium (prepared  by
adding 40 ml sterile 10% oxgall solution to 60 ml
sterile BHI broth), and PYC broth. The period of
the pre-enrichment was 2-3 hours. For the purpose
of FS verification, red and pink colonies (3 to 6
colonies per filter) were fished at random from the
membrane filter and inoculated onto a brain-heart
infusion agar (BHIA) slant, followed by a catalase
test. If the catalase test was  negative, then the
growth  on the  BHI A  slant was sub-cultured into
                      both  a BHI broth and into a bile  broth medium
                      for confirmation.

                          With slight variation all  bacteria assay  pro-
                      cedures followed Standard Methods (4). Generally
                      all the media used were freshly prepared; none of
                      the media used was more than four days old.

                               RESULTS AND DISCUSSION

                      Total Coliforms
                          Multiple-tube versus membrane filter. Consis-
                      tent with Standard Methods (4)  recommendations
                      that a comparison be made between MPN and MF
                      techniques before using the MF procedure, a series
                      of bacterial  assays on  unchlorinated samples from
                      a variety of sources was performed. This evaluation
                      included enumeration for total coliforms as well as
                      fecal  coliforms. Table  1 summarizes the results.
 TABLE 1.   COMPARISON OF THE MPN AND THE MF COLIFORM DENSITIES OF
            UNCHLORINATED WATERS FROM SEVERAL SOURCES
    Source*
 Spoon River
 Spoon River
 Spoon River
 Spoon River

 A. S. effluent
 A. S. effluent
 A. S. effluent
 A. S. effluent
 A. S. effluent
 A. S. effluent
 A. S. effluent
 A. S. effluent
 A. S. effluent
 A. S. effluent
 T. F. effluent
 T. F. effluent
 T. F. effluent

 Tert. pond effluent
      Total coliforms/100 ml
                            Completed MPN
                       LES-MF
     2,400
     3,300
       460
     1,300

92,000,000
13,000,000
 7,900,000
 5,400,000

   790,000
   350,000
   240,000
 3,500,000
     2,700
     3,100
     1,200
     2,600

80,000,000
12,000,000
 7,400,000
 6,700,000

 1,300,000
   670,000
   360,000
 4,000,000
                           Fecal coliforms/100 ml
                             MPN
                    MF
Illinois River
Illinois River
Illinois River
Illinois River
1,700
1,300
790
—
1,400
2,000
1,200
—
230
490
170
79
640
330
270
160
       490
       170
       140
       790
 4,900,000
 2,400,000
   460,000
    33,000
    79,000
    79,000
    49,000
    33,000

   490,000
   490,000

   790,000
      330
      270
      250
      790
35,000,000    19,000,000
1,000,000
2,100,000
  400,000
   30,000
   52,000
   80,000
   59,000
   60,000

  600,000
  410,000

  600,000
 *A. S.  =  Activated sludge process
 *T. F.  =  Trickling filter process
                                              116

-------
Using the paired data t-test  technique in testing
the hypothesis  (H0)  that  the  mean  of  the  first
population is equal to the mean of the second; the
results (Tests 1 and 3 of Table  2) do  not indicate
significant  differences  in  TC and  FC recoveries
determined  by  the MPN  and  MP  methods.  The
comparison  for the purposes  of this study, there-
fore, were considered acceptable.

     M-Endo (one-step) versus  LES M-Endo (two-
step).  Samples  of chlorinated  effluents  from an
activated  sludge process  were  evaluated  for TC
densities using  M-Endo one-step and LES M-Endo
agar two-step MF procedures. McCarthy et al. (14)
performed a similar assessment on unchlorinated
water samples from rivers, lakes, and ponds leading
to the development of the LES agar-based medium.
The results  obtained  on  chlorinated secondary
effluents were  comparable to  those  observed by
McCarthy et al. (14). As shown in Figure 1, the
plotted data lie above  the equality  line, indicating
that total coliform recovery  by the LES two-step
procedure  was  superior to the M-Endo one-step
method. A  better development of  sheen colonies
was also observed on the LES  medium.
  5 240
  o
  o
  £ 200
     160
  o
  Q
     120
  CO
  CO
      80
      40
               n=34
               Y = 0.968
CHLORINATED ACTIVATED
SLUDGE EFFLUENTS
             40    80   120   160  200   240
            ONE-STEP M-ENDO MF COUNT

  Figure 1. Comparison of Total Coliform Counts
           made on M-Endo Broth and on LES
           M-Endo Agar.


     The experiences  of McCarthy et al.  (14) and
McKee et al. (12) were similar with regard to total
coliform recovery  from unchlorinated wastewater
samples. The McCarthy group  found no advantage
in using an enrichment phase when compared with
a one-step agar  method  on unchlorinated  wastes
and  polluted  waters.  They suggested that the
recovery efficiency for total coliform&was a func-
tion  of the number of coliforms in the sample and,
therefore, in natural waters where smaller numbers
of coliforms are likely  to exist, an enrichment
phase in the MF technique is required, whereas
with polluted waters and unchlorinated wastewater
the enrichment two-step procedure can be omitted
without  significant effect  on coliform recovery.
McKee and  his  colleagues, however,  experienced
the lessening of coliform recovery on chlorinated
settled wastewater similar  to that described for
water with a  smaller number of coliforms. This
suggests that equivalent conditions are encountered
when temporarily inactivated  colonies exist or a
smaller  number  of colonies are present. In both
cases, an enrichment  phase  would more than likely
be required to attain satisfactory coliform recovery
using the MF technique.

     Although the number of colonies per filter
as depicted  in  Figure 1  exceeded the desirable
range of  20 to  80/filter,  they were considered
satisfactory for  comparison purposes. The  results
correlated well  (r =  0.968), and the  relationship
between  the two procedures can be expressed as

             TC2 = 0.64+  1.56TC.,

where  TC-j  and TC2 are,  respectively, the total
coliform  colonies determined by the one-step and
two-steo MF  techniques. The  total coliform  re-
covery on chlorinated effluents by  the  LES two-
step  procedure is about 1.5 times greater than that
attained by the M-Endo one-step method.

     LES (two-step)  versus multiple-tube.  The
multiple-tube method is considered  acceptable for
assaying the total coliform  densities in chlorinated
wastewater effluents.  A  comparison of the total
coliform  data resulting from  the  LES  two-step
method,  which was used in  this study, with  bacter-
ial densities obtained from parallel multiple-tube
observation was therefore  pertinent.  Using meth-
ods described  by Thomas (32)  the total coliform
data for all  chlorinated secondary effluents ex-
amined  are  shown in  Figures 2  and  3.  Figure 2
represents observations of  the  LES two-step MF
technique, and Figure 3 represents observations of
the  multiple-tube procedure.  The  figures  reflect
simply the geometric distribution of the bacterial
densities for all  effluents  using  two techniques.
Similar curves could  have been presented for each
type of effluent.
                                               117

-------
CO

O
I-
cc
LU
CO
00
o
Q
LU
CC
LU
LU
LU
00

LU
LU
DC
LU
LL
LU
O
O
CO

DC
O
CO
LU
LU
IE
I-
LL
O

CO
CO
LU
QC
CN
LU
_J
OQ

I-
                 CNI

                 II

                 a.
                 "o
                IE
                *
                H-
                T3

                LO"
                 CD
                CJ
              i_  iL  C
              QJ  QJ    ».
             -Q  y,  ^
              |°-B
             z*S  5
                    "8
                 « cS
                00 +j
 in
T3
 O
 +-*
 0)
 E
"D
 2?
 CO
 a
 E
 o
O
                    CD


                 tt E
                 CP 3
                 I- C
                             CD
                                     CO
                             oq
                             o
                                     ro
                                     o
                             _a>

                              CD
                             O
                                     O
                                  •a
                                  (D
                                            CN
                         O
                         LL.
                                                    O5
                                            CN     «-
                                            r-     LO

                                             O)      CD
                                                    CO
                                                            in*-
00


LO

CO
                                                                  cxD
                                                      cqr--
                                                      o     ^
                                                            r^
                                                            CM"
cxj
o
                                                                                              CO
                                                              >
                                                            COCO
                                                                          COCOCOCO
                                                                o
o

3
                          LLc/)
                                          o
                                          C
                                                  C\
                                                         o
                                                         c
                                                          CN
                                      CLL.(nLLv)LL.
                                                                CM
                                                                        CMC  CNC  CNc   CO
                                                                                                     a>
                                                                                                     o
                                                                         •D
                                                                          L_
                                                                          o
                                                                                                     cu
                                                                                                     a>
                                                                                                     o

                                                                                                     0}
                                                                                                    T3
                                                                                                    0)
                                                                                                    ~   'o
                                                                                                    "S   &
                                                                                 3   CD

                                                                                 3   «
                                                     118

-------
   106
 o
 o
   105
 CO
 uu
 Q
   EXPLANATION
   HIGH RATE ACTIVATED
I- SLUDGE
  ^CONTACT
   STABILIZATION
 _°TRICKLING FILTER  / Mg x dg,
   n=71           f   350,000
 P104
 O
 o
   103
Mg-Og

  2400
                 Mg= 29,000
                    .12.14
   102
     0.5       10       50       90     99.5
   PERCENT PROBABILITY OF NOT EXCEEDING
                                        o
                                        o
                                          106
                                                    105
                                        o
                                        0.
                                                   10*
                                                 O
                                                 o
                                          103
                                          102
                                                         •EXPLANATION
                                                          HIGH-RATE ACTIVATED
                                                        "^SLUDGE
                                                          CONTACT
                                                          STABILIZATION
                                                         oTRICKLING FILTER
Mg X dg

370,000
                                                                   Mg-28,000
                                                                   d  = 13.55
                                                    Mg-dg

                                                    2100
                                            0.5       10       50       90     99.5
                                           PERCENT PROBABILITY OF NOT EXCEEDING
Figure 2. Total Coliform Analysis on LES Two-Step    Figure 3.  Total Coliform most probable number
        Membrane Filter Samples from Chlorinated
        Effluents.
                                               Ana|ysis
                                                                         ch|orinated Eff|uents
     More important for comparative purposes is
the summary included  in Table 3. For the MF
technique, including all  data, the geometric mean
was 29,000 total coliforms/100 ml; the geometric
standard deviation was 12.14; and  the arithmetic
mean computed  from geometric parameters (32)
was  650,000/100 ml. Similarly,  the  MPN data
reflected  a  geometric mean  of 28,000 conforms/
100  ml, a geometric standard deviation of 13.55;
and an arithmetic mean of 700,000/100 ml. All of
the data, including  that for  each type of effluent
summarized in Table 2 suggest that the LES two-
step  MF method for chlorinated effluents is com-
parable  in  coliform recovery efficiency to the
multiple-tube procedure.

     Figure 4 is a graphical presentation for com-
parative purposes also.  For  the 71  examples ex-
amined, 32 of the  MF  results are  higher and  34
of the MF results are lower than concurrent MPN
results. Five observations were found to be identi-
cal. The ratios of MF:MPN varied from 0.44 to
5.03 with a median of 1.00.
                                        o
                                        o
                                        CO

                                        LU
                                        Q
                                         ;104
                                        o
                                        o
                                        2102I2
                                                        EXPLANATION
                                                        HIGH-RATE ACTIVATED SLUDGE!
                                                        CONTACT STABILIZATION
                                                        TRICKLING FILTER
                                                        n = 71
                                            102
                                                                   10*
                                                                           106
                                         Figure 4. LES Two-Step Membrane Filter and
                                                 complete most probable number
                                                 results on Chlorinated Effluents.
                                            119

-------
TABLE 3.   COMPARATIVE RESULTS FOR TOTAL COLIFORM DATA
Chlorinated
Effluent
High rate
Act. si.

Contact
Stabilization

Trickling Filter

Total

No. of
Observations

21


24

26

71

Test*

MF
MPN

MF
MPN
MF
MPN
MF
MPN
Geometric
Mean, Per
100ml

40,000
36,000

5,100
4,600
110,000
120,000
29,000
28,000
Geometric
Standard
Deviation

18.00
20.43

5.16
5.20
6.31
7.15
12.14
13.55
Arithmetic
mean t, Per
100ml

2,600,000
2,900,000

190,000
150,000
6,000,000
7,000,000
650,000
700,000
*LES (two-step) MF and completed MPN

tM - C Mg ag 1.15 log ffg. C = 1.0 for MF, C = 0.851 for 5-tube MPN
     As  shown  in test  2 of Table 2  the  TC re-
covered  by the MPN and LES M-Endo methods
are not significantly different. It is concluded that
the LES two-step MF technique was as good as the
multiple-tube  method for assaying total coliforms
densities in chlorinated  secondary effluents. From
the standpoint of time, convenience, freedom from
bias,  and  equipment  needs, the LES  two-step
technique would seem preferable to the multiple-
tube technique for chlorinated effluents.


     LES  (two-step)  and multiple-tube verifica-
tions.  Occasionally, coliform bacteria may  fail to
reproduce colonies on membrane filters  and  non-
coliform organisms may  develop sheen  colonies.
Verification   procedures  were  undertaken   for
coliform organisms on all membranes and multiple-
tube samples  in accordance  with procedures des-
cribed  by  Geldreich  et al.  (30). Calculations for
verification include:
Percent verified (MF)  =

              BGB verified sheen colonies
              total sheen colonies tested

Percent verified (MPN) =
     Coliform count by the completed test
     Coliform count by the confirmed test
X 100
X 100
The results of verification for the LES (two-step)
MF  and  confirmed  multiple-tube  methods are
summarized in  Table 4. From  263  membrane
filters,  1110  sheen colonies were  selected for
verification; 89.6 percent were verified as coliform
organisms.  Trickling filter effluent  displayed the
highest  verification (97.5 percent)  from the MF
technique.  It was  also the highest (93.3 percent)
in  using  the  MPN procedure. From  97  MPN
samples  a  wide  range of verifications (22-100
percent) were observed;  however,  80 percent of
the MPN samples  reflected  100 percent verifica-
tion. The average verifications for both the MF and
MPN methods were higher than reported by  Geld-
reich  et al. (30):  78.1 percent for MF and 70.3
percent for MPN on samples of natural waters and
sewages.

     Time effect after dechlorination. During the
course of the investigation, the question arose as to
whether or  not,   after  the  dechlorination  of
samples, the  observed bacterial densities signifi-
cantly  fluctuated   with  time. This  seemed  an
important consideration  because of the time ele-
ment involved in   performing  comparative  tech-
niques.  To investigate this, samples of three  types
of secondary effluent were chlorinated at varying
dosages for a contact time of 15  minutes after
which   they  were dechlorinated   as  previously
described and kept at room temperature (20 to
22  C).  Bacterial density assays were undertaken
using the  LES  two-step and the  multiple-tube
                                               120

-------
TABLE 4.   VALIDITY OF TWO-STEP MF AND CONFIRMED MPN TESTS

                                                                  MPN Confirmed Test
                         MF Test
  Chlorinated    No. of          No. of      Avg. %      Number     100%          % Verified
  effluent       Membranes      Colonies    Verified     Tested      Verified      Avg.     Min.
High rate
    act. si.           66

Contact
    Stabilization     101

Trickling
341


393
86.2


88.8
27


35
22


26
92.4


91.0
46.8


50.0
Filter
Total
96
263
376
1,110
97.5
89.6
35
97
30
78
93.3
92.0
22.0
22.0
methods at  15 minute  intervals  for  a  2-hour
period.  The  procedure not  only  permitted  an
assessment  of the time element but also provided
an opportunity for more comparative analyses of
the MF versus  MPN techniques. The results  are
summarized in Table 5.

    There  was no significant change in coliform
densities during the more than 2-hour  period. A
comparison of  the paired  MF  and MPN  results
indicates the inherent precision of the MF method
over that of the MPN.

Fecal Coliforms

    Comparison for assaying fecal  coliforms was
made using the M-FC MF technique recommended
by Geldreich  et at. (16) and  the confirmed MPN
procedures  (4). These procedures have been accept-
ed for fecal coliform  enumerations on  unchlori-
nated  wastewater. Four chlorinated  effluents were
examined.  One  effluent,  representative of  the
Bloomington-Normal,  III., Sanitary  District's acti-
vated sludge process, was collected from  a chlorine
contact tank  effluent stream  and immediately
dechlorinated; the  other  three were treated with
various dosages of chlorine as previously  described.

    The results of the two assay methods  on  the
four effluents are shown in  Figure 5. It is apparent
that most of the plotted  points lie  below the line
of equality. In fact, 78 are below,  12 are  above,
and 6 are on the line.  Adjusting the equality line
for MPN bias as described  by Thomas  (32) does
little to change the pattern; 74 points  are below
and 22 are  above the line. In several cases the dis-
       -4
       u_
       s
       u_
       O

       Q o
                    0
                       EXPLANATION
                      •HIGH-RATE ACTIVATED SLUDGE
                      ^CONTACT STABILIZATION
                      oTRICKLING FILTER          S-*
                      iCI2 CONTACT TANK
                       EFFLUENT            SI*
                                          ALLOWANCE FOR
                                          MPN BIAS
                     0246
                       LOG OF CONFIRMED MPN/100 ml

               Figure 5. Fecal Coliform densities as determined
                        by the Membrane Filter and most
                        probable number techniques on
                        Chlorinated Effluents.
               crepancy is by a factor of 10 or more which are
               not apparent in Figure 5.

                    It can be concluded that the M-FC MF tech-
               nique for fecal  coliform detection, when  applied
               to chlorinated wastewater effluents, is less efficient
               in recovery than the confirmed MPN  procedure.
               Other  media  or an  enrichment step  similar to
               that used in the LES two-step procedure for total
                                              121

-------
 TABLE 5.  COMPARISON OF TOTAL COLIFORM DENSITY (PER  100 ml) IN EFFLUENTS
            AFTER DECHLORINATION, DETECTED BY MF AND MPN TECHNIQUES
 Time (min.) after High rate act. si.
 dechlorination         MF      MPN
                  Contact stabilization
                    MF
        MPN
                      Trickling filter
MF
MPN
0
15
30
45
60
75
90
105
120
135
Arithmetic mean
Geometric mean
Median
Mode
Coeff. of
variation, %
% verified avg.
Range of %
verified
1,500
1,500
1,500
1,600
1,800
1,500
1,600
1,500
1,500
1,500
1,600
1,500
1,500
1,500

6.9
92.0

77.0-100
1,700
1,700
1,700
1,700
1,700
2,200
2,400
2,400
2,200
2,200
2,000
2,000
1,900
1,700

15.7
96.9

69.0-100
6,400
4,100
4,000
4,000
5,200
3,200
3,800
4,500
3,900
—
4,400
4,300
4,000
4,000

21.4
93.5

83.3-100
7,900
6,300
4,900
3,300
3,500
11,000
7,900
3,500
3,500
—
5,800
5,200
4,900
3,500

46.8
97.9

78.8-100
23,000
21,000
25,000
25,000
23,000
24,000
22,000
23,000
23,000
24,000
23,000
23,000
23,000
23,000

5.6
91.4

73.4-100
23,000
35,000
17,000
28,000
35,000
35,000
49,000
22,000
22,000
22,000
29,000
27,000
25,000
22,000

33.2
95.3

53.2-100
 Chlorine dosage
       mg/1
 Contact period,
       mm.
4.0

 15
2.0

 15
      3.0

       15
 *MF  = LES (two-step)
  MPN = Completed tests

coliform  might improve the recovery  efficiency.
Further investigations along these lines would seem
justified.  Braswell and Hoadley (21) found that the
use of Trypticase soy agar was superior to the MPN
and MF techniques for E. coli recoveries in chlori-
nated secondary sewage.

     The minimum and  maximum fecal coliform
ratios of  MF/MPN for all tests were 0.17 and 1.46
respectively. The median ratio was 0.70. Based  on
observations from 96 comparative runs, the rela-
tionship of fecal coliform densities in chlorinated
effluents for the two procedures can be expressed
as

        log MF = 0.012 + 0.942 log MPN
                       The correlation coefficient is 0.987. Until a more
                       precise  procedure  is  developed  for  using  MF
                       techniques  in  recovering  fecal  coliforms from
                       chlorinated wastewater, a mathematical expression
                       of this nature may be useful for estimating MPN
                       densities. The results should be multiplied by the
                       factor 0.851  as described by Thomas (32)  for an
                       estimate without bias.
                            Verifications of membrane developed colonies
                       were made using a phenol red lactose broth and EC
                       broth.  A total of 616 blue colonies were fished for
                       verification; 87.7 percent  were verified. This was
                       lower  than 93.2  percent verification reported by
                       Geldreich et al. (16) on pure cultures.
                                              122

-------
Fecal Streptococci

    Multiple-tube  versus  membrane  filter. To
compare the MPN and  MF techniques a series of
FS tests on unchlorinated samples from a variety
of sources were  performed. The results are sum-
marized in  Table 6.  A statistical test  of the ob-
served  data was  made  using the t-test of pairing
observations to  determine  whether there is  a
significant difference in  FS  recoveries by the
MPN and MF2 methods. The results indicate there
is no statistical difference in the mean value of the
bacterial counts  determined by  the two methods
(test 4 of  Table 2). The  FS  densities obtained
from both procedures are  comparable and  prob-
ably have the same sanitary significance. Therefore
the laboratory techniques of this study were con-
sidered acceptable.

    One  hundred   and  thirty-one  chlorinated
samples taken from  three  secondary  sewage  ef-
fluents were  concurrently  assayed  for FS  densi-
ties by the MPN  and  MF procedures. The colonies
developed  on the membrane  filter were counted
after 2,3,4, and 7 days  incubation and  were desig-
nated   MF2, MF3, MF4,  and  MFy, respectively.
The MF2  and MPN  method is recommended  by
Standard Methods (4).

    The comparative results of  MF2 and MPN on
chlorinated samples  are presented  graphically  in
Figure 6. It is apparent that most of the plotted
points  lie below  the  line of equality. In fact, 107
plotted points are below, 19 are above, and 5 are
on the line. From adjusting the equality line for
the MPN bias, as described by Thomas (32), most
of the plotted points (97 points)  are  below the
MPN bias reference line, 32 points are above and 2
are on the line. Statistically significant differences
were  found  in  FS  recoveries, when  the MPN
procedure with the MF2 method (Test 5 of Table
2).  It can be concluded that the MF2 procedure
gives lower FS recovery on chlorinated effluents
than does the MPN  technique.  It seemed reason-
able  that  enrichment and  prolonged  incubation
might improve FS recovery using the MF method.
TABLE 6.
MOST PROBABLE NUMBER AND
MEMBRANE FILTER COUNTS,
FECALISTREPTOCOCCI PER 100
ML IN UNCHLORINATED WATERS
           Source
                  MPN
Illinois River
Spoon River
Spring Lake

Twin  Lake


Havana Farm Pond

Fiatt Farm  Pond

High-rate Activated
  Sludge Process effluent
Contact Stabilization
  Process Effluent
                   140
                   140
                   130

                 4,600
                 2,200
                   790
                   700
                   540
                   350
                   280
                   230
                   220
                   170
                 1,300
                   340

                   110

                 1,400
                35,000
                33,000
                27,000
                22,000
                22,000
                 3,000

                 9,400
                 4,900
                 4,900
                 4,600
    64
    60
   140

 3,100
 1,600
   900
   830
   900
   300
   300
   240
   200
   230
 1,400
   420

    75

 1,200
34,000
30,000
30,000
26,000
24,000
 2,000

 7,700
 5,400
 4,700
 4,300
     Enrichment. Azide dextrose broth is the me-
dium used for the  presumptive test of the MPN
method for fecal  streptococci  in  waters. Brain-
heart infusion broth and  bile broth medium  are
the confirmation media of FS for the MF method.
These three media were used  in this study for en-
richment  purposes  in efforts to enhance FS re-
covery in  chlorinated  effluents. The results of FS
recovery on M-Enterococcus  agars (MF method)
Trickling Filter
  Process Effluent
                24,000
                11,000
                 9,400
                 4,600
24,000
13,000
 7,000
 5,400
*0ne-step M-Enterococcus agar MF count with
 two-day incubation
                                             123

-------
    10.5
 o
 o
    103
 CM
    101
 EXPLANATION
•HIGH-RATE ACTIVATED SLUDGE,
^CONTACT STABILIZATION
oTRICKLING FILTER
 n=131
                ALLOWANCE FOR
                MPN BIAS
      10'
                103
            MPN/100 ml
105
    Figure 6. Fecal Streptococci densities as
            determined by the One-Step
            Membrane Filter and most probable
             number technique on Chorinated
             Effluents.
                                                        240
                                                      —
                                                      0
                                                   o
                                                      o  80
                                              o
                                              o
                                              o
                                            Q- O
                                            LU or
                                            co £
                   CHLORINATED HIGH'RATE      ,
                   ACTIVATED SLUDGE EFFLUENTS
                                  n=24     /
                                                                       /
       0          80         160        240
ONE-STEP M-ENTEROCOCCUS AGAR MF2 COUNT
                                         Figure?. Comparison of Fecal Streptococci Counts
                                                  on M-Enterococcus Agars with and
                                                  without Azide Deztrose Broth Enrichment.
with  and without  enrichment  for  chlorinated
samples are shown in Figure 7, 8, and 9. All FS
counts in these figures were made after a two-day
incubation. Although the number of colonies per
filter  as  depicted in these figures  exceeded the
desirable range of 20 to 100 per  filter, they were
considered satisfactory  for  comparison  purposes.


    With AD broth enrichment, all  plotted points
lie  below the  equality  line (Figure 7).  In other
words, the FS recovery from chlorinated effluent
on  M-Enterococcus  agar  with  AD  broth enrich-
ment  falls far short of that without enrichment.
This is substantiated by the t-test (Test 6 of Table
2) and it is  concluded  therefore  that enrichment
with AD broth inhibits the FS  recovery of chlori-
nated  samples on  membranes.

    Figure 8  shows no appreciable difference in
FS counts with or without BHI broth enrichment.
Eleven plotted points lie above, 9 lie below, and 5
points are on equality line. A statistical test (Test 7
of  Table 2)  suggests no significant difference in
FS  recoveries from chlorinated  effluents deter-
mined by the MF method with  or without enrich-
ment. Using the  least square regression technique,
the plotted  points in  Figure 8 can be  fitted as
follows:
                                          x  9.
                                          LU «=
                                          LU CC
                                          cc. <
                                          Q. O
                                          03 <
                                          T CO
                                          ^  O
                                          ^  O
                                          V)
                  CHLORINATED TRICKLING
                     FILTER EFFLUENTS
                           n = 25
                                                                            i      /
                                                                                /
                                                                             •  /

                                                 OLL
                                                  0           40            80
                                          ONE-STEP M-ENTEROCOCCUS AGAR MF2 COUNT


                                         Figure 8. Comparison of Fecal Streptococci Counts
                                                  on M-Enterococcus Agars with and
                                                  without Brain.Heart Infusion Broth
                                                  Enrichment.
                                              124

-------
   o
   ^160
LLJ <£
CO

       80
&
        0
                                        /
         •»/_
             /
                               A   X
                              -;x
                  •  A*£/

                  ^A/X      EXPLANATION
                 .  'A         n = 53
                         o TRICKLING FILTER
                 ACONTACT STABILIZATION
             HIGH-RATE ACTIVATED SLUDGE
         0            80           160
 ONE-STEP M-ENTEROCOCCUS AGAR MF2 COUNT

Figure 9. Comparison of Fecal Streptococci Counts
         of M-Enterococcus Agars with and
         without Bile Broth Enrichment.


               Y = 0.097 + 0.98 X

 in which Y = FS counts by M-Enterococcus agar
 MF2 with  BHI medium enrichment,  in organisms
 per 100 ml; X = FS counts by one-step M-Enter-
 ococcus agar MF2 method, in  organisms per 100
 ml. The correlation coefficient  is 0.97. Equation 5
 shows  the  slope to be 0.97 with  an intercept  of
 0.097.  Thus  the regression  line expressed by
 Equation 1  is  almost identical with a 45 degree
 line. From these tests, it  can  be reasonably con-
 cluded  there is no advantage to BHI broth enrich-
 ment for the MF method on chlorinated effluent
 samples.

     It  is quite evident from the data depicted in
 Figure  9 that  the  FS recovery using bile  broth
 enrichment  is  higher  than FS  recovery by  non-
 enrichment  techniques.  Fifty-three  comparisons
 were made on three  effluents and  only two ef-
 fluent samples  showed the enrichment FS counts
 slightly  less than  the  nonenrichment. This is con-
 firmed  by statistical analyses (Test 8 of Table 2).
 It is concluded that  bile broth enrichment did
 improve FS  recovery  on  chlorinated effluent
 samples.

     A  peptone yeast-extract casitone enrichment
 broth was suggested by  Rose and Litsky (28) for
 use with the MF  method for  the  enrichment  of
 FS recovery  in unchlorinated waters. To deter-
 mine the efficiency of PYC broth on chlorinated
effluents samples,  parallel tests were made with
PYC  broth,  bile  broth  medium, and  without
enrichment on portions of the same samples.

     About  one-half of the experimental  results
were  discarded  due to  extremely  high  or low
counts.  The  results (68 samples),  where  filter
counts were in  the  desirable range of 20  to 100,
are summarized in Table 7.  The values of Table 7
represents a  two-day  incubation period.  For all
tested effluents, with few exceptions, the recovery
of  FS increased with enrichment and especially
with bile broth enrichment.

     The recovery  ratios of enrichment to non-
enrichment for each effluent are presented in Table
8. The highest ratios were 2.45:1  and 1.77:1 for
bile  broth   and  PYC  enrichment,  respectively.
Similarly,  the overall  average  ratios  for  the  68
samples were 2.14:1 and 1.60:1.  In  comparison
with the work done by Rose and Litsky (28), the
recovery  ratio  of  PYC  enrichment  to M-Enter-
ococcus agar was 2.44:1 in waters.

     Nineteen chlorinated samples were examined
for FS densities by both MPN and PYC enrichment
MF  methods.  The  results from these assays are
depicted  in  Figure 10.  The equality  line  was
adjusted for MPN bias as described by Thomas (32)
and  used for reference.  Fourteen  plotted  points
lie below the equality line, 4 are above and 1 is on
the line. A t-test analysis confirmed the differences
                                                    105
                                                 O    o
                                                 ^  103
                                                 CM
                                                    10
                                                         EXPLANATION
                                                         HIGH-RATE ACTIVATED SLUDGE /'
                                                       ACONTACT STABILIZATION    /
                                                       ^TRICKLING FILTER        / '
                                                         n=19               A/X&  '
                                                                            /
                                                                /°
                                                      101
                       103
                  MPN/100 ml
                                                 Figure 10. Comparison of Fecal Streptococci
                                                          densities determined by the MPN and
                                                          the PYC Enrichment MF methods.
                                             125

-------
TABLE 7.  COMPARISON OF RECOVERY OF FECAL STREPTOCOCCI ON
          M-ENTEROCOCCUS AGARS WITH AND WITHOUT ENRICHMENT
Sample
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Average
Activated Sludge
NE
12
75
61
17
32
57
36
19
25
12
29
65
12
17
45
19
31
28
42
52
46





35
PYC
22
74
95
30
57
73
47
30
37
20
46
71
22
28
52
30
48
45
66
83
86





51
Bile
50
180
160
54
111
97
52
54
68
26
59
90
26
35
74
36
58
48
70
93
96





73
Contact Stabilization
NE
54
28
48
72
61
23
69
44
20
53
59
44
34
88
39
24
53
31
23
40
72





47
PYC
75
32
80
95
86
29
91
62
30
104
90
65
36
91
72
34
71
46
38
65
100





66
Bile
83
38
90
112
91
29
107
63
30
90
104
80
52
95
62
38
82
70
58
98
122





76
Trickling Filter
NE
20
16
20
18
26
32
40
60
24
18
60
16
30
20
24
18
11
12
21
50
16
30
18
24
30
16
26
PYC
35
22
52
34
60
66
80
100
40
29
96
30
38
36
42
32
25
21
22
89
38
36
32
28
56
30
45
Bile
40
46
50
38
58
90
100
110
54
45
126
38
54
50
48
45
40
39
34
132
36
62
52
56
90
50
61
Incubation time was 48 hours for all cases; NE means M-Enterococcus agar without enrichment;
PYC means with PYC broth enrichment; and Bile means with bile broth medium enrichment.


TABLE 8.  FS RECOVERY RATIOS OF ENRICHMENT OF NON-ENRICHMENT
                                                     Ratio of*
                                          PYC/NE
  One-step M-Enterococcus agar MF count with two-day incubation

                                        126
Bile/NE
Chlorinated effluent
High-rate Activated Sludge
Contact Stabilization
Trickling Filter
Overall
Range
0.99-1.87
1.06-1.96
1.05-2.60
0.99-2.60
Average
1.54
1.44
1.77
1.60
Range
1.38-4.16
1.08-2.45
1.62-3.25
1.08-4.16
Average
2.24
1.67
2.45
2.14

-------
(Test 9 of Table 2).  From this test it is concluded
that the recovery of FS from chlorinated effluents
on PYC enriched membrane filters is less than for
the MPN  procedure. Although prolonged incuba-
tion  through  seven  days on  PYC enriched filters
showed  increasing counts with time.  No attempt
was  made  to  compare prolonged  PYC  enriched
MF counts with MPN values.

     Bile broth enrichment, as mentioned earlier,
gave  the highest  recovery of FS from chlorinated
effluent samples. To compare the bile broth en-
riched MF2 results with the MPN data, 45 chlori-
nated samples collected from three sewage  efflu-
ents  were subjected to FS assays,  in parallel, by
both methods. The results of the analyses are pre-
sented in Figure 11. The ratios of the bile enriched
MF2 to the  MPN  FS  densities  were calculated,
arrayed  in order of magnitude, and plotted on log-
probability paper.  The line of the best fit was
drawn.  The median, or 50 percentile of the 45
ratios is 1.00. In  fact, 4 ratios are equal to, 21 are
greater than, and 20 ratios are less than unity. This
indicates that the bile enriched MF2 data is in very
close  agreement  to that data obtained  by  MPN
techniques. It was also observed that there was no
significant  increase  in FS  count on  the  bile en-
riched filters for  prolonged incubation up through
seven days. It is concluded that the bile enrichment
MF2 method  is  superior to  the  PYC enrichment
MF2 method and comparable to the MPN proce-
dure  for  the  recovery  of  FS  in  chlorinated
secondary sewage effluents.
    3.0
 Q_
    1.0
 00
    0.3
  EXPLANATION
  HIGH-RATE ACTIVATED SLUDGE
^CONTACT STABILIZATION
o TRICKLING FILTER          ^
  n=45
      0.5       10       50       90
     PERCENT PROBABILITY OF NOT EXCEEDING

Figure 11. Analysis of Fecal Streptococci Recovery
          made on Chlorinated Effluents by use of
          Multipletube (MPN) Test and of
          M-Enterococcus Agar MF with Bile Broth
          Enrichment.
                                               Prolonged Incubation. As  stated earlier, the
                                          M-Enterococcus  agar  MF2  (non-enriched)  techni-
                                          que tends to produce lower FS recovery than the
                                          MPN  procedure  on  chlorinated  effluents  (see
                                          Figure 6). Colonies developed for two-day  incuba-
                                          tion were generally small. To check the effects of
                                          prolonged  incubation  on  FS  recovery  for  the
                                          M-Enterococcus   agar  MF   technique  all  filters
                                          were counted at the end of 2,  3, 4, and 7 days
                                          incubation periods.  Figure  12, a typical example,
                                          shows the general trend of the FS counts with
                                          incubation time. The FS recovery increased signifi-
                                          cantly  up through  the  three-day  period. After
                                          three days, the  FS counts leveled off for chlori-
                                          nated  effluents.  For the  unchlorinated  effluent
                                          sample, no significant increase was  found in FS
                                          counts  after a two-day incubation. The ratios of
                                          MF3 to MF2 for unchlorinated and chlorinated
                                          effluents  are summarized in Table 9. For chlori-
                                          nated samples the average MF3/MF2 values ranged
                                          from a low of 1.27 for contact stabilization efflu-
                                            10,000r;
                                              5000
o
o
o
o
o
£
cc
in
                                                  CJ
                                                  LU
                                                     1000
                                                      500
                                               100
               i    I    I     I    i    i    i  _
              UNCHLORINATED CONTACT   -
              STABILIZATION
                      -v	v—	

                     HIGH-RATE  ACTIVATED-
                     SLUDGE
                        	
                        TRICKLING FILTER
                                                           CONTACT STABILIZATION
                                                                                     -A—
                 CHLORINE DOSAGE: 2 mg/l
                  CONTACT TIME: 15 min
               I    I    I     I    I    I    i
                                                          2468
                                                          INCUBATION TIME, days
                                            Figure 12. Recovery of Fecal Streptococci on
                                                      M-Enterococcus Agar from one
                                                      Unchlorinated and three Chlorinated
                                                      Secondary Effluents.
                                              127

-------
TABLE 9.  FECAL STREPTOCOCCI COUNT MF3/MF2 RATIO
                              Unchlorinated
                                      Chlorinated


Type of Effluent
Number
of
Sample


Range


Average
Number
of
Sample


Range


Average
High-rate Activated
  Sludge

Contact
1.00-1.33
1.14
41
1.17-4.68
2.12
Stabilization
Trickling Filter
Overall
4
4
12
1.00-1.09
1.00-1.14
1.00-1.33
1.07
1.06
1.09
39
44
124
1.07-1.67
1.07-4.28
1.07-4.68
1.27
2.07
1.84
 ent, to a high of 2.12 for high-rate activated sludge
 with an overall average of 1.84.

     To compare the non-enriched MF3 data with
 the MPN results, 124 comparisons, made on three
 chlorinated effluents,  are depicted in  Figure 13.
 Seventy-six plotted  points are above the line of
 equality, and 38  are  below. Using the corrected
 MPN bias as a reference line, 101 points are above,
 21 are below, and 2 are on the line. The MF3 re-
 sults were  found to be slightly  higher than the
 MPN data, especially when the  FS counts were less
 than 500/100 ml (Figure  13). For the  124  in-
 stances, the  geometric mean  values  were 1,300
 MF3/100  ml and 1,000  MPN/100 ml. The geo-
 metric  standard deviations were 3.98 and 4.93 for
 the MF3 and the MPN methods, respectively. How-
 ever, a statistical test (Test 10 of Table 2) did not
 indicate a significant difference between the MPN
 and MF3 methods.

     When  comparing MF3 and  MPN results for
 124  chlorinated  effluent samples in a  manner
 similar to that depicted in Figure 11 the median, or
 50 percentile, for the MF3/MPN is 1.11. Although
 the MF3 values are  slightly higher than  the MPN
 data, the MF procedure  for 3-day incubation on
 M-Enterococcus agar, without enrichment, appears
 applicable for the FS assay of chlorinated efflu-
 ents.

      Verification. A total number of 967 colonies
 were fished from 306 membrane filters and  sub-
 jected  to the verification  procedure outlined  in
 Standard Methods (4). The results of the verifica-
 tion are summarized  in Table 10. These include
                    105
                 \ 103
                  CO
                         EXPLANATION
                        •HIGH-RATE ACTIVATED SLUDGE
                        ^CONTACT STABILIZATION
                        ^TRICKLING FILTER
                         n  124
                              ALLOWANCE FOR MPN BIAS
                      101
                         103
                      MPN/100ml
                             105
                 Figure 13. Comparison of Fecal Streptococci
                           densities determined by the MPN and
                           M-Enterococcus Agar MF3 methods.
                colonies grown on filters placed on M-Enterococ-
                cus agars with and without enrichment. After two-
                day incubation, all of 688 colonies isolated from
                unchlorinated  and chlorinated effluents were, veri-
                fied as fecal streptococci. Although  Kenner et al.
                (34) reported  similar 100 percent recovery of FS
                from  the  membranes for the fecal samples, Rose
                and  Litsky  (28)  experienced a 94.6 percent  FS
                verification from filters placed on  M-Enterococcus
                agars  with  and  without  PYC  enrichment for
                                              128

-------
TABLE 10. VERIFICATION OF FS GROWN ON FILTERS PLACED ON  M-ENTEROCOCCUS AGARS
           WITH AND WITHOUT ENRICHMENT
Sample
Unchlorinated Effluents
Chlorinated Effluents


Overall
Growth after
days of
Incubation
2
2
3
4
—
No. of
Filter
27
189
74
16
306
No. of
Colonies
Examined
92
596
234
45
967
Positive Verified
No. of
Colonies Percent
92
596
223
42
954
100
100
95.3
93.3
98.6
natural waters. From markings placed on the back
of petri dishes during this study it was possible to
distinguish two, three, and four days growth colo-
nies. About 5-7 percent of the three and four day
growth colonies were not verified as FS (Table 10).
        SUMMARY AND CONCLUSIONS

     Two series of  laboratory assays were per-
formed to determine whether or not the standard
membrane  filter  (MF)  procedure  for total coli-
forms,  fecal coliforms, and  fecal  streptococci
detections on chlorinated secondary sewage efflu-
ents was comparable to  that obtained  by the mul-
tiple-tube (MPN)  method. If not found to be the
case,  efforts were made to improve  bacteria  re-
coveries  using various modifications  of  the MF
method.

     Grab samples  of  secondary  effluents  were
collected with  up through  6  mg/1  of chlorine,
stirred, and  dechlorinated by sodium thiosulfate.
After varying periods of contact the samples were
assayed for bacteria. Based upon the results derived
from  this work,  the  following conclusions  were
developed.

     For chlorinated secondary sewage effluents,
the recoveries of TC, FC, and FS by  the standard
MF  (one-step  nonenrichment) method is signifi-
cantly less  than  that obtained  by the standard
MPN procedure.

     The use of the  LES two-step  MF method is
comparable to the completed MPN  procedures for
total coliform detection. Total  coliform recovery
by  the LES two-step MF technique is approxi-
mately 1.5 times that obtained using the M-Endo,
one-step MF procedure.  From 273 filters using the
LES two-step  MF  procedure  and  1,110  sheen
colonies, 89.6  percent were verified as colifnrm
organisms.

     Estimates  of FC MPN  densities may be de-
rived from the MF procedure by using a  mathe-
matical relationship similar to  log MPN =  1.062
log  MF — 0.014. For FC verification, 87.7 percent
of 616  blue colonies were verified.

     The  use  of   azide-dextrose  broth,   brain-
heart  infusion  broth  and  peptone yeast-extract
Casitone  for enrichment purposes,  with the M-
Enterococcus agar  MF2 procedure,  did not satis-
factorily  increase the sensitivity of the procedure
for  FS  assays. Enrichment with bile broth medium
of the  M-Enterococcus agar MF2 procedure signifi-
cantly  increases the  FS recovery to the extent that
the  procedure is comparable to the multiple-tube
method.

     The recovery of FS using the membrane filter
technique with M-Enterococcus  agar  increased
significantly after  three days incubation  (MF3>
compared to two days incubation (MF2>; and the
MF3 procedure is comparable to the multiple-tube
method for FS detection. The membrane filter
technique preferred for FS  assays  is the MF2
procedure using M-Enterococcus agar with bile
broth enrichment. All  of 688 colonies for two-day
incubation  on filters were verified  as fecal  strep-
tococci.
                                              129

-------
           ACKNOWLEDGEMENTS

     The writer wishes to express appreciation to
Ralph  L.  Evans, Head of the Water Quality Sec-
tion, Illinois  State Water Survey, Peoria, for his
encouragement  and review of this paper. Special
acknowledgements are also extended to Davis B.
Beuscher,  Jack W. Williams, Pamella A. Martin and
Donald H. Schnepper of the Water Quality Section
for their assistance during the study; and to John
W.  Brother, Jr. of the Survey who prepared the
illustrations.

                REFERENCES

1.    Illinois  Environmental   Protection Agency.
     Water Pollution Regulations in Illinois. 36 pp.
     March 7, 1972.
2.    Stanier,  R.Y., M. Doudoroff, and E.A. Adel-
     berg.  The Microbial World.  Prentice  Hall,
     Inc., Englewood Cliffs, N.J., 43 pp., 1963.
3.    Kabler,  P.W.  Microbial  Considerations  in
     Drinking Water. J. Amer. Water Works Assn.,
     60(10): 1173-1180, 1968.
4.    Standard  Methods  for  the  Examination  of
     Water and Wastewater,  13th  ed. Amer.  Pub.
     Health Assn., Inc., New York, 1971.
5.    The Working Group of Water Quality of the
     Subcommittee  on   Water  Quality,  Interde-
     partmental Committee  on Water.  Guidelines
     for Water Quality Objectives and Standards, A
     Preliminary Report. Technical Bulletin No.67,
     Inland Water Branch, Department  of the En-
     vironment, Ottawa, Canada, pp. 14-25., 1972.
6.    Gallagher,  T.P., and D.F. Spino. The Signifi-
     cance of Numbers of Coliform Bacteria as an
     Indicator of  Enteric Pathogens.  Water  Re-
     search,  Pergamon   Press, Great   Britain  v.
     2(1):169-175, 1968.
7.    Geldreich,  E.E. Applying Bacteriological  Para-
     meters to  Recreational  Water Quality. J.
     Amer. Water Works Assn. (2):113-120, 1970.
8.    Smith, R.J.,  and R.M.  Twedt.  Natural Rela-
     tionships of Indicator and Pathogenic Bacteria
     in Stream  Waters. J. Water Poll. Contr. Fed.
     43(1):2000-2209, 1971.
9.    Smith, R.J.,  R.M. Twedt, and L.K. Flanigan.
     Relationships of  Indicator  and  Pathogenic
     Bacteria  in  Stream  Waters.  J. Water  Poll.
     Contr. Fed. 45(8): 1736-1745, 1973.
10.  The  Committee  on Environmental Quality
     Management  of the Sanitary Engineering Divi-
     sion. Engineering Evaluation of Virus  Hazard
     in Water. Amer. Soc. & Civil  Eng., Journal of
     Sanitary Engineering Division,  96(SAI):111-
     161,  1970.
11.  Geldreich, E.E.,and N.A.  Clarke. The  Coli-
     form Test: A Criterion for the Viral Safety of
     Water.  In: Virus and Water  Quality: Occur-
     rence  and Control.  Proceedings  13th Water
     Quality  Conference,  Department  of  Civil
     Engineering,  University  of  Illinois at Urbana-
     Champaign, ML, pp. 103-113,  1971.
12.  McKee, J.E.,  R.T.  Mclaughlin, and P. Les-
     geurgues.   Application  of  Molecular  Filter
     Techniques to  the Bacteria Assay of Sewage.
     III.  Effects of Physical and  Chemical  Disin-
     fection. Sewage Ind. Wastes, 30(3):245-252,
     1958.
13.  McKee, J.E., and R.T.  Mclaughlin.  Applica-
     tion  of Molecular  Filter Techniques to the
     Bacteria Assay  of  Sewage.  II  Experimental
     Results for Settled  Sewage.  Sewage & Ind.
     Wastes 30(2):129-137, 1958.
14.  McCarthy,  J.A.,  J.E.  Delaney,  and  R.J.
     Grasso. Measuring Coliforms  in Water. Water
     & Sewage Works 108(6):238-243,  1961.
15.  Geldreich, E.E., H.F. Clark,  C.B. Huff, and
     L.C.  Best. Fecal-Coliform-Organisms Medium
     for the Membrane Filter Technique. J. Amer.
     Waterworks Assn. 57(2):208-214, 1965.
16.  Geldreich, E.E. Sanitary Significance of Fecal
     Coliforms  in the Environments. U.S. Depart-
     ment of Interior, FWPCA publication WP-20-
     3, 122pp., 1966.
17.  Evans, F.L. Ill, E.E. Geldreich, S.R. Weibel,
     and G.G. Robeck. Treatment of Urban Storm-
     water  Runoff. J. Water  Poll. Contr.  Fed.
     48(5)R162-R170, 1968.
18.  ORSANCO Water  Users  Committee.  Total
     Coliform:  Fecal Coliform  Ratio  for Evalua-
     tion  of  Raw  Water  Bacterial  Quality.  J.
     Water Poll. Contr. Fed. 43(4)630-640, 1971.
19.  Maxey, R.B.  Non-lethal Injury and Limita-
     tions of Recovery of Coliform  Organisms on
     Selective Media. J. of Milk and Food Technol.
     33(9):445-448, 1970.
20.  Scheusner  D.L., F.F. Busta, and M.L. Speck.
     Injury of  Bacteria by Sanitizers. Appl. Micro-
     biol. 21(1):41-45, 1971.
21.  Braswell,  J.R., and A.W. Hoadley. Recovery
     of Escherichia coli from Chlorinated Second-
     ary Sewage. Appl. Microbiol. 28(2):328-329,
     1974.
22.  Hufham, J.B. Evaluating the Membrane Fecal
     Coliform Test by Using Escherichia coli as the
     Indicator  Organism.  Appl.  Microbiol. 27(4):
     771-776, 1974.
23.  Presswood, W.G., and L.R. Brown. Compari-
     son of Gelman and  Millipore Membrane Fil-
     ters for Enumerating Fecal Coliform Bacteria.
     Appl. Microbiol. 26(3):332-336, 1973.
                                              130

-------
24.  Lattanzi, W.E., and E.W. Mood. A Compari-
     son of Enterococci and Escherichia coli as In-
     dices  of Water Pollution. Sewage  &  Ind.
     Wastes. 23(9) :1154-1160, 1951.
25.  Litsky, W., W.L. Mailman, and C.W. Fifield.
     A  New Medium for the Detection of Enter-
     ococci in Water. Amer. J. of Pub. Health.
     43(7):873-879, 1953.
26.  Slanetz,  L.W., and C.H. Bartley. Number of
     Enterococci  in  Water,  Sewage  and Feces
     Determined  by the Membrane Filter Techni-
     que with an Improved  Medium. J. of Bact.
     74(5):591-595, 1957.
27.  Kenner, B.A.,  H.F. Clark, and P.W. Kabler.
     Fecal  Streptococci I. Cultivation and Enumer-
     ation  of Streptococci in Surface Waters. Appl.
     Microbiol.9(1):15-20, 1961.
28.  Rose,   R.E.,  and  W.  Litsky.  Enrichment
     Procedure for  Use  with  the Membrane Filter
     for the  Isolation and Enumeration  of Fecal
     Streptococci  in  Water.  Appl.  Microbiol.
     13(1):106-108, 1965.
29.  Pavola,  M.T.,  E. Beauvais,  FT. Bresenski,
     and W.  Litsky. Rapid Assessment of Water
     Quality  by Flourescent  Antibody Identifica-
     tion of Fecal Streptococci. Proceeding of the
     6th International Water Pollution  Research,
     Pergamon Press Ltd. London,  Great Britain,
     A/3/5/1 -A/3/5/8, 1972.
30.  Geldreich, E.E., W.L.  Jeter, and J.A. Winter.
     Technical  Considerations  in   Applying  the
     Membrane Filter Procedure. Health  Lab.  Sci.
     (4)2:113-125, 1967.
31.  Seiz,  D. Personal  Communication.  Illinois
     Department of Public Health, Springfield.
32.  Thomas,  H.A.  Jr.  Statistical  Analysis of
     Coliform  Data.  Sewage & Ind. Wastes 27(2):
     212-222, 1955.
33.  Dixon, W.J., and F.J. Massey, Jr. Introduc-
     tion   to  Statistical  Analysis.  McGraw-Hill,
     New York, pp. 384, 1957.
34.  Bartley, C.H.,  and  L.W. Slanetz.  Types  and
     Sanitary  Significance  of Fecal Streptococci
     Isolated  from  Feces,  Sewage and  Water.
     Amer.  J.  Pub.  Health.  50(10): 1545-1552,
     1960.
      QUESTION AND ANSWER SESSION

Geldreich:  Dr.  Lin,  with  enterococci, we find
           when  we try to go through  enrich-
           ment with extended incubation time,
           we run into  pediococci   and other
           organisms which  are not fecal streps.
           I don't recall whether you had a table
           containing this data. Did you  check
            verification when you extended either
            the time of incubation or  the use of
            these other enrichment devices before
            you  went  to   M-Enterococcus  to
            show that this increase was still  from
            fecal  strep rather than possibly  some
            of these other species that we know
            will   grow  on   M-Enterococcus  or
            some other strep media?  What was
            that increase,  80% or 90%? That was
            good.

Lin:        Eighty  to  90% of the increase was
            from fecal strep.


Geldreich:   So  even  though you extended the
            time and you  added enrichments you
            didn't get  any  more  false positives.
            Good.

Bordner:    Dr. Lin, Did you evaluate media other
            than  M-Enterococcus  agar for  fecal
            streps,  for example  PSE agar or the
            KF medium?
Lin:       We restricted our studies to the  M-
           Enterococcus.

Bordner:   Shifting back to the M-FC medium, I
           understood   you  to say  that  you
           haven't looked  into enrichment, for
           fecal  coliforms.  Is  this  your  next
           plan?

Lin:       You mean this one?  Yes, but I didn't
           succeed.

Bordner:   Do you plan to do this in the future?

Lin:       Yes.

Bordner:   May I  ask what brands of filters you
           usually use?

Lin:       Millipore.

Brodsky:   Perhaps  I should  direct this  short
           question  to  Geldreich. I  have read in
           the  literature  that  M-Enterococcus
           agar is   selective for certain  groups
           or species of streptococci. In compar-
           ison with PSE agar  and  KF agar the
           terms  fecal  streptococci  and enter-
           ococci tend  to  be  used  interchang-
           ably.  Could you  clarify these  terms?
                                              131

-------
Geldreich:  We find that M-Enterococcus agar is
           very  selective for  enterococci  but
           there are other fecal streptococci that
           we are concerned  about.  These are
           from   other  warm-blooded  animals.
           Feed  lots have a tremendous number
           of streptococci which do not recover
           too well on M-Enterococcus agar. KF
           agar  and  PSE  agar  recover  these
           species much better. We know we get
           equivalent  results with  M-Enterococ-
           cus and KF when we use it on domes-
           tic sewage  because  we are  looking at
           enterococci,  the sub-group  of  fecal
           strep. Since we are working  on  ratio
           development  of  fecal  coliform  and
fecal strep in a stream, there are times
when we do have animal feedlot dis-
charges  and  slaughter house waste. It
would  be easier to  stay  with  one
medium that will  recover all  of the
members of  the fecal strep group be-
cause otherwise  your  ratio  won't
mean a  darn thing.  We have  always
recommended the KF agar. Recently
PSE  agar which is being introduced,
looks like an excellent medium.  I am
sure  Warren  and Fran  have  used it
and with excellent results.  It may be
far superior  to M-Enterococcus  agar,
particularly   for  relationships  with
fecal  coliforms  and  pollution  from
feedlots or from domestic wastes.
                                                132

-------
                 THE ASTM PROPOSED MEMBRANE FILTER TEST PROCEDURE
                         FOR THE RECOVERY OF FECAL COLIFORMS
                                      Margareta J. Jackson
                                          Microbe One
                                      Ann Arbor, Michigan
                                        Donald W. Davis
                                        George R. Kinzer
                                   Johns-Manville R&D Center
                                        Denver, Colorado
                  ABSTRACT

     Results of  a  collaborative study of the pro-
posed  ASTM test  procedure for  the recovery of
fecal  coliforms  on membrane filters (MF)  were
presented.  The test procedure is one of a series
planned  by the ASTM  subcommittee  D  19.08
.04.02 to  evaluate MF materials.  Recoveries of
fecal  coliforms on MF's were compared to those
on pour or spread  plates using M-FC agar. Analysis
of variance indicated  differences  in filter brands
and  in the results obtained in different  labora-
tories. The laboratory effect could not be sep-
arated from sample effect.

     The test procedure did not satisfy all of test
objectives  and   must  be  rewritten  to eliminate
variables and options.
               INTRODUCTION

     Since the acceptance of the membrane filter
technique for the  isolation and enumeration of
total and fecal coliform and fecal streptococcid),
conflicting papers (2,3,4,5,6) have appeared in the
literature about the use of membrane filters as a
method of evaluating the quality of water. In June
of 1973, ASTM  brought together microbiologists,
membrane  filter  manufacturers, media  manufac-
turers, regulatory agencies, consumers, and users of
membrane filters to discuss the problems that con-
fronted them. At this meeting, it was decided to
examine the following  parameters of membrane
filters:
              1.  Inhibitory effects
              2.  Recovery
              3.  Retention
              4.  pH
              5.  Shelf life
              6.  Sterility

    Since existing test procedures for inhibitory
effects and recovery were based  on 35C (7,8), it
was agreed to proceed  with a draft for inhibitory
effects and recovery using the elevated tempera-
ture of 44.5±0.2C for fecal coliforms.

    The standard test method  for recovery was
designed to determine  the ability of a membrane
to recover  fecal coliform  organisms from  un-
treated  water samples  on a selective differential
medium. The test method was based on recovery
by the M-FC  method currently being used in water
testing  laboratories (1)  compared  to pour and
spread plate procedures.

         METHODS AND MATERIALS

    Four polluted  waters and  one  raw  sewage
were collected by  each participating laboratory and
serially  diluted to obtain  fecal coliform densities
of 20-60 organisms per ml.

    Five replicate dilutions  of each  sample were
filtered through  test  membranes,  transferred  to
M-FC  agar and incubated  at 44.5 C ± 0.2 C for 22-
24 hours. The same  dilutions of each sample were
also tested on M-FC  agar by the pour plate  or
                                             133

-------
spread  plate technique.  The  blue colonies  were
counted  with  a  stereomicroscope for the  MF
technique and  with  a Quebec colony  counter for
the pour and spread plates. Twenty colonies per
sample from one representative pour and/or spread
plate were verified  using  lactose broth and EC
broth according to the test procedure.

     By  comparing  relicate  membrane   filter
counts  with replicate  pour and/or spread  plate
counts,  the recovery rate of fecal coliforms on
membrane filters was determined. If the arithmetic
mean counts on five membrane filters  was 85 per-
cent or greather than the arithmetic mean of the
five pour and/or  spread plate counts, the mem-
brane filter  had met the criteria for  recovery of
fecal coliforms.

     A   preliminary  study of round  robin was
initiated  to determine whether a valid test proce-
dure had been drafted.  A  common source  of
media and  membrane filters was essential for the
study.   Six  manufacturers supplied   membrane
filters (0.45 micron pore, white, gridded, sterile
47 mm). Each participating laboratory  received the
same lot  number of membrane  filters  from the
manufacturers.  Difco  supplied  media  with the
same control numbers to each laboratory.
      Media
Code
1.
2.
3.
4.
5.
Bacto-Peptone
M-FC Agar
Rosolic Acid
Lactose Broth
EC Medium
0118-01
0677-01
3229-09
0004-02
0314-02
Control No.

  602509
  586112
  596061
  597469
  598803
    Several laboratories were able to test all six
membranes, whereas, the other labs tested two to
four membrane filters. A common data sheet was
used to record each  laboratory's evaluation. These
data sheets were then  submitted  for statistical
analysis to determine whether  a valid test proce-
dure had been drafted.
LABORATORIES PARTICIPATING IN
PRELIMINARY ROUND ROBIN - COMMITTEE
D 19.0804.02
1.   Methods Development and Quality Assurance
     Research Laboratory
     U. S. Environmental Protection Agency
     Cincinnati, Ohio
                          2.   Canada Center Inland Waters
                              Microbiology  Laboratory
                              Burlington, Ontario

                          3.   Millipore Corporation
                              Bedford, Massachusetts

                          4.   Department of Environmental Sciences
                              University of Massachusetts
                              Amherst, Massachusetts

                          5.   Division of Laboratory Services
                              Illinois Environmental Protection Agency
                              Chicago, Illinois

                          6.   Gelman Instrument Company
                              Ann Arbor, Michigan

                          7.   Sartorius Membrane Filter GMBH
                              West Germany

                          8.   Ministry of the Environment
                              Division of Laboratories
                              Bacteriology Branch
                              Rexdale, Ontario

                          9.   Johns-Manville R&D Center
                              Denver, Colorado
                                                  MEMBRANE MANUFACTURERS OR
                                                  DISTRIBUTORS  PARTICIPATING IN
                                                  PRELIMINARY ROUND ROBIN TESTS
                          1.   Sartorius
                          2.   Gelman
                                   Beckman
                                   Anaheim, California
                                   (Manufactured in Germany)

                                   Gelman Instrument Company
                                   Ann Arbor, Michigan
                                                  3.   John-Manville  John-Manville Corporation
                                                                    Denver, Colorado
                          4.   Oxoid


                          5.   Millipore


                          6.   S&S
                                   Med-Ox Chemicals Limited
                                   Ottawa, Canada

                                   Millipore Corporation
                                   Bedford, Massachusetts

                                   Schleicher and Schuell, Inc.
                                   Keene, New Hampshire
                                              134

-------
                   RESULTS
              (DATA ANALYSIS)

     The raw count data from the supplied forms
were converted to a machine-readable form and the
mean and standard deviations were computed for
each set of replicates. These calculated values were
then used to determine the recovery of fecal  coli-
form  as a percentage of  the  pour plate (PP) and
spread plate (SP) results.  In addition, the recovery
calculations  were repeated,  rejecting  all counts
outside of the ranges of 20 to 60 counts on mem-
brane  filters and 30 to  300  colonies on pour or
spread plates. These calculations are listed on Table
1. The "ERR" listed  is determined  from  the
standard  deviations of both measurements by cal-
culating the standard deviations as a percentage of
the mean. These percentages were summed and the
recovery  was multiplied  by  this percentage  to
obtain the "ERR." The recovery data was used in
the remaining statistical analyses.

     The data from  all  participating laboratories
were  plotted. In general, the data appeared  scat-
tered for all laboratories  except one. This labora-
tory showed no differences between filters.

     Data from  three laboratories  were selected
for  the  analysis of  variance.  The  laboratories
were selected on the basis that:

     1.  All six filters were tested.

     2.  At least 5 water samples were run.

     3.  Both  spread and pour plate standards
         were run on all five samples tested.

     These laboratories included one filter manu-
facturer,  one  university, and  one government
agency.

     Two analyses of variance were run, one using
the spread  plate standard and one using the pour
plate  standard.  The  three  variables used in  the
analyses were:

     Laboratories (L)
     Filter Membrane (F)
     Water Samples (S)


     Since  the  water samples selected  by   each
laboratory  were  different, the  variability due  to
the samples, the laboratory-sample interaction, and
the filter-sample interaction  are meaningless for
these analyses.
     The  results of the  analysis of variance  are
summarized in Tables 2 and 3.

     The following conclusions can be drawn from
the analyses:

     1.    In both cases the F x  L interaction was
          not  statistically  significant,  i.e., the fil-
          ters  behaved the same in all laboratories.

     2.    There  are  differences between  filters
          with manufacturer 3 supplying the best
          filter and manufacturer 2 supplying the
          poorest filter.  All  other filters supplied
          were about the same.  Statistical signifi-
          cance represented the 95 to 97.5 percent
          level.

     3.    The difference between laboratories was
          highly  significant  (over 99  percent). It
          is not clear from this analysis if these dif-
          ferences  are  due  to  technique in  the
          laboratories  or  due to  differences  in
          water samples used, since these sample
          differences  are included  in  this effect
          and  in the residual and cannot be sepa-
          rated.

     The  contribution of the various sources  to
the total test variance were determined using com-
ponents of variance analysis for both sets of data.
Since the F x  L interaction was not significant, a
better estimate of  residual error can be made by
pooling the sum of squares  for  this term and for
the "residual". The calculation of these variances
is   listed in Tables 4 and 5.  It should be noted
that the variance contribution of the filters is only
5.5 percent in the case of spread plates or 7.3 per-
cent  in the case of the  pour plates  of the total
variance.  The  large sources of  variance are  the
"residual" (unexplained  or random) variance and
the inter-laboratory differences.
     The  net  effect of these large variances is to
render the results of the test suspect. The expected
value of the test of any  filter by  any laboratory
would fall  in a range of ± 2 standard deviations 95
percent of the time. This  factor is ± 78 for spread
plates and ±  85 for pour  plates. This variability
makes the proposed test useless.

     The  proposed test procedure permitted  the
use of either spread  or  pour plates  as the standard
at the discretion  of   the  testing laboratory. A
scatter plot of the results obtained in this test from
spread and pour plates on the same samples from
                                                135

-------
all  laboratories is shown in Figure  1, although
higher  counts were obtained using spread plates.
Statistical analysis of the data  indicated  that the
variability  of the data  from  the  two  methods
was not statistically significant.
   200
 O
 o

   100
 LLJ
 cc
 Q_
 CO
      0                 100               200
             POUR PLATE COUNT

       Figure 1. Scatter Plot of Spread and
                Pour Plate Results.

                CONCLUSIONS

     We can reach the following conclusions based
on the data analyses:

     1.   There are differences between filters.

     2.   Different results are reached in different
         laboratories although  filters behave simi-
         larly in all laboratories.

     3.   Unknown causes contributed  markedly
         to  the variability of  test results. This
         could  be due to  differences in bacterial
         population  and  pollutants  in the  test
         samples.

     4.   There  is no evidence that pour or spread
         plates  are  more  variated. Spread plates
         have a higher recovery than  pour plates.

     5.   We do not have a satisfactory test.
            RECOMMENDATIONS

     The test as presently set up is not satisfactory.
Since we  cannot separate the  effects of samples
from  the  effect of laboratories with  the  present
data, we cannot be sure if the variability is due to
differences between  technique  or to  sample dif-
ferences. If the same samples  could be run  at a
limited  number of  laboratories  with a  limited
number of filters, these effects could be clarified.
This would require  either a mixed pure  culture
approach or the shipping of samples by air from a
common point  with all  its inherent problems.
Neither proposal is ideal.

     The procedure  should  be  rewritten  to  leave
nothing to the discretion of the person running
the test before another round  robin is proposed.
A  decision should be made  limiting the standard
to  either  spread or  pour plates to  reduce the
amount of work required.
                REFERENCES

1.   Standard Methods for the Examination of
     Water and Wastewater,  13th  Ed., Am. Pub.
     Health Asso., Nqw York, N.Y., 1971.
2.   Levin, G.W., V.L. Strauss, W.C. Hess. Rapid
     Coliform Organism Determination With C14,
     J.  Water Pollut.  Cont.  Fed.  33:1021-1937,
     1961.
3.   Hufham, James B. Evaluating the Membrane
     Fecal Coliform Test by Using Escherichia coli
     as  the Indicator Organism.  Appl. Microbiol.
     27:771-776, 1974.
4.   Presswood, W.C.,  and  L.R.  Brown. Compari-
     son  of   Gelman  and  Millipore  Membrane
     Filters for Enumerating Fecal Coliform Bac-
     teria. Appl. Microbiol. 26:332-336, 1973.
5.   Dutka,  B.J.,  M.J. Jackson,  and J.B. Bell.
     Comparison  of  Autoclave  and   Ethylene
     Oxide-Sterilized  Membrane Filters  Used in
     Water Quality  Studies. Appl.  Microbiol. 28:
     474-480, 1974.
6.   Schaeffer, D.J.,  M.C.   Long,  and K.G. Jan-
     ardan. Statistical  Analysis  of the  Recovery
     of   Coliform   Organisms  on  Gelman  and
     Millipore Filters.  Appl.  Microbiol.  28:605-
     607, 1974.
7.   Interim  Federal  Specification, NNN-d00370
     (DSA-DM) dated 13, April, 1965.
8.   Specification  NIH-01-119, dated  March 25,
     1071.
                                              136

-------
Table I. Fecal Coliform Recovery Data
L
A
B
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
I
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
H20
SPL
TYP
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-2
S-2
S-2
S-2
S-2
S-2
S-?
S-2
S-2
S-2
S-2
S-2
S-2
S-2
S-2
S-2
S-2
S-2
M
F
G
1
1
1
1
2
2
2
2
3
3
3
3
4
4
5
5
5
5
£
6
6
6
1
i
1
1
2
2
2
2
3
3
3
3
4
4
5
5
5
5
S
T
D
GA
GA
PA
PA
GA
GA
PA
FA
GA
GA
PA
PA
GA
GA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
GA
GA
PA
PA
T
S
T
PP
SP
PP
SP
PP
SP
P?
SP
PP
SP
PP
SP
PP
S?
PP
S?
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PO
SP
OP
SP
PP
SP
P?
SP
PP
SP
*
*
«
ft
*
ft
ft
*
*
ft
ft
ft
ft
ft
ft-
ft
ft
ft
ft
ft
ft
*
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
RECOVERY - PERCENT OF STO
ALL DATA * SELECTED DATA
AVG ERR * AVG ERR
92.92 57*32 * 83.33 28.46
80.15 43.27 * 88.49 27.87
46.01 4S.66 *
39.69 38.92 *
12.38 9.97 *
10.68 7.77 *
77.87 68.70 *
67.17 54.08 *
114.15 54.80 * 86.00 32.38
93.47 39.68 * 91.32 31.96
139.82 113.48 * 105.33 74.58
120.61 88.59 * 111.85 76.23
69.91 65. 86 *
60.30 52.16 *
111.50 49.28 * 84.00 28.42
96.18 35.10 * 89.20 27.82
107.07 6!f>.P>6 * 87.30 40.13
92.36 5C.30 * 92.92 40.15
69. C2 54. f^ ft 71,66 19*24
59.54 42.54 * 76.10 18.42
109.73 98.65 * 104.44 72.84
94.65 77.80 * 110.91 74.41
42.94 20.14 * 35.79 10.13
29.13 15.84 *
33.76 19.64 * 39.52 16.14
22.89 15.03 *
3.63 1.18 *
2.46 0.99 *
32.47 10.95 * 35.52 8.83
22.02 9.08 *
66.66 10.36 * 61.43 6.22
45.21 10.42 ft
60.68 15.86 * 53.41 11.14
41.15 13.84 *
36.53 18.73 * 41.39 16.82
24.78 14.56 *
48.07 23.60 * 39.79 12.74
32.60 18.45 *
1)8.76 21.02 # 44.33 24.76
39.85 17.25 *
*
#
ft
#
*
*
ft
ft
ft
»
ft
*
#
#
ft
*
«
*•
*
«
#
Jt
*
«•
ft
ft
«
#
#
ft
*
«
*
«
»
ft
«
ft
*
*
ft
*
*
CODES -

H20
SPL TYP COLUMN
TST/STD COLUMNS:












: P = POLLUTED WATER S = SEWAGE
GA = GRIDDED MEMBRANE ON AGAR
PA = PLAIN MEMBRANE ON AGAR
SP = SPREAD PLATE, PP = POUR PLATE




                    137

-------
               Table I. Fecal Coliform Recovery Data (Cont'd.)
L
A
B
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
H20
SPL
TYP
S-2
S-2
S-2
S-2
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P- 1
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
M
F
G
6
6
6
6
1
1
1
1
2
2
2
2
3
3
3
3
4
/;
5
5
5
5
6
6
6
6
1
1
1
1
2
2
2
2
3
3
3
3
4
S
T
D
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
GA
GA
PA
DA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
T
S
T
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
£P
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
pp
SP
PP
SP
PP
SP
PP
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
*
ft
ft
ft
*
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
*
ft
ft
ft
*
RECOVERY -
ALL
AVG
9.40
6.37
33.76
22.89
68.55
68.55
44.92
44.92
44.63
44.63
bl.59
bl . 59
74.20
74.20
76.66
76.66
bO.72
bOo 72
62.17
62.17
65.21
65.21
bO.OO
bO.CO
62.60
62.60
62.61
62.33
72.52
72.19
b6.75
b6.50
68.46
68.16
119.81
119. ?8
90.99
90.58
72.97
DATA
ERR
4.70
3.66
11.25
9.35
45.89
24.87
47.50
33.72
27.25
13,56
34.81
18.99
44.84
22.09
47.81
24,30
37,12
21.57
37.32
18.25
46.16
26.16
40.26
24.93
37.76
18.56
22.14
21.82
30.30
29.91
17.57
17.29
22.76
22.42
61.68
60.97
29.73
29.28
26.86
PERCENT OF STD *
ft
*
#
#
ft
ft
*
ft
ft
ft
ft
#
*
ft
ft
»
#
ft
*
a
#
*
#
ft
ft
ft
ft
*
ft
ft
ft
ft
#
«
*
ft
*
#
»
ft
*
SELEC
AVG


33.76



39.64

58.57













53.57



62.61
62.33
72.52
72.19
56.75
56.50
68.46
63.16
103.60
103.13
90.99
90.58
72.97
TED DATA
ERR


11.25



14.72

5.84













12.35



22.14
21.82
30.30
29.91
17.57
17.29
22.76
22.42
37.01
36.48
29.73
29.28
26.86
*
#
*
#
«
*
*
«
*
ft
*
*
*
«
#
«
«
*
#
«
*
*
*
*
#
*
ft
*
*
«
*
#
ft
*
#
*
*
ft
*
ft
ft
CODES -
     H20  SPL  TYP COLUMN:
     TST/STD  COLUMNS:
GA
PA
SP
= POLLUTED  WATER  S = SEWAGE
GRIDDED MEMBRANE ON AGAR
PLAIN MEMBRANE ON AGAR
SPREAD PLATE,   PP = POUR  PLATE
                               138

-------
Table I. Fecal Coliform Recovery Data (Cont'd.)
L
A
B
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
H20
SPL
TYP
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-4
P-4
P-4
P-4
P-4
P-4
P-4
P-4
M
F
G
4
5
5
5
5
6
6
6
6
1
1
1
1
2
2
2
2
3
3
3
3
4
4
5
5
5
5
6
6
6
6
1
1
1
1
2
2
2
2
S
T
D
GA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
T
S
T
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
FP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
*
*
*
*
*
*
*
*
*
ft
*
*
*
*•-
*
*
*
*
*
*
•*r
*
*
*
*
*
*
*
*
*
*
*
*
*
#
*
*
*
*
*
*
*
RECOVERY - PERCENT OF STD
ALL
AVG
72.64
79.72
79.37
90.99
90.58
44.14
43.94
71.17
70.85
110.06
84.69
84.75
65.21
79.26
60.99
100.00
76.95
i 1 0 i 3 6
84.92
114.32
87.97
80.18
61.70
92.68
71.31
90.24
69.44
65.54
bO.43
96.95
74.60
37.32
25.27
38.52
26.08
6.13
4,18
42.71
28.91
DATA
ERR
26.48
25.71
25.31
26.02
25.58
15.44
15.21
25.84
25.47
51.02
43.29
26.66
23.62
23.96
21.34
35.79
31.20
26.46
25.94
27.63
25.45
35.63
30.35
26.11
23.49
22.37
20.52
27.39
23.48
23.87
21.92
14.88
8.01
16.74
9.20
9.26
5.93
21.57
12.24
#
*
#
*
#
#
*
#
*
*
*
*
*
*
#
*
*
*
#
*
#
#
#
«•
*
#
#
*
*•
*
*
»
#
*
*
*
*
*
jt,
*
#
SELECTED DATA
AVG
72.64
79.72
79.37
90.99
90.58
48.79
48.57
71.17
70.85
89.17
75.97
75.71
64.50
74.69
63.63
80.03
68.18




72.02
61.36
81.55
69.48
87.27
74.35
65.54
55.84


44.70

46.13



51.15

ERR
26.48
25.71
25.31
26.02
25.58
13.77
13.53
25.84
25.47
17.41
10.20
18.38
11.73
17.58
11.09
15.95
9.44




27.41
19.61
14.04
7.73
19.06
11.71
27.39
19.93


10.90

12.90



17.91

*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
#
#
*
#
#
*
*
*
*
*
*
*
*
*
*
*
#
*
*
*
*
#
*
*
#
*
*
CODES -

H20
SPL TYP COLUMN
TST/STD COLUMNS:
: P =
GA = GR
POLLUTED
WATER S
IDDED MEMBRANE ON
PA = PLAIN MEMBRANE
SP = SPREAD PLATE,
= SEWAGE
AGAR


ON AGAR
pp =
POUR PLATE

                 139

-------
               Table I. Fecal Coliform Recovery Data (Cont'd.)
L
A
B
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
2
2
2
H20
SPL
TYP
P-4
P-4
P-4
P-4
P-4
P-4
P-4
P-4
P-4
P-4
P-4
P-4
P-5
P-5
P-5
P-5
P-5
P-j>
P-5
P-5
P-5
P-5
P-5
P-5
P-5
P-5
P-5
P-5
P-5
P-5
P-5
P-5
P-5
P-5
S-2
S-2
S-2
V
F
G
3
3
3
3
4
4
5
5
6
6
6
6
1
1
1
1
2
2
2
2
3
3
3
3
4
4
5
5
5
5
6
6
6
6
1
1
1
S
T
D
GA
GA
PA
PA
GA
GA
GA
GA
GA
GA
PA
PA
GA
GA
PA
PA
GA
CA
PA
PA
GA
GA
PA
PA
GA
GA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
T
S
T
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
S?
Dp
SP
PP
i,p
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
pp
SP
PP
*
*
*
*
*
*
#
*
*
*
*
*
*
*
*
#
*
*
*
*
*
*
*
*
#
*
*
*
*
*
#
*
*
#
*
*
#
*
*
*
RECOVERY -
ALL DATA
AVG
t>2.49
35.54
50.89
34.45
43.51
29.45
48.90
33.10
33.33
22.56
35.92
25.00
83.50
31.21
34.09
12.0?
24.52
5.64
115.70
40.81
125.28
44.18
111.11
39.18
87.73
30.94
102.29
36.08
109.19
38.51
76.24
26.89
62.45
22.02
lt>2.00
145.01
117.06
ERR
19.18
10.08
19.56
10.43
15.66
8.19
23.84
13.43
17.32
9.83
22. RO
13.39
25.21
11.47
25.33
9.93
1C*5<.
4.43
17.79
9.64
34.37
15.77
38.37
16.77
22.77
10.58
23.55
11.28
28.59
13.26
28.38
12.23
222.23
80.20
30.00
29.33
44.56
PERCENT OF STD #
* SELECTED DATA *
* AVG
* 62.86
*
# 60.95
*
* 52.11
*
* 54.68
*
* 39.92
#
* 50.79
*
*
*
*
*
#
•1C
#
#
*
*
*
*
*
#
*
*
*
*
*
*
*
*
*
*
*
ERR *
13.23 *
*
13.98 »
*
10.68 *
*
17.52 *
*
14.56 *
*
12.36 *
*
*
«
*
*
*
if
#
*
*
#
*
*
#
*
*
*
*
*
*
*
*
*
*
*
*
CODES -
     H20 SPL  TYP COLUMN:   P  =  POLLUTED WATER   S  =  SEWAGE
     TST/STD  COLUMNS:  GA  =  GRIDDED MEMBRANE ON  AGAR
                        PA  =  PLAIN MEMBRANE ON  AGAR
                        SP  =  SPREAD PLATE,  PP  =  POUR PLATE
                              140

-------
               Table I. Fecal Coliform Recovery Data (Cont'd.)
L
A
B
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
7.
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
H20
SPL
TYP
S-2
S-2
S-2
S-2
S-2
S-2
S-2
S-2
S-2
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-3
P-3
P-3
P-3
P-3
M
F
G
1
3
3
3
3
6
6
6
6
1
1
1
1
3
3
3
3
6
6
6
6
1
i
1
1
3
3
3
3
6
6
6
6
1
1
1
1
3
S
T
D
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
T
S
T
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
ft
RECOVERY -
ALL DATA
* AVG
#
ft
*
*
«
*
ft
ft
*
*
ft.
*
*
*
ft
«•
ft
*
*
*
*
ft
«•
»
*
ft
ft
ft
ft
ft
*
#
*
#
*•
*
*
*
111
135
125
120
110
100
79
120
95
179
113
1*4
93
182
133
167
127
301
293
2*8
2*6
164
147
1*5
140
271
2*1
101
94
190
165
205
178
764
126
749
123
1275
.68
.59
.13
.05
.79
.69
.65
.25
.12
.60
.88
.97
.?&
.47
.57
.43
.15
,30
.96
.11
.10
.48
.22
.55
.12
.84
.74
.55
.04
.08
• oc
.50
.39
.70
.21
.01
.62
.86











1

1



1

1



1
1
2
1




1

ERR
43.
21.
27.
40.
43.
41.
40.
28.
30.
88.
29.
22.
54.
11.
44.
90.
32.
52.
69.
07.
37.
84.
87.
02.
03.
02.
09.
80.
45.
95.
37.
31.
64.
433.

3

69.
69.
59.
922.
06
69
16
30
52
54
01
26
89
66
43
02
26
64
75
64
10
02
88
87
68
18
21
77
28
68
13
43
13
81
27
25
31
57
66
43
12
44
DE
*
*
ft
#
ft
ft
ft
#
*
*
*
ft
»
*
ft
ft
#
*
*
*
*
#
*
ft
#
#
*
ft
ft
ft
#
*
«
#
*
#
*
*
#
«
                                            :E/NT  OF STD       *
                                            SELECTED DATA    *
                                               AVG      ERR    *
                                                              *
                                                              «
                                                              *
                                                              #
                                                              #
                                                              *
                                                              *
                                                              «
                                                              *
                                                              *
                                                              *
                                                              *
                                                              *
                                                              *

                                                              *
                                                              *
                                                              #
                                                              *
                                                              *
                                                              *
                                                              ft
                                                              «
                                                              ft
                                                              *
                                                              *
                                                              ft
CODES -
     H20 SPL  TYP  COLUMN:  P = POLLUTED  WATER  S = SEWAGE
     TST/STD  COLUMNS:   GA = GRIDDED  MEMBRANE ON AGAR
                        PA = PLAIN MEMBRANE ON AGAR
                        SP = SPREAD PLATE,   PP = POUR PLATE


                              141

-------
               Table I. Fecal Coliform Recovery Data (Cont'd.)
L
A
B
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
H20
SPL
TYP
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-4
P-4
P-4
P-4
P-4
P-4
P-4
P-4
P-4
P-4
P-4
P-4
S-l
5-1
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-2
S-2
S-2
S-2
S-2
V
F
G
3
3
3
6
6
6
6
1
i_
i
i
3
3
3
3
6
6
6
6
1
1
2
2
3
3
4
4
5
5
6
6
1
1
2
2
3
S
T
D
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
PA
PA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
T
S
T
SP
PP
SP
PP
SP
PP
SP
pp
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
*
*
*
*
*
*
#
«•
*
#
#
*
*
#
*
•*
•*
,t
•H-
*
*
tt
*
*
•w
*
»
•R-
#
*
#
*
#
#
*
*
*
«
*
RECOVERY -

ALL
AVG
187
1117
164
1516
237
1530
239
391
lb8
117
47
447
169
840
318
375
248
349
231
83
92
95
105
96
107
DO
56
68
76
227
2t>2
92
S8
107
103
118
.81
.24
.46
.27
.09
.23
.27
• 86
.96
.44
.64
• 56
.90
.24
• 9fl
.18
.30
.63
i f°*
. 4w
.60
.72
.08
.45
.7?
.27
.81
.36
.85
.36
.86
.72
.07
.15
• 92
.31
.81
DATA

ERR
55
856
56
1196
115
1221
119
136
88
127
61
426
137
702
221
439
93
507
157
41
55
2S
43
29
^4
20
28
23
34
132
174
38
35
30
27
36
.91
.65
.15
.57
.82
• 86
.12
.18
.53
.21
.58
.09
.35
.96
.43
.85
.97
.80
.03
.21
.67
.70
.16
.91
.69
.17
.42
.28
.02
• **6
.04
.19
.26
.57
.75
.56
PERCENT CF STD
* SELECTED DATA
* AVG ERR
*
#
*
#
*
*
*
*
* 113.26 30.49
#
*
* 196.55 52.80
* 113.43 55.21
*
•*
*
*•
#
*
*
«•
*
*
#
#
*
*
«•
#
*
*
* 92.07 38.19
* 88.15 35.26
* 107,92 30.57
* 103.31 27.75
* 118.51 36.56
*
*
#
*
*
*
*
*
#
*
*
*
#
#
*
#
*
*
*
#
*
*
*
#
*
*
*
*
*
*
«
*
*
#
#
#
*
*•
#
CODES -
     H20 SPL  TYP COLUMN:   P  =  POLLUTED WATER   S  = SEWAGE
     TST/STD  COLUMNS:  GA  =  GRIDDED MEMBRANE  ON  AGAR
                        PA  =  PLAIN MEMBRANE ON AGAR
                        SP  =  SPREAD PLATE,  PP =  POUR PLATE
                              142

-------
                Table I. Fecal Coliform Recovery Data (Cont'd.)
L
A
8
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
H20
SPL
TYP
S-2
S-2
S-2
S-2
S-2
S-2
S-2
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-i
P-l
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-3
P-3
P-3
P-3
P-3
P-3
P-3
M
F
G
3
4
4
5
5
6
6
1
1
2
2
3
3
4
4
5
5
6
6
1
1
2
2
3
3
4
4
5
5
6
6
1
1
2
2
3
3
4
S
T
D
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
T
S
T
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
pp
SP
pp
SP
pp
SP
pp
SP
pp
SP
pp
SP
pp
SP
pp
SP
pp
SP
pp
SP
pp
SP
p?
SP
?p
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
*
ft
ft
RECOVERY -
ALL
AVG
113.74
103.46
99.05
103.46
99.05
82.17
73.67
114.59
122.47
87.12
93.11
102.14
109.17
83.69
89.44
88.84
9^.95
109e 01
116.51
76.74
72.52
103.43
97.80
111.62
105.49
73.25
69.23
89.53
84.6,1
P.O. 23
75.82
107.65
161.06
106.63
Ib9. 5^*
119.89
179.33
79.08
DATA
ERR
33.33
50.23
46.63
50.22
46.63
31.84
29.32
39.13
45.54
32.22
37,27
31,33
36,80
32.92
37.90
37.90
43.40
2 6 i 9 2
43.06
17.20
19.20
26.05
28.59
22.03
25.11
19.52
21.26
13*19
15.91
22.23
24.09
45.04
24.36
52.29
26.12
64.6**
49.35
41.79
PERCENT OF STD *
*
*
ft
ft
#
*
#
*
ft
ft
ft
ft
#
ft
ft
•a-
ft
ft
#
«.
ft
ft
ft
#
ft
ft
ft
ft
ft
ft
#
»
ft
»
#
*
*
#
#
*•
SELEC
AVG
113.74
103.46
99.05
103.46
99.05
82.17
78.67
96.56
103.21
87.12
93.11
102.14
109,17
83.69
89.44
88.84
94.95
1 ri -a A o
110.09





99.18






107.65
161,06
106.63
159.54
110.96
166.03
79.08
TED DATA
ERR
33.33
50.23
46.63
50.23
46.63
31,84
29.32
14.28
18.40
32.22
37.27
31.33
36,80
32.92
37. 9C
37.90
43,40
2 4 • ? 9
40.64





6.23






45.04
24.86
52.29
36.12
53.56
36.7^
41.79
ft
*
#
ft
#
ft
*
»
ft
«
«
«
#
ft
ft
ft
ft
*
#
JC
ft
ft
ft
*
*
ft
ft
ft
ft
ft
*
ft
*
ft
ft
ft
ft
ft
ft
ft
CODES -
     H20 SPL  TYP COLUMN:
     TST/STD  COLUMNS:  GA
                        PA
                        SP
P = POLLUTED  WATER  S = SEWAGE
= GRIDDED MEMBRANE ON AGAR
= PLAIN MEMBRANE ON AGAR
= SPREAD PLATE,   PP = POUR PLATE
                               143

-------
Table I. Fecal Coliform Recovery Data (Cont'd.)
L
A
B
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
H2C
SPL
TYP
P-3
P-3
P-3
P-3
P-3
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-l
5-1
S-l
S-2
S-2
S-2
S-2
S-2
S-2
S-2
S-2
S-2
S-2
S-2
S-2
P-l
p-1
P-l
P-l
P-l
P-l
P-l
M
F
G
4
5
5
6
6
1
1
2
2
3
3
4
4
5
C,
^
6
6
1
1
2
2
3
3
4
4
5
5
6
6
1
1
2
2
3
3
4
5
T
D
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
T
S
T
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
pp
SP
PP
SP
PP
SP
PP
*
*
*
*
ft
*
ft
*
*
ft
*
*
ft
ft
•fr
«•
ft
*
*
*
ft
ft
ft
ft
*
*
ft
*
ft
*
*
ft
*
*
*
*
tt
*
*
RECOVERY -
ALL
AVG
118.32
79.08
118.32
61.73
92.36
74.48
99.31
57.14
76.19
75.00
100.00
62.75
83.67
68.37
91.93
55.61
74.14
58.67
77.96
45.40
60.33
68.36
90. 34
i>0.00
66.44
51.53
68.^7
52.04
69.15
94.09
85.35
12.42
11.26
102.48
92.95
105.59
DATA
ERR
31.29
41.79
31.29
34.87
27.79
37.93
36.47
29.60
28.64
33.30
30.23
3?. 76
31. 50
30.95
28.23
32.50
32.80
34.26
30.30
27.24
24.41
32.16
24.99
25.91
21.46
25.46
20.46
23.43
17.63
28.82
34.98
13.77
13.66
35. ?5
42.15
27.72
PERCENT OF STO
*
*
*
*
*
#
#
»
#
#
ft
*
*
*
*
ft
ft
*•
it
#
#
#
*
»
#
»
ft
ft
#
»
#
ft
#
#
#
»
*
#
SELECTED
AVG
118.32
79.08
118.32
66.32
99.23
74.48
99.31
60.58
80.78
75.00
100.00
66.96
69.28
68,87
91.33
60. 5R
80. 7S
68.02
90.29
58.67
77.96
68.36
90.84
56.12
74.57
59.94
79.66
58.67
77.96
81.52
83.33


85.40
87.30

DATA
ERR
31.29
41.79
31.29
33.63
24.12
37.93
36.47
28.56
26.61
33.30
30.20
30.68
28.22
30.95
28.23
29.98
28.50
32.26
25.21
19.49
10.68
32.16
24.99
21.20
13.60
25.33
18.10
19.49
10.68
27.41
23.43


18.16
13.75

ft
*
#
»
ft
*
ft
#
ft
»
#
ft
#
ft
*
»
ft
#
»
«
*
*
#
ft
»
•»
ft
*
#
«
ft
»
*
»
»
#
#
*
«
CODES -
     H20 SPL  TYP COLUMN:
     TST/STD  COLUMNS:  GA
                        PA
                        SP
              P  =  POLLUTED WATER   S  =  SEWAGE
              =  GRIDDED MEMBRANE ON  AGAR
              =  PLAIN MEMBRANE ON  AGAR
              =  SPREAD PLATE,  PP  =  POUR PLATE
                 144

-------
               Table I. Fecal Coliform Recovery Data (Cont'd.)
L
A
B
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
H20
SPL
TYP
P-l
P-l
P-l
P-l
P-l
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-4
P-4
P-4
P-4
P-4
P-4
0.4
P-4
P-4
V
F
G
4
5
5
6
6
1
1
2
2
3
3
4
4
R
5
6
6
1
i
2
2
3
3
4
4
5
5
6
6
1
1
2
2
3
3
4
4
5
5
^
D
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
T
S
T
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SD
PP
SP
PP
SP
= p
SP
PP
SP
pp
SP
PP
*
*
*
*
*
*
*
*
*
*
*
#
*
•&
*
*
*
*
a
•a-
#
*
*
*
#
#
*
*
#
*
*
#
*
*
«
*
*
#
*
*
*
RECOVERY -
ALL
AVG
95.77
65.52
i>9.43
66.14
60.00
126.5*
40.28
139.82
44. 5C
114.15
36.33
100.00
31.83
96.46
30.70
112.38
35.77
144.73
46.47
111.40
35.77
142.10
45.63
86.84
27.88
121.05
38.87
122.80
39.43
214.03
lt»4.43
142.10
102.53
317.54
229.11
223.07
164.55
238.59
DATA
ERR
35.06
16.91
21.49
22.05
26.21
68.27
25.65
50.95
20.55
31.77
13.65
29.48
12.48
35.27
14.22
39.22
15.97
44.96
20.39
60.65
24.06
42.11
19.37
25.52
11.77
30.26
14.70
58.72
23.91
58.72
110.42
55.25
85.05
100.36
173.38
56.72
113.45
72.50
PERCENT OF STO
*
*
*
*
*
#
*
*
*
#
*
*
*
*
*
*
*
*
*
#
*
*
*
*
*
*
#
*
*
#
*
*
*
*
*
*
*
*
*
*
SELEC
AVG

65.52
66.98
66.14
67.61
134.04
50.00
134.46
50.15
109.78
40.95
100.00
37,30
103.54
33.62
114.89
42.85
137.50
52.38
129,16
49.20
135.00
51,42
91,66
34,92
115,00
43.80
131.25
50.00
203.33
112.61
135.00
74.76


216.66
120.00
216.66
TED DATA
ERR

16.91
13.59
22.05
13.81
62.56
20.77
47.65
15.20
29.45
8.83
25.86
7.73
31.23
9.67
32.39
9.89
34.39
14.67
50.00
20.52
31.84
13.67
11.13
5.29
21.79
9.61
36*44
15.38
43.48
47.27
44.32
39.94


40.77
47.29
51.83
*
*
«
*
*
*
*
*
*
#
*
*
*
«
•»
*
#
#
*
•w-
*
*
*
#
#
*
#
*
*
#
*
#
*
*
#
#
*
*
*
*
*
CODES -
     H20  SPL  TYP COLUMN:
     TST/STD  COLUMNS:  GA
                        PA
                        SP
P = POLLUTED  WATER  S = SEWAGE
= GRIDDED MEMBRANE ON AGAR
= PLAIN MEMBRANE ON AGAR
= SPREAD PLATE,   PP = POUR  PLATE
                               145

-------
               Table I. Fecal Coliform Recovery Data (Cont'd.)
L
A
B
4
4
4
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
H20
SPL
TYP
P-4
P-4
P-4
S-l
S-l
S-l
S-l
S-l
S-l
S-l
S-?
S-2
S-2
S-2
S— ?
S-2
P-l
P-l
r>- 1
P-l
P-l
P-l
P-2
P-2
P-2
P-2
P-2
P-2
P-3
P-3
P-3
P-3
P-3
P-3
M
F
G
5
6
6
1
1
2
3
4
5
6
1
2
3
A
tj
6
1
2
3
A
5
6
1
2
3
A
5
6
1
2
3
A
5
6
S
T
D
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
T
S
T
SP
PP
SP
PP
SP
SP
SP
SP
SP
SP
PP
PP
PP
PP
PP
p?
SP
SP
SP
SP
SP
SP
SP
SP
SP
SP
SP
SP
SP
SP
SP
s°
SP
SP
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
*
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
RECOVERY -
ALL
AVG
172.15
289.47
208.86
763.15
100.11
76.00
131.46
66.51
9A.76
122.90
71.31
5A.28
SB, 16
35.15
1 n •* e r, c,
81./>5
156. Al
159.85
211.12
191.78
259.50
297.80
129.23
158. A7
196.61
133.24
186.77
157.89
98.13
10A.62
136.71
107.32
103.02
152.77
DATA
ERR
128.17
75.17
1A6.28
205.21
A0.85
23. 1A
A9.65
A6.00
18.08
40,18
24.42
20.42
23*01
27.27
32s 00
41 .48
37. 3A
113.77
115.59
84.91
155.43
100.47
70.02
29.71
78.66
28.25
69.31
56.53
33.16
53.62
47.19
22.63
32.60
6A.61
PERCENT OF S
#
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
ft
#
*
ft
*
«.
it
ft
ft
ft
*
«
*
*
ft
»
*
ft
ft
#
*
ft
*
*
ft
SELECTED
AVG
120.00














73.68






116.77
158. A7
161. ?6
133.24
172.19
147.36
98.13
92.59
119.16
107.32
108.02
136.38
TD
DATA
ERR
53.42














1'A.IC






53.49
29.71
33.^0
28.25
53.43
^3.50
33.16
40.43
29.45
22.63
32.60
30.07
ft
ft
*
*
*
*
*
ft
ft
ft
ft
«
ft
«
«
«
ft
«.
*
»
ft
*
*
#
#
*
«
*
«
ft
*
ft
ft
«
#
«
*
CODES -
     H20  SPL  TYP COLUMN:
     TST/STD  COLUMNS:  GA
                        PA
                        SP
P = POLLUTED  WATER  S = SEWAGE
= GRIDDED MEMBRANE ON AGAR
= PLAIN MEMBRANE ON AGAR
= SPREAD PLATE,   PP = POUR  PLATE
                              146

-------
                Table I. Fecal Coliform Recovery Data (Cont'd)
L
A
B
5
5
5
5
5
5
6
6
6
6
6
6
6
6
&
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
7
7
7
H20
SPL
TYP
P-4
P-4
P-4
P-4
P-4
p-4
S-2
S-2
S-2
S-2
S-2
S-2
Sr2
S-2
S-2
S-2
S-2
S-2
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
S-2
S-2
S-2
M
F
G
1
2
3
4
5
6
1
1
2
2
3
3
4
4
5
5
6
6
1
1
2
2
^
3
4
4
5
5
6
6
1
2
3
5
T
D
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
G A
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
T
S
T
PP
PP
PP
PP
PP
PP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
PP
SP
DP
SP
PP
SP
PP
SP
pp
SP
pp
PP
pp
*
*
#
*
*
*
#
*
*
*
#•
*
*
*
*
*
*
1C
*
*
*
#
*
#
*
#
*
*
*
#
*
*
*
*
*
*
RECOVERY -
ALL
AVG
78.49
80.89
108.89
87.09
102.51
100.74
97.68
106.28
80.30
53.80
128. P7
95.00
78.12
75.75
9 s * 0 3
82.67
126.31
124.85
81.84
92.11
4.89
4.37
121.44
111.71
b8. 22
66.02
65.33
89.13
131.74
2bl.29
86.07
5.52
92.65
DATA
ERR
11.91
22.11
29.63
23.68
24.85
17.06
38.12
43.17
40.20
39.63
59.05
46.72
46.85
52.36
30.92
26.92
53.47
94.07
83.86
133.94
11.65
12.00
91 .54
126.30
22. *7
34.48
39.53
93.34
64.40
197.60
10.30
1.86
10.30
PERCE.N
T OF
STD
* SELECTED DATA
*
* 78
* SO
* 108
* 87
* 102
* 100
* 97
* 106
* 97
* 69
* 120
* 95
* 96
* 93
* 9S
* 82
* 126
* 111
*
*
*
*
*
*
* 53
* 66
* 66
* 76
#
#
*
*
AVG
.49
.89
.39
.09
.51
.74
.68
.28
.34
.79
.90
.00
.35
.43
*03
.87
.31
.48






.22
.02
.93
.71




ERR
11.91
22.11
29.63
23.68
24.85
17.06
33.12
43.17
24.57
39.00
49.96
46.72
20.21
23.15
3 0 • 9 2
26.92
53.47
74.62






22.87
34.46
11.86
30.56




#
*
#
*
*
#
*
*
*
#
#
*
*
*
*•
*
*
A
*
*
#
*
#
#
#
*
*
*
*
*
*
*
*
*
#
*
CODES -
     H20 SPL  TYP COLUMN:
     TST/STD  COLUMNS:
   P = POLLUTED WATER  S = SEWAGE
GA = GRIDDED  MEMBRANE ON AGAR
PA = PLAIN  MEMBRANE ON AGAR
SP = SPREAD PLATE,   PP = POUR  PLATE
                              147

-------
               Table I. Fecal Coliform Recovery Data (Cont'd.)
L
A
R
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
9
9
9
9
9
9
9
9
9
H20
SPL
TYP
S-2
S-2
S-2
P-l
P-l
P-l
P-l
P-l
P-l
P-2
P-2
P-2
P-2
P-2
P-3
P-3
P-3
P-3
P-3
P-3
P-4
P-4
P-4
P-4
P-4
S-2
S-2
S-2
S-2
S-2
S-2
S-2
S-2
S-2
M
F
G
4
5
6
1
2
3
4
5
6
1
2
4
5
6
1
2
2
4
5
6
1
3
4
5
6
1
1
2
2
4
4
5
5
6
S
T
D
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
T
S
T
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
SP
PP
SP
op
SP
PP
SP
PP
*
*
*
ft
*
*
*
*
*
ft
*
ft
*
*
*
*
*
ft
ft
#
ft
»
*
ft
»
ft
*
*
*
*
ft
*
ft
ft
*
*
*
RECOVERY -
ALL
AVG
71.13
75.44
91.13
94.63
0.56
90.11
77.96
88.13
85.87
94.51
85.65
72.15
75.94
77.21
89.14
1.31
90.78
62.50
77.96
79.27
89.04
90.41
73.97
85.61
86.30
238.35
84.26
245.20
86.68
272.60
96.36
2t>4.10
89. S3
236.30
DATA
ERR
12.16
8.15
12.63
8.49
0.80
12.00
9.22
11.97
13.69
12.10
10.21
10.70
11.92
17.13
19.08
2.30
22.65
19. 5C
14.92
13.48
18.05
18.99
18.98
21.23
19.65
29.27
27.45
46*^4
34.72
44.06
35.14
57.27
38.45
50.17
PERCENT OF STD
«•
#
ft
*•
ft
ft
*
ft
»
#
ft
#
ft
ft
ft
ft
*
ft
ii-
fc
ft
#
ft
ft
ft
#
li-
ft
ft
ft
#
#
*
#
#
ft
SELECTED
AVG
63.67
73.83



84.74
77.96
81.92
76.97
94.51
85.65
72.15
75.94
77.21
86.34

86.75
62.50
77.96
79.27
89.0^
90.41
77.91
8 5-. 61
86.30









DATA
ERR
9.92
7.34



5.24
9.22
7.06
5.76
12.10
10.21
10.70
11.9?
17.13
16.77

19.09
19.50
14.92
13.48
18.05
18.99
16.43
21.23
19.65









#
ft
ft
ft
*
*
ft
*
#
*
*
*
ft
ft
»
*
#
#
«
*
#
»
ft
ft
«
ft
»
ft
ft
*
#
»
«
*
»
*
ft
CODES -
     H20  SPL  TYP COLUMN:   P  =  POLLUTED WATER   S = SEWAGE
     TST/STD  COLUMNS:  GA  =  GRIDDED MEMBRANE  ON AGAR
                        PA  =  PLAIN MEMBRANE  ON AGAR
                        SP  =  SPREAD PLATE,   PP = POUR PLATE
                              148

-------
Table I. Fecal Coliform Recovery Data (Cont'd.)
L
A
B
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
H20
SPL
TYP
S-2
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-l
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-2
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-3
P-4
P-4
P-4
P-4
P-4
P-4
M
F
G
6
1
1
2
2
4
4
5
5
6
6
1
1
2
2
4
4
5
5
6
6
1
1
2
2
4
4
5
5
6
6
1
1
2
2
4
S
T
0
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
4 GA
T
S
T
SP
PP
SP
PP
SP
pp
SP
pp
SP
pp
SP
pp
SP
pp
SP
pp
SP
pp
SP
pp
SP
pp
SP
pp
SP
pp
SP
pp
SP
pp
SP
pp
SP
po
SP
pp
SP
*
RECOVERY - PERCENT OF STD
* ALL DATA
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
#
*
#
*
*
»
*
*
*
*
*
*
*
*
*
*
#
*
*
*
*
AVG
83.53
189.72
78.02
188.35
77,46
193.83
79.71
202.05
83.09
195. 89
80.56
200.00
73.73
20s. 21
76.76
197.26
72.72
219.17
80.80
215.75
79.54
186.98
80.29
191.09
82.05
205.47
88.23
206.16
88.52
210.27
90.29
204.79
77.86
199.31
75. 7B
195.89
74.47
ERR
34.69
47.87
22.29
34.90
16.93
43.50
20.54
47.15
22.16
32.91
16.22
38.74
32.00
45.53
35.23
49.44
35.70
42.25
34.99
41.61
34.45
35.59
25.83
31.01
24.14
37.18
27.61
41.71
29.59
52.59
34.49
53.12
19.93
42.51
15.90
33.35
12,42
* SELECTED DATA
* AVG
*
* 171.23
* 70.42
* 183.21
* 75.35
* 178.08
* 73.23
* 172.94
* 71.12
* 187.21
* 76.99
* 188.35
* 79.42
* 184.93
* 77.97
* 187.50
* 79.06
* 202.05
* 85.19
* 195.20
* 82.31
* 180.65
* 77.57
* 191.09
* 82.05
* 188.35
* 80.88
* 184.93
* 79.41
* 189.49
* 81,37
* 132.64
* 69.44
* 1P.6.07
* 7C.7<*
* 191.73
* 72.91


36
17
31
15
29
14
19
10
26
13
29
24
38
28
41
29
25
23
29
25
28
22
31
24
19
18
28
22
29
23
32
12
36
13
30
11
ERR

.36
.30
.16
.33
.85
.71
.89
.55
.31
.59
.30
.66
.05
.13
• 59
.91
.25
.85
.40
.15
.47
.46
.01
.14
.02
.84
.36
.66
.60
.45
.29
.04
.03
.45
.90
.50
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
#
*
*
«•
*
*
*
*
*
*
*
#
*
*
*
*
#
*
*
*
*
*
*
#
*
CODES -

H2
0
SPL TYP COLUMN: P = POLLUTED
TST/STD COLUMNS:






GA = GRI
PA = PLA
WATER S
DDED MEMBRANE ON
= SEWAGE
AGAR


IN MEMBRANE ON AGAR
SP = SPREAD PLATE, PP =
POUR
PLATE

                   149

-------
               Table I. Fecal Co I i form Recovery Data (Cont'd.)
L
A
B
9
9
9
9
10
10
10
10
10
10
10
10
10
10
10
10
10
H20
SPL
TYP
P-4
P-4
P-4
P-4
S-2
S-2
S-?
S-2
S-2
S-2
S-2
P-l
P-l
P-l
P-l
P-l
P-l
M
F
G
5
5
6
6
1
1
2
3
4
5
6
1
2
2
4
5
6
S
T
D
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
*•
S
T
PP
SP
PP
SP
PP
SP
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
PP
*
#
*
#
•»
*
*
*
*
*
#
*
#
*
«•
#
*
*
ir
#
RECOVERY -
ALL
AVG
191.78
72.91
180.13
68.43
121.63
54.16
131.49
130.40
111.69
103.50
143.09
li>2.55
107.82
128.94
135.00
i:>2.72
ib6.57
DATA
ERR
45.45
17.03
37.04
13.85
44.90
15.63
53.12
41.07
34.16
37.12
36.19
43.56
68.74
42. 12
55.00
50.62
39.49
PERCENT OF STD
#
*
*
*
*
*
*
«•
*
#
*
*
#
#
#
#
#
*
*
SELEC
AVG
183.21
69.66
172.94
65.75
121.63
54.16
118.09
130.40
111.69
103.50
143.09
152.55



133.63

TED DATA
ERR
38.96
14.57
28.82
10.73
44.90
15. 6S
25.27
41.07
34.16
37.12
36.19
43.56



26.07

*
*
*
*
#
*
•*
*
*
#
*
*
#
#
*
*
*
*
#
*
CODES -
     H20 SPL  TYP COLUMN:  P  =  POLLUTED WATER  S  =  SEWAGE
     TST/STD  COLUMNS:  GA =  GRIDDED MEMBRANE ON  AGAR
                        PA =  PLAIN MEMBRANE ON AGAR
                        SP =  SPREAD PLATE,  PP =  POUR PLATE
                               150

-------
Table 2     ANOVA TABLE - SPREAD PLATE STANDARD
  Source
Sum of Squares
DF
Mean Squares
Filters (F)
Samples (S)
FS
Labs (L)
FL
SL
Residual
Total
10257.6269
22436.6289
9214.7500
39769.3829
7386.6250
74822.8908
32624.6289
196512.5316
5
4
20
2
10
8
40
89
2051-5253
5609.1572
460.7375
19884.6914
738.6625
9352.8613
815.6157

2.51
6.87
0.56
24.37
0.90
11.46


Significant - 97
#
*
.5%


Significant - 99+%
Not Significant
#






  Meaningless - since they refer to differences attributed to samples which we know are different.
 Table 3     ANOVA TABLE - POUR PLATE STANDARD
   Source
Sum of Squares
DF
Mean Squares
Filters (F)
Samples (S)
FS
Labs (L)
FL
SL
Residual
Total
14464.0000
21314.5000
10630.3769
47449.2579
8669.1269
104435.5158
37086.0000
244048.7816
5
4
20
2
10
8
40
89
2892.8002
5328.6250
531.5188
23724.6289
866.9125
13054.4394
927.1500

3.12
5.74
0.57
25.58
0.93
14.08


Significant - 97
*
*
.5%


Significant - 99+%
Not Significant
*






  Meaningless - since they refer to differences attributed to samples which we know are different.
                                            151

-------
Table 4    COMPONENTS OF VARIANCE        Table 5     COMPONENTS OF VARIANCE
          ANALYSIS - SPREAD PLATE                    ANALYSIS - POUR PLATE
          STANDARD                                  STANDARD
02 Residual = 7386.625 + 40011.245 = 800          a2 Residual = 8669.+ 37086 = 915
                     50                                         50

   02 I OK = 19885 - a2 Residual = RQC                              o
      L8b  	™	  b  b             a2 Lab - 23725 - a2 Residual = 760
                                                              30
   a2 Filters =  2052 - a2 Residual = 83

                     15                       a2 Fj|ters = 2893 - a2 Residual =  132
                                                               15

 a2  For a single determination on any filter
     in any laboratory is:
                                             a2 for a single determination on any filter
                                               in any laboratory is:
  a2 = a2  Residual + a2 Lab + a2 Filter =


           800   +   636  +   83      1519    a2 = a2 Residual + a2 Lab + a2 Filter =


       a = ± 39               2a = ± 78                915   +   76° +    132  = 18°7


                                               a = ± 42.5                2a  = ± 85

          GRAND MEAN = 81.5

                                                    GRAND MEAN = 95.5
                                          152

-------
            CRITIQUE ON ASTM TEST FOR RECOVERY OF FECAL COLIFORMS AND
                              PROPOSAL FOR MODIFIED METHOD

                                       Norman  H. Goddard
                                   Sartorius Membrafilter GmbH
                  ABSTRACT

     As both a manufacturer of membrane filters
and  as  a participating  laboratory,  in  the  ASTM
Collaborative Study on MF's, Sartorius (Goddard),
made the following suggestions for change in the
study plan:

1.   Conduct the  study with a flora more favor-
     able  than  polluted  waters  in sewage.   For
     example, use a potable water  spiked with  a
     standard pure culture.

2.   Standardize sample mixing conditions.

3.   Require separate counts  of  coliforms   and
     non-coliforms.

4.   Randomize sample positions in the incuba-
     tion.

5.   Verify  accuracy  of thermometer and incuba-
     tor settings.
     The requirements of a routine method for the
testing for the recovery of fecal coliforms include
the following:

     Agreement must be obtainable with a compar-
ison method such as  the pour  plate  method.

     The results must be reproducible irrespective
of the location  of the testing laboratory or of the
person performing the test.

     Evaluation must  be clear,  without the possi-
bility  of  subjective errors, so  that  the  culture
medium must not only be selective but also allow
simple  recognition  of  the  resulting colonies.


     The method itself must be simple and practi-
cal,  capable  of  use  in any  testing  laboratory.
     As  the  leading  European manufacturer  of
membrane filters, we welcome this attempt by the
American Society for Testing and Materials and the
United States Environmental Protection Agency to
standardize a method for testing the suitability of
membrane filters for fecal coliform determinations,
and for comparing  different brands of membrane
filters. We were therefore very pleased to be invited
to participate actively in this work and to carry out
the  round   robin   comparison tests  on  various
filters. Such  a  filter comparison  requires that  all
influences which  can affect the result,  excepting
the filters,  be removed. Unfortunately, we believe
that this was  not  the case with the procedure
used, and  that modifications are  required  for
meaningful  results to be obtained.
     One  point of  interest which  arose from  our
tests was,  however,  that there appeared to be some
correlation  between  the  flow rate  through  a
filter under standard conditions and the efficiency
of recovery of that  filter. The faster the flow rate,
the higher the efficiency.
     We heard yesterday that higher recoveries can
be obtained by increasing the average surface open-
ing diameter of the filter to an optimum size. Now
the surface opening diameter is directly related  to
the pore size.  The former is measured by electron
microscopy, the  latter  is characterized  by the
mercury  intrusion  method, by  retention charac-
teristics or by  flow rate measurements. In general,
the faster the  flow rate, the larger the pore size,
and  the larger the average surface opening diame-
ter.

     It was also stated yesterday that a reason for
the higher recovery with a larger pore size filter
may be the  better provision of nutrient to a micro-
organism sitting in  a  surface pocket compared  to
one  perched on the top of the filter and exposed
to different evaporation effects.
                                               153

-------
     The structure of a membrane filter is uniform,
the pore size being the same throughout the width
of the filter. Another  possibility is therefore that
the use of a larger pore size filter simply results in
an easier passage of the nutrient from the bottom
to  the top surface  of the  filter, thus favoring
growth.

     Our experiments in this direction are not yet
completed, but it could be that an optimum  sur-
face opening diameter, or pore size, for the re-
covery of fecal coliforms can  be decided upon,  and
that comparisons between competitive filters need
then be made only on the basis of the measure-
ment of physical parameters: the sizes, and  the
reproducibility of the sizes, of the surface openings
or pores on each side  of  the filter, and the flow
rate through the filter under defined conditions.

     This is purely conjecture at present, and I  will
therefore  return to  the  test  under  discussion.

     We are not happy with the samples suggested,
polluted water and raw sewage, as a number of E.
coli and coliform types can occur which affect the
reported recovery because of subjective evaluation.
Different investigators may not agree as to which
colonies are fecal coliforms.

     Additionally, such samples can contain  high
concentrations  of  sub-lethally  injured  micro-
organisms,  and  there  is  a large variation in the
samples taken at the different locations of the
investigating laboratories.

     The culture medium, M-FC agar, is apparently
not  sufficiently   selective.  Not  only coliform
colonies but also  numerous other colonies are
obtained, which may  or may not work as antago-
nists, either by affecting the growth of the bacteria
we are searching for, or by changing the pH of the
surrounding nutrient  by  alkali  formation and so
affecting blue color formation.

     We filtered lake  water  through  a filter  and
cultivated as usual. We then selected four of the
blue colored colonies and  three different non-blue
colonies and subcultured them.

     We streaked the blue colony culture on each
of  three different M-FC agar plates and streaked
the culture of the non-blue  colonies vertically to
the blue colonies, one  on  each of the three plates.

     On incubating, differing effects of the non-
blue on the blue colonies could be seen. This shows
that according to  the  type and  the amount of
alkali producing  bacteria which  are  present, the
typical blue color of the coliforms may be weak-
ened or even completely prevented from occuring.

     Fewer non-col if orms appear to grow on Endo
media  than  on M-FC and strong alkali producers
appear to have less effect.

     Returning  to the  present  procedure,  the
length  of incubation using M-FC agar appears to be
insufficient, as we have often  noted blue color
developing after completion of the standard incu-
bation  time.  The  incubation  temperature  of
44.5 C.  is  difficult to  maintain  in  routine  use.

     With regard to the reference method, using
samples containing high concentrations of antago-
nists, their effect depends on the contact area with
the nutrient. The relationship of sample volume to
contact area should be  the same with  the filter
and the reference method, i.e. smear plates should
be of the same diameter as the filter diameter used.
Pour plates  of the same diameter have a far larger
contact  area,  and  conditions within  the culture
medium  may affect different antagonists in differ-
ent ways.

     The sample suggested, polluted  waters, con-
tain bacterial agglomerates.  Shaking will dispense
them to  an  extent, depending on the amount and
method  of  shaking.  Some aggregate  break-up
occurs in the pour  plate method on mixing the
sample with the warm viscous agar.

     We  would therefore suggest that a sample be
used which  contains  favorable bacteria  flora for
this test. We recommend one in  which the numer-
ous types of protein  decomposing, alkali forming
bacteria  occurring in raw sewage do not occur, so
that the difference  in contact area between  test
and reference method is not important.

     Such a  sample could be potable water with
added pure culture.

     The conditions of shaking the sample are less
important with potable water than with polluted
water  but  should  nevertheless be  standardized.
     M-FC agar would be suitable as culture med-
 ium for this sample, whereas with a sample con-
 taining a bacteria spectrum approaching that of
 raw sewage a more  specific culture medium is re-
 quired. Regardless of which culture  medium  is
 used, the total number of non-coliform colonies
                                                154

-------
occuring should  be  stated  in  addition  to the
coliform count.

     Again,  using such a  sample, the  incubation
time of 22-24 hours would  be sufficient and dif-
ficulty in evaluation would not occur.

     Streaked spread  plates  of the same diameter
as the filter would be optimal, but not practical be-
cause of the  small sample  volume which can be
spread.  Pour plates in 60  mm.  Petri dishes should
therefore be used with a good relationship between
sample volume and contact area.

     Randomization of sample positions during
incubation is necessary, as is a check on the accur-
acy of the thermometer used.
         QUESTIONS AND ANSWERS

Geldreich:   Was rosolic acid added to the medium
            which you used?

Goddard:    (I don't really know.)  I  must apolo-
            gize, I am  not a microbiologist and if
            I  were timid, I  would  not have read
            this technical paper, but we did want
            to comment on comparisons of mem-
            brane filters.

Geldreich:   I  would like to know  if rosolic acid
            was added,  to  cut down the back-
            ground of other organisms?

Bordner:    The method carried out is exactly as
            in the ASTM draft procedure given us
            and the draft procedure did  include
            the addition of rosolic acid. Any other
            questions related to this paper?

Brezenski:   One of the criteria established in  the
            comparison of the replicate membrane
            filters was  an 85% recovery. Why was
            this number chosen - what is the basis
            for  it?  I just want to know what  the
            reference is, that's all.

Bordner:    This was  just a  goal  that  was  set
            arbitrarily  by the  subcommittee.  I
            think the percentage was from the  De-
            partment of Defense specifications for
            membrane  filters.  That  percentage
            was also quoted for previous recovery
            tests that were identified in the DOD
            specs  that  we  discussed  yesterday.
            Any other questions?
Litsky:      I  cannot  leave  this  room  without
            asking the question I  asked yesterday
            morning,  and that is the problem of
            the  extractables.  We  have  perfectly
            good procedures  on  extractables.  If
            the  specifications  require changing,
            let's try to get a uniform product.  If
            the  definition needs  changing,  let's
            change  it. But its rather ironic  that
            three months ago I  sat in  an EPA
            committee  concerning  disinfectants
            and  detergents, and they stuck to the
            letter of the law! Any deviation from
            the  specifications  caused  a  riot.  If
            the specs written ten years ago do not
            apply now, let's change them.

Sims:       During  the last two days we have seen
            a  dilemma  develop for the  manufac-
            turer. We  are responsive to you, the
            customer,  and we can  provide  you
            with  whatever you want. The repre-
            sentative  from Millipore pointed out
            this  morning  that  if  you create this
            surface  - what they are calling greater
            surface  pore size - you create a differ-
            ent  atmosphere.  It can  be provided.
            Also, cells  in the surface cavity will
            have a  tendency to run  and  colonies
            won't  look the same. In purchasing
            membranes you are going to  have to
            be more particular about what results
            you   require  and  not as particular
            about cell appearance.

            The specs were  written ten years ago
            and  our membrane was  developed to
            meet these specs,  whether correct or
            not, it is the law.  If you are selling to
            the  US Army, they  buy by those
            specs.   You  live  with  extractables.
            Everybody  does the  same  thing.  If
            somebody adds more extractables and
            they  can justify  it and  give  you a
            product,  fine.  If you  want  zero
            extractables, the  membrane  can be
            washed. It will be brittle and it might
            have  usage effects that  would really
            turn you off. You, the customer, are
            deciding what you want to buy. The
            manufacturer will  meet  your de-
            mands.   What  we want from  this
            symposium is to find  out what your
            demands are.

            Right now, from  the attitudes I have
            seen, you want  specific microbiologi-
                                               155

-------
cal  properties.  You  will accept, per-
haps, not as even a gridline, not as
dark a print. You will  take perhaps a
little more brittle product or a prod-
uct packaged in a different way. May-
be you want to  autoclave the mem-
brane  and  it  wrinkles, maybe   it
doesn't.  I think  what  we have been
discussing is a  two-way street. I am
looking  forward  to getting the  re-
sults of  the committee meeting over
the next few days and  as manufactur-
ers  we will  provide  what you want.

The  thing that  did interest me about
this  surface  phenomenon   is  that
membranes  can be manufactured  to
create  a  very tight system, a uniform
pore throughout  where bacteria  sit
right on  the top, fluid  has to nuture
the cell from both sides and the cell
grows  into a little round colony.  If
you  have a sponge-like network  the
cell is down in  it, the fluid surrounds
Bordner:
            it almost on four sides, but always on
            three, the colony growth might spread
            out   differently,  depending  on the
            structure.  If you  are affected  drasti-
            cally by colony appearance and how
            easy  it is for your technician to count,
            you  might  get  lower  counts. It's
            something,  I think,  that is very vital.
            The  manufacturer can  provide you
            with  either structure - its no problem.

            The only other question  I have  for the
            representative from Millipore concerns
            the  0.7  Aim filter. Will  Serratia mar-
            cescens  still be  the species used  to
            standardize  filters?  This  is what the
            specs now read and  as a manufacturer
            I  live by these specs. Are we going  to
            vary  the way  we control retention?
We hope  to provide some of these
answers  in  the  ASTM  committee
meeting.
                                    156

-------
                                   SUMMARY OF SYMPOSIUM

                          John A. Winter, Chief, Quality Assurance Branch
                 Environmental Monitoring and Support Laboratory, Cincinnati, EPA
                                              and
                        Francis T. Brezenski, Chief, Technical Support Branch
                           Edison Water Quality Laboratory, Edison, EPA
         SUMMARY AND DISCUSSION

     Bordner  (Moderator)  —  This  morning  the
agenda is shorter than yesterday. We plan to end
the Symposium  by noon. It will be  followed this
afternoon by  a  meeting of ASTM subcommittee
D19.08.04.02. You are welcome to stay this after-
noon and participate. You are also encouraged to
join the subcommittee as permanent members.

     The last item on the program this morning is
a summarization of the papers in the Symposium.
It will be given by Fran Brezenski and John Winter,
both microbiologists in EPA.
     The papers are placed into four categories,
however, the audience should be aware that there
is  a considerable  overlap  of  subject matter  be-
tween categories.

     I.   General papers.

     II.  Stressed/injured cells  and  observed  ef-
         fects on MF recoveries.
Comparisons   between
brands.
                                 MF   lots  and
     IV.  Solutions to apparent low recoveries.

     General Papers

     1)   Bordner  - The Membrane Filter Dilemma

         Bordner  described the general preference
     in  water  analyses  for MF techniques based
     on simplicity  and  speed. Increasingly,  how-
     ever, the  literature has reported  low  or vari-
     able MF results, particularly for fecal coliform
bacteria  and has attributed these results to
brand  or lot differences in MFs, stressing of
bacteria  in  natural  waters, inhibitive incuba-
tion at 44.5 C, chlorination,  and other test
factors.

     ASTM  Committee  D-19  and  the U.S.
EPA sponsored  this seminar to:  disseminate
known information on the problem, identify
the  better  possible solutions, establish the
standard test  procedures needed for validly
comparing   media,   membrane  filters,  and
other test factors, and select the best solution
using these standard tests.

2)   Geldreich  - Performance  Variability of
     Membrane Filter Procedures

     The variable results reported from mem-
brane filter  procedures are ascribed not only
to  differences  in membrane filter  materials
and  methods of MF sterilization  but also to
inconsistencies  in  absorbent pads,  commer-
cially-prepared media, and the  knowledge and
experience  of the technician.  The  need for
cooperation  between manufacturers and users
and  the  need for improved quality control
programs by both were emphasized.

3)   Powers - Quality Control of Media

     Powers  reviewed   the   manufacturing
techniques and the quality control techniques
practiced by the manufacturer. He suggested
that the  manufacturer  loses  control  over
product quality  once it  leaves his plant. He
then described the  abuses in laboratory test
conditions which influence microbial recover-
ies,  growth  and colonial  characteristics  and
stated that most media problems result from
mishandling  of media in  the laboratory. This
                                               157

-------
    attitude moves  final  QC responsibility from
    the manufacturer to the user.

     4)   Sladek - Statistical  Interpretation of MF
         Counts

         Sladek cited the importance of statistical
     design  in  any  comparison of  microbial re-
     coveries and reviewed the main factors to be
     measured in such studies.

         He used examples of actual studies to
     describe measures of  precision and to demon-
     strate  the  importance of using large numbers
     of  replicates. He cited the problems of stabi-
     lizing  samples for true replication over time.
     Sladek  emphasized   that  the  theoretical
     minimum  deviation  inherent in the method
     must  be established  to  compare with subse-
     quent  experimental  deviations. He  stressed
     that valid  studies  can  only be made by re-
     moval  or  control through  randomization of
     unwanted  variables which  include sampling,
     types  of samples,  procedures,  materials and
     media.

         Finally,  Sladek  recommended  the use
     of a control which is independent of the test
     methods, i.e., non-MF control for comparing
     MF tests.

II.   Injury Papers/Effect Papers

     1)   Hoadley - Effects of Injury on Recovery
         of Indicators on  Membrane Filters

         Hoadley studied the recovery of E. coli
    and S.  faecalis as control cultures and as cul-
    tures stressed by  exposure to stream condi-
    tions and  to chlorination. He used  spread
    plate and  membrane filter tests on  selective
    and  non-selective media.  Control cultures of
    S. faecalis  showed  low recovery on  selective
    media  while E.  coli controls  recovered well.

         After stressing,  E.  coli yielded low re-
    covery  on  MFs,  while S. faecalis showed good
     recovery. E. coli counts on MFs were reduced
    with increased stress.

         Hoadley concludes S.  faecalis may be  a
     better indicator than  E. coli  because  it does
     not show  low recovery after  stress.  Further,
     he recommends that any studies for improved
     methodology should  include  evaluation of
     stressed cell populations.
     2)   Hufham  -  Effects  of Temperature on
         Recovery of Fecal Coliform

         Hufham  cites extreme effects of  tem-
     perature on recovery of fecal coliforms, sug-
     gesting that these high temperature effects
     (44.5  C)  can overlap  effects  of membrane
     differences. Temperature may  have been a
     major  factor  in  earlier papers citing severe
     differences in membranes.

         For  future   recovery  studies,  Hufham
     proposes the use  of a standard  E. coli culture
     which  has not been selected from MF cultures
     at 44.5 C. He proposes that a  standard plate
     count  of cells grown at 35 C is  necessary to
     avoid temperature-related losses and to estab-
     lish  the reference cell  numbers  for MFs or
     other comparisons.

III.  Comparison Papers

     1)   Brodsky  - A Comparison of Membrane
         Filters  and  Media  Used  to  Recover
         Coliforms from Water

         Two  to  seven   laboratories   analyzed
     water  samples and cultures for coliform bac-
     teria using  membrane  filters  from Johns-
     Manville of  Canada,  Millipore  Corporation
     and  the Sartorius Company.

         Cultures   were   both  routine  water
     samples and total and  fecal coliform mixed
     cultures obtained by  passage  through  Mac-
     Conkey broth and EC  broth.  Analyses  were
     performed in parallel  using LES Endo  Agar
     and  M-Endo agar.

         Although the Johns-Manville  and  Milli-
     pore filters  showed  higher recoveries  than
     Sartorius in  Phase I  of the study, Phase II
     results  showed the three filters to be equal.
     Recoveries on LES  Endo agar and M-Endo
     agar were  similar. The  authors reported that
     their results varied with  the test conditions
     used.  They concluded  that test design and
     quality control must be carefully selected and
     standardized before the results are  meaning-
     ful.

     2)   Stuart - Comparison  of MFs in Recovery
         of Naturally-Injured Coliforms

         In a series of studies, Schillinger et al. ex-
     posed   pure  cultures  of  E.  coli  using  MF
                                              158

-------
chambers, in situ in a stream.  Cells were re-
covered, enriched and  tested  by the M-FC
procedure,  using  Gelman,  Millipore  and
Muclepore  filters. Results to date show no
significant difference in recovery of E. coli on
Gelman or Millipore filters. Nuclepore filters
gave lower recoveries than the  other brands.
It was concluded the Nuclepore brand should
not be used for culture work.

3)   Harris - Efficiency of Coliform Recovery
     Using Two Brands of MFs

     Harris  and  Bailey compared three  lots
each  of  Gelman  and Millipore filters in the
testing of aerobic lagoon and sewage samples
for total and fecal coliform bacteria. Gelman
filters were steam-autoclaved and Millipore
filters were sterilized  with ethylene  oxide.
Total coliform counts showed  no significant
difference with lot or brand. However, twice
as many fecal coliforms  were  recovered on
Gelman filters as on Millipore filters. Colonies
on  Gelman  filters  were  smaller than  on
Millipore. Millipore filters were  blue and  Gel-
man  filters  beige, suggesting a  pH effect on
the dyes.

     This  work  confirms Presswood  and
Brown's work. In later work by  Harris  and
Bailey,  experimental membrane filters from
Millipore  showed recoveries comparable to
Gelman. The authors urge more quality con-
trol   for  uniformity  between  batches  and
brands.

4)   Dufour -  Comparison of MF  Brands for
     the Recovery  of the Coliform Group
     Dufour and Cabelli  studied  the recov-
eries  of pure  cultures  of E. coli, K. Pneu-
moniae and fecal and  total  coliforms from
natural  samples, using  membrane  filters  pro-
duced by Gelman,  Millipore, S&S, Sartorius
and Nuclepore.

     They found that strain differences in the
organisms and differences between  lots of
membranes  confused the  recovery compari-
sons  between  brands.  They  noted  that  the
precisions were consistent  with brands when
using pure cultures.

     The accuracy of recoveries was the same
whether pure cultures or natural samples were
used. Because  precision decreased  with natu-
ral samples, studies done  with  these samples
should use a larger number of replicates than
for pure cultures. Acceptable total coliform
counts were obtained with all filters except
Nuclepore.

5)   Glantz  - Comparison of  Millipore and
     Gelman Filters, Culture Media,  Incuba-
     tors and Escherichia coli strains

     Pure culture isolates of E. coli were test-
ed for recovery on Trypticase soy agar (TSA)
pour plates, violet red bile  agar (VRB) spread
plates and  M-FC  broth membrane filtrations,
using different brands and lots of membrane
filters. Recoveries were compared at  incuba-
tion temperatures of 35  C, 43 C  and  44.5 C.

     Although some cell  counts were reduced
at 44.5  C, counts varied  most  significantly
between  E. coli strains and between different
MFs.  M-FC broth gave lower recoveries than
VRB. Standard strains of E. coli are recom-
mended for future evaluations.

6)   Davis  - The  ASTM  Proposed Membrane
     Filter Test Procedure  for  the Recovery
     of Fecal Coliforms

     Jackson and Davis described the pre-
liminary  test procedure developed   by the
ASTM subcommittee 019.08.04.02  for  re-
covery of fecal coliforms by membrane filters
and detailed the  round robin test performed
by ten laboratories using the procedure. They
summarized  the  statistical  analysis  of  the
results and concluded that the  preliminary
procedure did detect differences in filters but
was unsatisfactory because  it did  not separate
the  effects  of  individual  laboratories and
techniques  from  the natural  sample differ-
ences and other procedure variables.
7)   Goddard  -  Critique on  ASTM Test for
     Recovery of Fecal Coliforms and Propo-
     sal for Modified Method

     Goddard  feels that the  necessary stand-
ardization and control  of  variables was not
attained in this first ASTM test effort.

     He discussed briefly membrane  charac-
teristics and  suggested that the  better  re-
covery  reported with  large surface pores
could be due to the cradling  effect or simply
be due to easier passage of nutrients.  He sug-
                                          159

-------
    gested that an  optimal  pore size  for  fecal
    coliforms be  established  as  a  standard  and
    comparisions of filters then  made  based on
    other  physical characteristics of the  filter.

         Goddard critiqued the ASTM procedure
    with  the  following suggestions for improve-
    ment:  Use  a  standardized  potable  water
    sample to obtain a more  uniform and favor-
    able flora for the test, report noncoliform and
    coliform counts, use spread  plates of  same
    size as filters for fairer  establishment of ref-
    erence count, place samples  randomly in in-
    cubator and verify accuracy of thermometers.


IV. Solutions

    1)   McFeters - Recovery Characteristics of
         Bacteria Injured  in Natural  Aquatic
          Environment

         Bissonnette,  et  al.  continued  the MF
    chamber  studies  described  by  Schillinger,
    et al.. They placed raw sewage, E. coli and S.
    faecal is cultures in the chambers in  situ in
    stream, and sampled the chambers daily over
    time.  They plotted  recoveries  in Trypticase
    soy yeast  extract agar  (TYSA) and desoxy-
    cholate lactose agar (DLA) and calculated the
    number  of bacterial cells injured (unable to
    grow).

          Further  studies  added  MPN  lactose
    broth,  MPN  BGB  broth, M-Endo MF  and
    M-FC membrane  filter media.  Comparative
    recoveries were:

    MPN - TSY = MPN - LB > MPN - BGB > DLA
    >M-Endo-MF>M-FC-MF

         The numbers of survivors increased with
    increased time of exposure.  Membrane filter
    media were  less efficient than  other  proce-
    dures in recovering injured bacteria. Exposure
    of  uninjured cells to TSY broth repaired the
    cells  so  that  they could  reproduce.  A two-
    hour enrichment on TSY agar was suggested
    before enumeration  of  indicator bacteria on
    a selective medium.

    2)    Geldreich - An Improved MF Method
         for FC Analyses

          Rose et al. described the  development
    and testing by  three laboratories of a two-
    layer  agar method.  The  procedure used an
overlay of lactose broth agar on M-FC agar
and two-hour incubation at 35 C prior to the
44.5 C incubation for 22-24 hours.  Limited
work with raw and chlorinated wastewaters,
and results from reservoir, stream and marine
samples showed improved recoveries in 59 of
61 samples and a median improved recovery
ratio  of 1.9.  More work must be done to
verify these results.

3)   Grasso - Measurement of Fecal Coliform
     in Estuarine Water

     Stevens, Grasso  and Delaney  cited  at-
tempts to improve recoveries of fecal coli-
form  from seawater using Millipore filters in
two-step  procedures.  The  first  series using
minimal  media at 25 C failed to yield con-
sistent recoveries.

     The second approach utilized a two-step,
two-day  procedure of incubation at 25 C for
18  hours on a  minimal LES medium  then
transfer to M-FC for incubation at 44.5 C for
24 hours. Grasso reported an average increase
of 2.9 in the  recovery  of fecal coliforms as
compared with EC counts. Ninety-three per-
cent of the picks of  fecal coliform colonies
did verify.

     Later work  suggested  that batch and
brand  differences  in  membrane filters may
have  influenced the  study  data which was
generated using  Millipore filters only. Com-
parison  of  Millipore  and   Gelman   filters
showed a high level of dissolved solids in the
Gelman  filters.  Examination of water  ex-
tracts from the filters showed a low pH and
significant mg/liter  levels of ammonia-nitro-
gen  and  orthophosphate   in  the  Gelman
filters. Gelman filters showed higher  recov-
eries than  Millipore  filters on regular M-FC
medium  but  lower  recoveries with the new
LES procedure. Further tests are being con-
ducted.

4)   Lin - Evaluation of Method  for Detect-
     ing  Coliforms and  Fecal Streptococci in
     Chlorinated Secondary Sewage Effluents

     In a massive series of tests, Lin studied
total   coliform,  fecal  coliform, and fecal
streptococci  recoveries  from unchlorinated
waters and wastewaters  and chlorinated efflu-
ents,  using membrane filters and MPN tech-
niques.
                                              160

-------
          MF  recoveries with  chlorinated second-
     ary effluents were significantly lower for total
     coliform, fecal coliform and fecal streptococci
     than the counts by the standard MPN. M-FC
     counts were  lower than  EC recoveries.  No
     fecal  coliform  enrichments  were reported.

          Low recoveries of total coliforms using a
     single-step  procedure  were  improved   1.5
     times by use of the LES two-step procedure
     and were comparable  to  the  MPN values for
     these effluents.  Equal   recoveries  of total
     coliforms were made with the LES two-step
     MF  procedure  and the standard MPN con-
     firmed test when testing  chlorinated second-
     ary effluents.

          Use of a bile broth  enrichment for two
     and  three  days increased fecal  streptococci
     counts to a level comparable to  MPN values.

     5)   Sladek - Optimim Membrane Structures

          Sladek studied test factors  affecting re-
     covery of fecal coliforms in MF tests and con-
     cluded that the  most critical  factor in  re-
     covery was the structure of the membrane
     filter. Studies on maximum recovery of fecal
     colifrom  bacteria showed that there was an
     apparent optimum top surface  opening of
     2.4 jum accompanied  by a 0.7  /zm  internal
     pore size. This  surface opening was reported
     as critical for maximum recovery  of fecal coli-
     form organisms because the cells are exposed
     to a high evaporation effect at 44.5 C. Evap-
     oration  produces hypertonic solution areas
     unless the  cells are cradled in surface open-
     ings. Millipore feels this theory explains the
     low M-FC recoveries on 0.45 /im  membranes.
     The  better recoveries on the new  2.4 ^m
     surface pore, 0.7 ;um internal pore membranes
     confirm this.  Data are limited and must be
     substantiated.
            GENERAL COMMENTS

Winter:      Let's forget about any more compari-
            sons of MF's. There have been enough
            such studies. If we do any more work
            we had  better look at the solutions
            suggested in this Symposium.

            Almost all papers given here were fo-
            cussing on some variable and  ignoring
            another  variable  of lot,  batch, medi-
            um, incubation temperature, time and
Editor's
Note:
Winter:
so forth. A number of solutions were
proposed  and  these  proposals  need
more supportive data, but  only after
factors which affect the MF test are
controlled or randomized as described
by  Karl Sladek, so that we can get a
valid comparison of numbers.

We would urge everyone  to partici-
pate  in   collaborative  studies  per-
formed  with ASTM,  EPA  and other
testing groups. No proposed method
should be accepted until it has had a
satisfactory  collaborative study. Fail-
ure to do so was  how mistakes were
made in the past. If methods get into
books, there is a  great deal of diffi-
culty getting them  out.

At  this stage we shouldn't try for an
instant solution.  We  have  a serious
problem of test recoveries made more
serious because of compliance  moni-
toring regulations  promulgated   by
EPA and implemented by the States.
Specific industrial  and municipal dis-
chargers are  required  to get NPDES
permits  and  prove that they are not
exceeding the  limits  for  fecal  coli-
forms set in their  permit. The fecal
coliform test that must be used at this
time  for  chlorinated   wastes  is  the
MPN. The MPN is  the only technique
we  know of which seems to withstand
the effects of chlorination and give us
the maximum recoveries from chlori-
nated effluents.

Since the  time of  the Symposium in
January 20-21, 1975 EPA has review-
ed  public response to  the proposal
that only the MPN be used  to analyze
effluents  for fecal coliforms  in  the
presence  of  chlorine.  The  current
proposal is that both the MPN and MF
may be used but  the  method  used
must  be  identified. The MPN  is the
method  of choice  if  controversy is
anticipated because the MF has been
reported  to  yield low  and  erratic
results from  chlorinated  effluents.

The users need the cooperation of the
MF manufacturers  in solving the prob-
lem  of  recoveries on MF's. If they
come out with a new formulation of
                                              161

-------
           membranes  such as the one spoken
           of  yesterday,   the   HC  membrane
           filter from Millipore,  I  hope that we
           don't find out that it has phosphate
           or  some  other  stimulating  factor in
           it.  We  don't  want  more  surprises
           about membrane  filters.  We   were
           told that membrane filters were inert.
           The specs have always said they shall
           not enhance or  inhibit  growth. What
           should we do now? If  user and sup-
           plier  exchange  information and work
           together they can  find hopefully  a
           solution to the MF recovery problems.
Brezenski:  Bob Bordner started the Symposium
           with a question:  why are we  here? I
           think the answer was to cope with the
           problems of the  lack of recovery  by
           the membrane  filter systems, and the
           number  of  conflicting reports in the
           literature. After reviewing the papers
           and tapes and  listening  to the discus-
           sion yesteraday we  concluded those
           two objectives were met. So from that
           point of view, the meeting was suc-
           cessful, but the ASTM Committee  on
           MF's  should  continue  its  work.  I
           would like to make a suggestion to the
           manufacturers.  One  company contri-
           buted  two papers to the Symposium.
           There  are other companies  manufac-
           turing MF's. I would  like  to see a
           greater  research contribution  by the
           companies themselves. I am sure that
           most  of them develop data in the pro-
           cess  of controlling  the membrane
           manufacturing  process.  This data can
           contribute  to  the knowledge about
           the physical and  chemical character-
           istics of the membranes.

           A  number  of papers described the
           phenomenon of the stressing of  cells.
           There were a  number of terms used
           yesterday: stressed cells,  attentuated
           cells,  resuscitated cells, damaged cells,
           debilitated  cells,  unresuscitated  cells.
           All of these terms have a somewhat
           different connotation.  I would urge
           some group, government, academic or
           private,  to  investigate  what these
           debilitated  cells  are physiologically.
           Once  we have this  information, the
           medium and  the whole testing pro-
           cedure  can be  modified  to recover
           these cells.  Right  now  we see a lag
           time in growth and say that the cells
           were debilitated or they were stressed.
           In what way? This is a major problem
           that was not discussed in the Sympo-
           sium. Yesterday, Ed  Geldreich men-
           tioned the tremendously large number
           of MF tests run in the US each year. It
           brings to mind that if we  have defec-
           tive MFs or problems of recovery, we
           have  had  many  defective  numbers
           produced  and  used  over  these past
           years.

           I  would like to close by saying that it
           was an  enjoyable  experience.  I  have
           been in a  lot of other  meetings and
           this is the first time I  have seen  so
           many worthwhile  contributions and
           such active participation in a one-day
           session. I  think we  are well on our
           way to solving this  problem.  Thank
           you.

FINAL DISCUSSION

Cotton:    I  would like to commend the review-
           ers for putting the program of yester-
           day in such great perspective. I would
           like to speak  briefly on behalf  of
           membrane manufacturers. I  was with
           Miliipore  in the beginning when the
           company  first  took  the  initial con-
           tract and developed the membrane. In
           fact, I  was responsible for part of the
           development and for the  initial work
           on the coliform test done  by Milli-
           pore. One point which was  left out
           was that although  MF's are a big busi-
           ness for other purposes, the commer-
           cial development  of the membrane
           filter  stemmed from  the  coliform
           test. It was felt that the coliform test
           provided a significant commercial use
           for the membrane filter. This is where
           it all started. Although the use of the
           membrane filter in sanitary bacteriol-
           ogy is not the major commercial appli-
           cation,  in spite of the millions of tests
           that are being  run, it is a significant
           one and should be of great concern to
           the membrane manufacturers.

           Mr. Geldreich made a point on filter
           specifications and  quality control that
                                              162

-------
 I  think needs  clarification.  I  know
 that  our Company and  other  mem-
 brane manufacturers are aware of the
 need  for extreme quality control tech-
 niques  in their  own laboratories. The
 problem is that there are many as-
 pects of the filter that need to be con-
 trolled. Sometimes  when  you find  a
 little  piece of paper, as we saw yester-
 day,  or when a box of filters comes
 out of production  without the filters
 and only the separator pads, you can
 get a  good laugh, but this represents a
 very  small  percentage  of the total
 products produced.  It is only one of
 about 14 different aspects which need
 careful control.  All the manufacturers
 have brought some fine control condi-
 tions  to the filter manufacturing pro-
 cess.
One point brought  out was extract-
ables  in filters. I don't think this was
completely  clarified  yesterday.  One
manufacturer's representative pointed
out that glycerol was  put into the
filter  to plasticize it. This is not cor-
rect. Glycerol is used in pore forma-
tion. It is a  means to control the pore
size of  the filter, and  after manufac-
ture it  is removed. The  degree of re-
moval  will  have an effect on the elas-
ticity   of the filter and for  certain
applications  it  is not  important to
remove all  of the  glycerol  or  even
most  of  it.  For  some  applications
however,  it is very  important,  espe-
cially  in gravimetric analysis. In the
case of the coliform test, if too much
glycerol  remains in the  filter,  false
positives will  result  because glycerol
is  broken down to  aldehydes  which
react with the fuchsin sulphite system
and produce sheen.  So,  it is very im-
portant to limit the amount of glycer-
ol  that remains in the filter. This is
one of  the  reasons  for  the  tight
government specifications. Originally
the specifications on total extractables
in   the  membrane were set at  2.5%
 Recently, the  extractable  condition
on  specifications  in the  government
specs  was dropped,  and  we are  look-
ing into this.  We  think it should be
re-instituted primarily  because of the
false  positives  in  the coliform  test.
Also,  in the  gravmetric analysis, which
 is done in  water and  in fuels, false
 readings can result because of extract-
 ing some of the weight of the filter
 which will counteract the-dry weight.

 Another   membrane   manufacturer
 mentioned  the  inclusion  of the wet-
 ting  agent,  which  is  an  alkylaryl-
 polyether  alcohol.  This  is  well  re-
 searched. It was selected because of
 low toxicity and is in the formulation
 primarily for pore formation and con-
 trolled  pore size. The amount which
 is  removed  in   the  manufacturing
 process  must be controlled so that
 the   filter  is  wettable.  Otherwise,
 filtration won't  take  place or non-
 wetting  spots will result. The  state-
 ment was correct,  but the primary
 reason  for  it is  control of pore for-
 mation  in  membrane  production.

 Subsequently, the procedure calls  for
 removing both the wetting agent and
 the   plasticizer   in  the  processing.


 I want to point out that during  the
 development of  the total coliform  MF
 test,  thousands  of  tests were run.
 There was a competition between  the
 membrane filter and  the MPN pro-
 cedures.  The problem  was  to  get
 acceptance  from the Standard Meth-
 ods Committee  of the MF method in
 comparison  to the MPN. Hundreds of
 papers, many contributed by some of
 the   researchers   here  today,  were
 written before this test was accepted,
 first as a tentative standard and then
 as an alternate to the standard MPN
 test. Subsequently, however, when  the
 fecal  coliform test was developed as a
 better sanitary  indicator,  something
 slipped  in.  It was assumed  that  be-
 cause this was also a coliform test,  the
 same  filter should be the right product
 for the test. Other aspects of this test
 were   examined.  Recently*  reports
 were  published  that observed  a dif-
 ference  in  recoveries  in  the  fecal
coliform test. Dr. Presswood's paper
appeared  first  and  other  followed,
 but there was  a lot of confusion.
When you look  at these reports  to-
gether you find that different research
found diametrically  opposed  results.
                                    163

-------
The  ASTM  committee that will  be
meeting here this afternoon to estab-
lish  methods  for  evaluating  mem-
brane filters has decided that the fecal
coliform test would be the optimum
test for evaluating all microbiological
membranes because of its  sensitivity.
They  worked on the  test procedure
and modified it and as is the final step
in all   ASTM  methods,   planned  a
round robin study.  A number of labo-
ratories, including our own, agreed to
run the test and compare  the results
to make sure that the test works. The
results of the round-robin were rather
confusing and  you heard  about that
in the report by Mr. Davis. It was dif-
ficult to  determine  what  had hap-
pened.

Our  company  began  a research pro-
gram to find  the  problem. We were
looking for what we  call  the big red
X, this  confusing  factor that led  to
what seemed to be total  confusion in
the fecal coliform  test.  Dr. Sladek's
group spent  a  great deal of time  re-
searching the  point. They  found out
that there can be an effect of the dif-
ferent factors. The question of differ-
ent  methods  of   sterilizing  filters,
autoclaving   versus  ethylene   oxide
versus  high  voltage were discussed.  I
think this point was just a little con-
fused yesterday. Proper  sterilization,
regardless of the method,  should not
be a significant factor.  If you  leave
ethylene oxide in  a filter and test it,
it's going to be toxic to the organisms,
there  is no  question  about it. That's
what the  ethylene  was there  for, to
kill organisms. But if you properly
remove  the  ethylene oxide,  it will
have no subsequent effect.


Let  me  re-draw  the curve that Dr.
Sladek showed in the presentation, be-
cause I feel that surface effects are not
the only factor, but the key factor.
After we found out about the surface
pore phenomenon, we ran tests on  a
tremendous number of different pore
size  filters,  comparing  the  surface
pore openings and pore sizes against
recoveries of fecal coliform in the test.
With the fecal coliform test, there was
a very clear pattern.  If the ordinate is
recovery shown as percent recovery or
cell numbers and the absissa is the sur-
face pore   size  measured with  the
electron  microscope, then there  is a
very pronounced curve of this nature.
We repeated the tests many, many
times. Because of time limitations, Dr.
Sladek  showed  only  a  few of the
results,  but what was shown clearly
was that the .45 urn filter which we
manufacture has  a pore  size opening
which lies close to the top of  this very
steep curve. I will not speak about the
other membrane manufacturers' prod-
ucts. However, a similar  curve should
develop for their products. If there is
a slight difference in the surface pore
openings from lot to lot, and  batch to
batch, it is a factor that has never be-
fore been quality controlled. We did
not realize  and  no  one knew that  it
was an  important factor, so  we were
looking at retention  pore size and not
surface  pore size. In the future  this
will change.

Slight  changes in the surface  pore
size  or  pore openings could  create a
tremendous difference in  the recov-
eries and this, we are convinced, is the
reason  why  attempts  to  compare
autoclaved  packed filters with ethy-
lene oxide  packed filters showed one
thing one time, and  something oppo-
site the next time. We were not care-
ful to use filters which were the same,
from the same batch or with the same
surface  pore size. What we are doing
as a company  is recommending  that
for the fecal coliform test we go up
into this region where slight changes
in the surface pore size will have very
little  effect in the  recovery  of  fecal
coliform. In the limited tests we have
run comparing this for optimum sur-
face pore size filter, we found a very
good uniformity of results  and we
have found recoveries usually higher
than what  we are using as the control,
the  spread plate. That's  where  our
technology  stands at this particular
time.

In this research we came up with a
few other  points which are worth-
                                    164

-------
while reiterating.  I think  Dr. Sladek
mentioned  these  yesterday  but  in
terms  of an appropriate control  for
this type of work the thickness of agar
in the spread plate control is an  im-
portant factor. We were just casually
pouring  our  spread  plates without
carefully  measuring the amount  of
thickness and found that this was an
important  factor.  We  use a 90 mm
diameter petri dish. We ran tests using
different quantities from 15 ml up  to
35 ml  of M-FC agar and we found a
recovery which looked essentially like
this with a peak at about 30 ml in the
dish. But if we had too little, we had
low recovery so that  is a factor that
we  are now carefully measuring  for
our controls.

A second factor mentioned  by several
people is the dilution water. Initially,
the fecal coliform  test was set up with
phosphate  buffer  dilution  water and
phosphate buffer is toxic to fecal coli-
forms.  It is  now  written as peptone
dilution  water  and it  should be cor-
rected,   incidentally,   to  phosphate
buffer  peptone water (we will  talk
about  that this  afternoon).  If  you
have good  pH  7 distilled water, pep-
tone water is fine,  but if your distilled
water is off in pH, you need a buffer.
This is another factor  keeping  your
control sample count low.


To finish off, I have another point for
clarification.  The question  was raised
about the effect of vacuum on bacter-
ial cells.  Let me say clearly that the
degree  of  vacuum has no  effect on
the bacterial cells. When the cells are
in the water they don't know whether
your vacuum pump is pulling one inch
or  three atmospheres  of  pressure.
They are not affected by the vacuum
unless they  are on the surface of the
water where the  vapor  pressure has
an effect. So during filtration they do
not know  what the  vacuum is and
don't care. The rate of  impingement
against the  filter  is not significantly
different. It is very slow regardless of
whether  it  is  high vacuum or low
vacuum.  At the end of the filtration,
when there is  no more water going
            through the filter, there is a blockage
            effect  because of Poiseville's  law of
            capillarity. The vacuum is  below the
            filter  but they again don't know to
            what  degree the vacuum is below the
            filter. So  unless you let the filter sit
            there  and dry for a long time  where
            the drying effect could kill the cells,
            there  is no effect of degree of vacuum
            on the cells.


Question    Is there anything  in the literature that
from        we could  read to find out all of this?
Audience:   I've heard you say this often,  and I
            have heard others talk that way but
            have  never  read  anything about  it.

Cotton:     Well, that particular point isn't worth-
            while  doing a research study on, be-
            cause  there is no possible effect.  It
            can't  happen. The cells  don't  know
            what the vacuum is. However, I don't
            know of literature on  that particular
            study.

            I think I  have just about covered  all
            the points that  I  had here,  except
            that as a result of our information we
            are  going to make  available a  filter
            which has a surface opening pore size
            which  will have  a retention pore size
            of  .7 urn.  I thank you for allowing me
            this time  to  speak on behalf of the
            membrane manufacturers. I would be
            happy to answer any questions.

Bordner:    Mr. Cotton,  you  have just given  us
            another paper which was not  on the
            agenda  but provided  additional  in-
            sight  into filter  manufacturing pro-
            cesses.

Winter:     We are looking at a graph showing the
            recoveries of  fecal coliform relating
            to  the surface pore opening. One of
            the points in the  paper"was passed
            over yesterday. It shows that passage
            of  microorganisms begins, interesting-
            ly enough, at 2.4 /zm. This means that
            cells pass  through your membrane.  If
            we will take Sladek's  description  of
            the standard deviation as  being the
            square root  of  the mean, and the
            mean  pore  opening size  is 2.4 vm,
            then one  standard deviation is  about
            1.2 //m and that  means that half the
                                   165

-------
            pore opening  sizes are greater than
            2.4 /im and hence are passing cells?
Sladek:     I was speaking of a  certain contribu-
            tion  to  scatter in  measuring counted
            data, such as you have in bacteriology
            where you count  a  number of colo-
            nies. Now  this is  not counted data,
            this is a measurement which is more
            like  a  measurement of temperature,
            and on  something like  a pore distri-
            bution there is no necessity  for the
            standard deviation to have any partic-
            ular size. That is, there isn't a Poisson
            distribution involved. This  is not  a
            counted sort of phenomenon.  In fact,
            the width of the pore distribution in
            membrane  filters is controlled entirely
            by the  detail of  the manufacturing
            process.  It  is considerably  narrower
            than  what you   were  just  saying.


Winter:     Well, how narrow is  it? This is a point
            that  has always bothered us. In any of
            the  literature,  manufacturers  talk
            about A5nm openings. They show
            neat pore  size distribution  in  their
            literature  which has  no tails. I  have
            always been curious  as to how they
            do  something  that  nobody can do,
            which is to cut off  the tails of the
            distribution.  It  shows  that as you
            go up in pore size opening on the sur-
            face, you  are also enlarging the size
            of the internal pore so that at 2.4 ;um
            the internal pore  size is not .45 but
            0.7 /urn.  So pores are  coming  very
            close. Are we retaining all of the cells?

Cotton:     May  I  speak  about  that question
            since  I'm  the  person that orginally
            cut off  the tail of  that curve.  That
            curve comes from work in measuring
            pore size with  the mercury intrusion
            technique.  I won't go into  detail be-
            that takes quite a  while, but there is a
            tail on  both ends of that curve and
            that was cut  off  for the purpose of
            explaining  our case.  The key factor
            as Karl  pointed out yesterday,  is  re-
            tention  of particles which are filtered
            through the filter and the criteria for
            the .45 p.m filter as established by the
            government was   100%  retention  of
            Serratia marcescens  cells which  aver-
            age in size of .6 to .7 micrometers in
Bordner:
diameter.  The coliform cells that we
know  about,  and  as  described  in
Bergey's manual  are all  larger than
that. I  think Bergey states 1.2/zm  to
2 ^m. They may expand on the range,
but  our  experience is  with  filters
which  average this size and in which
there is no passage. That is, the filters
made at the optimum size in the mid-
dle of  this  curve  showed no passage
whatsoever. We  began to see passage
at the  3 jum size, so if you look care-
fully at the graph  we showed you will
see some  passage at  the 3 urn surface
pore opening  and this would answer
the spread question  also. The passage
is  probably from  the  larger  pores  in
that particular filter. Its  not that the
precise size shows  no passage, but it's
the filter  at  that  pore size  that has
pores larger or smaller which show no
passage.

Are  there any  questions  related  to
Dick Cotton's comments?
Seidenberg:  Do I  understand that you are develop-
            ing another membrane with a surface
            pore  size which will  be suitable for
            fecal  coliform?

Cotton:     Yes.   Developing may  not  be  the
            proper term, because we can now pro-
            duce membranes with any graded pore
            sizes.  Let's say we have another mem-
            brane which  we are  going to  make
            available.

Seidenberg:  Does this mean that we'll have to have
            one membrane for total coliform and
            another for fecal streptococci?


Cotton:     Not  necessarily.  It  is our current
            feeling that  the filter which we are
            going to recommend for  fecal coli-
            form  will  be completely satisfactory
            for total coliform, although not  neces-
            sary  for total  bacteria count. As for
            fecal   streps,  we can't  answer  your
            question just yet because we will have
            to evaluate that parameter.

Seidenberg:  Then  it's possible that  when we do
            tests  we may have to use one of more
            different types  of membranes, is that
            right?
                                               166

-------
Cotton:    That's possible.

Seidenberg: That's more confusion. Thank you.

Cotton:    This  is true, of  course, with a great
           many of the procedures  even now.
           There are  a  variety  of membranes
           available; you may just be concentra-
           ting  on  the one .45 jum membrane,
           but there  are not  14 different pore
           size membranes.

Seidenberg: I'm  only  thinking  of  bacteriology.

Cotton:    For  the  optimum  test, it  may  be
           necessary  to  use   different  filters.

Bordner:   Dick, can  we  review?  Is  this new
           membrane  with the  wider  surface
           pores,  the HC  experimental mem-
           brane?

Cotton:    At the present time, our company is
           calling  this  the  HC   experimental
           membrane.

Bordner:   Is  this a membrane that has  an  ap-
           proximate mean  pore  size, if we can
           use those terms,  of 2.4 jum on the top
           surface  and .7  Aim  in the  bottom
           which we  can call the retention pore
           size?

Cotton:    Surface  pore size of 2.4 micrometers,
           plus  or  minus,  we  don't know  ex-
           actly yet, and a retention pore size of
           .7 micrometer.

Brodsky:   I wanted to ask whether Millipore or
           other membrane filter  manufacturers
           in  their quality control work on mem-
           branes have looked  at the filtrate to
           see the rate of  passage of the organ-
           sims  through the filter. I  discussed
           this,  this morning with somebody, and
            I think it  is a valid point. When you
           show on this point  of the graph that
           any slight  variation  in retention pore
           size can have a tremendous effect on
           surface pore size, and we know that
           some organisms  are always  going to
           pass  through, they don't always line
           up in any particular direction; they
           don't polarize   themselves,  so that
           you get  an increased loss of organisms
           through the filter due to slight varia-
           tions in surface pore retention size.
Cotton:     Let  me clarify that  point. Passage of
            organisms through the filter is a very
            important factor with us. We do check
            very carefully  to  determine  whether
            or  not  the organisms  have  passed
            through, both  in  this test and more
            significantly  in sterile filtration pro-
            cedures. Passage test data were pre-
            sented  on  the graph that Dr. Sladek
            showed, and all filters with the larger
            pore size-surface pore openings were
            checked for passage. This explained to
            us why we got a tail-off on this curve,
            it was  due  to passage. That  passage
            data was presented here and if you got
            a copy  of Dr. Sladek's paper I  think it
            is explained more clearly.  What we
            found was that at the optimum  sur-
            face pore size, there was no passage.
            Actually, we got just  as high a  re-
            covery  with larger surface-pore size,
            but  because  of passage,  the recovery
            dropped.  If  you  added the  passage
            data to the  recovery data you came
            out about the same until you reached
            the  point  where  the pores  were so
            large that  media transports were ex-
            cessive  and you got  sloppy colonies,
            spreading and difficult-to-read colo-
            nies.  So that  was  really  the upper
            level, except for  the passage factor
            which was carefully studied.


Bordner:    Thank you, Dick. Because Mr. Cotton
            is from Millipore  Corp., I feel  a need
            to  invite comments from other MF
            manufacturers.  If none, we welcome
            any additions  or comments  on  this
            summary.

Litsky:      You  know,  there is an  old  trick:  If
            you want  to  keep  a guy quiet  you
            make him  a moderator.  As long  as  I
            am not a moderator, I'm  going to talk.
            I can't help thinking that this is the
            first time I have seen so many big-wigs
            from EPA in one room and I am going
            to  take the opportunity  to  remind
            them what they want to forget. Years
            ago  we woke up one morning with  a
            book  published  by  the APHA  and
            sanctioned  by  EPA which  required
            everyone to  use a fecal  coliform MF
            test. We could not fight this  because
            if you  ever  went   to  college  you
            learned  that you  follow what  the
                                               167

-------
Bordner:
coach  believes  and the coach in  this
case accepted only fecal coliforms. It
was pushed down our throats without
proper evaluation  because, if we  had
had  time to evaluate we would have
found  the test did not work for all
types of samples. Most of the fault lies
with EPA; however,  we  in the uni-
versities and in the other  non-federal
laboratories  are   at   fault  because
microbiologists  are very lazy people.
When  someone publishes  a  method
which  gives two  steps, we accept it
without   question  because  1)  the
coach  accepted it  and 2)  no  one had
the guts  to question  it. I  was a little
teed off  with Presswood  because we
were content. We were very  comfor-
table although we knew it  wasn't such
a good method,  but  EPA wanted it
and it  saved us  some time and money,
so we  accepted it.  I hope to  God we
don't repeat this mistake.

We  (Litsky,  Rose,  and  Geldreich)
proposed  a  method  yesterday.  We
proposed  the method only to ask all
of you people  to try it  and tell  us
whether we are on the right  track. I
think we  are trying to make  a  point.
I  think we all look to EPA for guid-
ance and this is the way it should be;
however,  I hope EPA takes the warn-
ing that  we are not  going to accept
any other methods without the proper
round-robin testing, the proper field
testing and without the proper data to
prove  that the method proposed is
better  than the  method we have now.
I  hope EPA takes the lead because I
speak  for all  the  private, city, and
state laboratories.  We are  anxious to
cooperate. Use  us, but don't abuse us.


Thank  you Dr.  Litsky. I feel that not
only have our minds  been stimulated
but also,  at least psychologically, that
area of anatomy that young children
usually have  stimulated after getting
into trouble. It occurs to me that EPA
has  always  tried  to work  through
the concensus of opinion and  through
Standard Methods of  which you and
several  others within  and  outside of
EPA are  members.  So we try to pro-
tect  the  Agency  from single  interest
                                                             pressure. We will accept the guidance
                                                             and,  I  hope,  the help of interested
                                                             microbiologists,  academic  and  non-
                                                             government. Are there other additions
                                                             or critiques on the summary?
Hendricks:  I  really don't know what to say after
           Dr. Litsky's comments. Certainly I'm
           not an  EPA big-wig. I'm big, but the
           stature and the agency position do not
           correlate. I would like to say to Dr.
           Litsky that his comments are true in
           many respects.  However, I  would say
           this  about the  fecal coliform tests.
           They have served us well. We are for-
           tunate  that  we have  reduced  the
           intestinal disease rates in this country
           and  other parts of the world where
           this problem has existed.


           There is no doubt that we can do bet-
           ter and  I would like to make three
           comments. One is about the organisms
           which  are used  in tests to  evaluate
           media, filters and culture procedures.
           We all  know that  when we grow a
           pure culture in a broth and recover it
           under  stress  conditions, whether the
           stress is temperature, nutrient concen-
           tration, or inhibitors, the  results will
           be low. Organisms that you introduce
           in the  environment become stressed
           by  environmental parameters as  Drs.
           McFeters and Stuart have shown. To a
           great extent we are going to have to
           control the way we use our organisms,
           whether they  are  pure  or  natural
           cultures,  if  we  are ever  going to
           achieve uniformity  in the procedures
           that we  have been  talking  about this
           week. There is plenty of data to dem-
           onstrate this. We need a standard way
           of treating our cultures.

           Secondly, the  papers of Presswood
           and  Brown and others, I think show
           quite clearly that there is a  problem
           with membrane procedures. Let's look
           at the reasons why these results occur.
           I  think  we  have had enough of the
           observations.

           Thirdly, Ed Geldreich mentioned yes-
           terday that there can be one coliform
           per 100  ml in drinking water. So if we
                                              168

-------
           recover additional coliforms by  im-
           proved procedures, what is this going
           to  do to our standards? We should
           keep  in mind  the general significance
           of an improved technique. Something
           has to be done with the data and we
           had better know what we are going to
           do with it once we get it. The concept
           that recovery  may  be a function of
           survival is a valid one.

           Fourthly, I  am  aware  of a  paper by
           Kerr,  University  of   Georgia,  that
           showed visual  differences by electron
           microscopy  in  Millipore, Nuclepore,
           and  I  think,  Sartorious membranes.
           This technique might be of value to
           those of  you  who want to observe
           bacteria after culturing by membrane
           procedures. Thank you.


Bordner:   I  might  add that Dr.  Kerr was in-
           vited  to  give a  paper at this sympo-
           sium and  would have accepted had he
           had the travel funds.

Power:     I  would like to just make a  comment
           or two. We had this discussion yester-
           day on media. The point of the discus-
           sion was to tell you that we have spe-
           cifications for all our media and that
           we  use the best ingredients we  can
           obtain.  Speaking for my own com-
           pany, we are  not a chemical manu-
           facturer. We buy many of the ingred-
           ients  from reputable  companies.  On
           receipt, we do  quality control tests
           to insure that the product is what we
           ordered and what it says on the label.
           Just  as  the  user can  take a good
           product and ruin it in the preparation,
           we  are capable of taking good ingred-
           ients  and ending up with  a product
           that doesn't perform. That's the pur-
           pose of the quality control laboratory.
           We do not  release  any product that
           we  feel won't  perform  for the pur-
           pose  for  which  it  was   intended.

           From  the discussion,  I  understand
           that some of you have had problems. I
           personally have handled in my depart-
           ment all product reports for four of
           the last six years. I am not aware of
           any  great number of problem  prod-
           ucts so I  would ask that you would
let  us  know if you feel that you are
having a problem. This can be done in
several  ways.  If you  ever see a Bio-
Quest  representative you  can  inform
him. Most  of you probably never do,
so please call or write and tell us of
the  problem.   I  will  send you  an
authorization  form   to   return  the
product. A tremendous  amount  of
material comes  to   our  receiving
department and to  insure that it gets
into the right channels I  would  like
to provide  an authorization form. We
would  like to have return goods for
testing. If that's not possible, at least
provide the lot  number. Then I  can
check the production record and see
if  there was anything in there that
would  indicate that the product might
be any different or any suspicion that
it  has  deteriorated.  I  can't do any-
thing without the lot number.


If I can get returned goods we will set
up  an  evaluation with our reference
shelf sample of the same  lot and the
current lot and see  the  results.  We
may be able to find the problem. If
we  don't know about  it, we can't do
anything.

You are paying good money for  the
product  and we are putting a lot of
effort into  them. There is no sense for
anyone not to  use the best available.
We  can make changes. We are making
media    according   to   established
formulations of course. We are here to
serve you.  We hope  to give  you a
degree  of  stability and a  lot-to-lot
uniformity  and to take out some of
the variables. I would  encourage you
to contact  us. Of course,  one of  the
problems is turn-around  time,  but I
will  inform my people that  if any
reports  come  in  on  these types  of
products to please  let me  know.  I
don't handle every one of them indi-
vidually  any more.  I  will try to get
you  an answer as quickly  as possible.

There are some comments that I have
heard about the storage problem. If
the medium starts to  cake, you have
moisture and can have deterioration.
So  I encourage the  purchase of small
                                              169

-------
           quantities, tight covers and the use of
           a  dessicator  particularly  for small
           bottles. I'll go back and see if there is
           anything  that  we can  do  as far as
           stability.
           Someone  said  that  if  they  reorder
           there is a long lag time before they get
           the  product. This  is a  problem.  It
           could  be  our  fault through poor
           scheduling. It could be a problem due
           to lack of availability of some ingred-
           ients. We  have  made an effort in the
           past year  to improve our scheduling
           and  cut down on the backorders,  but
           I  know it  is a problem  at times. I
           would encourage you to call my of-
           fice  or write if  you prefer.  If you will
           keep in touch with us we  will do all
           we can to help you.  All of the phone
           calls, letters, product reports, etc. of a
           technical  nature come  to  my atten-
           tion. We  are there to help  you and
           you  can take advantage of  it. Thank
           you.

Winter:     I have a question for you, Dave, and
           for Aaron Lane, if he is here, EPA is
           now coming out with its own manual
           of methodology which includes a sec-
           tion on quality control.  In  it we are
           urging  a   limited holding  time  for
           media  and one  of  the problems  is
           that we can't  get the manufacturers
           to put a date on the bottles.

Power:     By law, we do  have a date on them.
           There  is  an  expiration date and  I
           don't  have the figures with me but  I
           would imagine  it is two years on these
           dehydrated materials.

Winter:     Do  you have a list  of recommended
           times  for holding different media?
           The recommended holding  times dif-
           fer with each media. Some, of course,
           like  lactose broth are pretty stable.  Is
           this list available,  I  think  it would
           guide  the  labs.  One  problem with
           some  laboratories, particularly state
            or  federal  laboratories, is  the need
            for  unique  type  media which they
            buy in quarter pounds and use only
            once in a  year.

Power:     The stability and expiration date that
           we  put on will be for the  unopened
Winter:
Power:
Winter:
Power:
bottle, because after it is opened, I
don't know. The more it is opened the
faster  it  is  going  to deteriorate. I
don't think that we can say, we have
to have data to show that a bottle of
unopened medium is good for a given
time period and expiration date. Once
it is opened, each bottle is unique and
stability  depends on whether you are
in a humid climate or a dry climate.
Of course the product has been stored
at a distributor; also it may be stored
in your  facility  and then  stored in
your  lab before its opened. The user
opens it, some put the cap on tight,
some put it on loose and some put it
on  crooked; so there is no way that
we  can  control  it.  I think  in some
cases  the  criticism  is somewhat un-
justified  because  we feel the product
that we  send out  is okay. So what
happens  after that? We have not had
expiration  dates until this year. As of
September  15, under  the  new  FDA
regulations,  every   newly  manufac-
tured product should have an expira-
tion date and every product must have
a lot  number. Perhaps you can come
up with some guidelines or some help
from  that  standpoint. I don't think
that we can.

You  are saying  that your responsi-
bility  essentially  ends  when  the
product leaves your plant, but that we
could supply guidelines for storage.
In your  manual,  you could certainly
use guidelines. We have enough to do
to justify the dating. Perhaps we are
talking abour relatively  few products
of interest  to  the  people  here. Our
company  and  our  competitors  are
talking about thousands of products.
Once  the bottle is opened, everybody
is going to handle it a little differen-
tly. I  have no objections to anything
you might  want to say about storage
or handling media in the laboratory.
We are only going to take it up to that
point and verify how long it should be
good unopened.

Could you  provide  the data on your
recommended guidelines?

Yes.
                                               170

-------
Bordner:   Could we have the comments from
           the   other  media  manufacturers?
           Would you gentlemen mind remaining
           up front for a minute because I think
           we might have questions related to
           media.

Lane:      Concerning  Dave Power's  paper, the
           comment was made this morning that
           there is a shift  in responsibility of
           quality  assurance from the manufac-
           turer to the user. I  don't  believe that
           was the intent of the paper. Certainly,
           we don't ask the user to do all the
           quality  assurance testing of a culture
           medium.  Both the  manufacturer and
           the user must have a quality assurance
           program.  Our  quality  assurance, as
           Dave pointed out  for his company,
           stops when  the bottle  leaves  our
           plant.   Every  component  of  the
           culture   medium - protein  source,
           carbohydrate  source,  buffer, indica-
           tor,  selective agents -  is tested alone
           and  in  combination.  The complete
           medium is  tested  for productivity,
           pH, and appearance  when it is bottled.
           Then the bottle is sealed,  capped and
           shipped to your laboratory.

           The  user has  responsibilities for the
           media.  Before you  use it, you must
           make sure that the medium is placed
           into   a  refrigerator if  the label  so
           states. When  the medium is delivered
           into  the laboratory  the date of receipt
           and  the  date  that the  bottle  was
           opened  are both on the labels. Don't
           open  the bottle  until the previous lot
           has been used up. That will give you a
           longer period of use. Store the medi-
           um at temperatures below 25 C in a
           low humidity environment and out of
           direct sunlight. Do not store the medi-
           um near an autoclave or drying ovens.


           When you weigh the medium, use a
           balance  and  weights  which  are  fre-
           quently checked for accuracy. Do not
           weigh  in a draft or  high humidity
           area.

           Do not  leave the dehydrated  medium
           exposed to the air. Get  it into solution
           quickly  or it can harden, then prompt-
           ly return the cap to the bottle and
tighten  securely.  Return unused por-
tions to their  proper storage. Most
dehydrated  media  are  very  hygro-
scopic and  the  ingredients may  be
sensitive to  excessive moisture and
light  and  heat.  Exposure to  such
conditions, especially  when the cap
on  the  bottle  is not  securely tight-
ened, may result in moisture  uptake
which alters the  physical, chemical
and  bacteriological  properties  of the
medium. The result could  be harden-
ing of the freeflowing powder, darken-
ing of the powder, oxidation of some
of the  components,  change  in pH,
change in solubility, change in  the ap-
pearance of the color  of the dissolved
medium and reduction or loss  of pro-
ductivity, selectivity   or differential
characteristics.

Dissolving of the medium  is another
point. If you use glassware that has
not  been thoroughly washed, residual
detergents  may  cause  low  counts.
Dirty or improperly washed glassware
can change the color of a medium,  in-
crease or decrease in pH, cause a pre-
cipitate reaction between the  residue
on  the glassware  and the medium
component, or  produce toxicity from
residual detergent.

Impure  water or water which has been
stored in a soft glass bottle  or exposed
to the laboratory air,  can  easily alter
the quality of the medium. A good
distilled  water  which  has  been in a
bottle on the shelf for two or three
weeks with a lose stopper will take up
C02  and that, too, can effect  the pH
of the medium. It can yield a precipi-
tate  in the medium or impart toxicity.


Test  the water that you are using to
dissolve the medium for conductivity,
metals  and  other  inhibitory  sub-
stances.   A  pH  of a good distilled
water is usually  between  6 and 6.5.
Do  not  use water  suspected of con-
taining  chlorine,  copper, lead or de-
tergents. We have run into this prob-
lem  when we had complaints  of cul-
ture  medium. We would send a man
to check how the distilled water was
made. In one instance, we found dis-
                                              171

-------
tilled water being made in a copper
still.

Avoid  scorching the  medium on the
bottom of a flask or other container
by stirring during  heating;  and if you
are heating an agar medium be very
careful. Distribute the medium  uni-
formly  and dissolve  completely. To
avoid overflow of the medium, do not
prepare media,  especially  agar, in a
flask less than 2% times the volume of
the medium.  Prepare  only sufficient
quantities  suitable for use  in a week
or less.

Do not attempt to readjust the pH of
a  medium  unless  you  have proper
equipment and  know the  procedure.
Adjusting the pH  of a broth  is not
really difficult,  but adjusting the pH
of an agar is quite difficult. The final
pH of a medium after autoclaving and
cooling to  25 C  is on the  label.  A
medium prepared  according  to direc-
tions in  distilled or  deionized water
and  which has not overheated during
dissolving  and  has  not   been  over
autoclaved  should   have  that  pH.

Be sure  that  the temperature  and
pressure gauges on the autoclave are
accurate.  Careful timing during auto-
claving  is essential. Remember not to
start  timing  when  the steam starts
entering the autoclave. Avoid under
or  over-autoclaving,  especially  over-
autoclaving.   Frequently  check  the
efficiency  of the  autoclave with a
biological  indicator such  as  Bacillus
stereothermophilus.  Do not keep a
sterilized  agar medium in a  water bath
at 50 C for more  than 45  to 60 min-
utes.  The agar can  settle  out  and
phosphates can precipitate. Do not
autoclave a medium  containing heat
labile enrichments or additives which
precipitate  by  heating. Since heat
penetration is slow in culture media,
especially media containing agar, it is
important that the recommended ster-
ilization period  be strictly adhered to.
The time  required  to autoclave a
medium  depends  not only  on the
efficiency of the autoclave  but the
volume of medium in the  bottle and
           its size and shape. Over-autoclaving a
           medium, especially an  agar  medium,
           can  cause:  development of  precipi-
           tate,  change in  pH, carmelization  or
           darkening,   depolymerization  of the
           agar and reduction  in gelation, reduc-
           tion in productivity, selectivity and in-
           crease in inhibitory substances.

           After the   preparation, some media
           can  be left at  room temperature  in
           screw-capped tubes. Agar plates must
           be in sealed plastic bags, preferably in
           the refrigerator. As Dave pointed out,
           the best prepared dehydrated medium
           can  be destroyed if it  is not handled
           properly, so  I  do take exception. I
           don't believe that the  manufacturer
           of  dehydrated  media wants  to  shift
           the responsibility of quality assurance
           to you. You are part of  it.

Bordner:   Aaron, would you  be willing to share
           that  material that you  read  with us?

Lane:      This  is  a   quality  assurance manual
           that  we  recently  completed.  It   is
           available. We can send  it to you  at
           any time.

Power:     In the transcript of the paper I  gave
           yesterday  is  a  reference to  a  paper
           published a couple of years ago by Dr.
           Vera who  at the  time was  head  of
           quality  control  laboratory.  It  says
           very  much the same thing  that Mr.
           Lane read.  If you write to us we will
           send  you  a copy  of our manual  as
           well.

Geldreich:  The  points you are  making are not
           new. We have evaluated state labora-
           tories for over 15 years. We have train-
           ed  state people who in turn evaluate
           the   laboratories   examining  public
           water supplies throughout  the state.
           The  laboratory certainly has an  im-
           portant  responsibility   in this area.

           Once a  medium leaves the  manufac-
           turer there are lots of things  that can
           go  wrong with  it.  I recently had one
           laboratory throw out over $200 worth
           of media which  was  caked.  We write
           strong  reports on  these subject mat-
           ters.  If we  don't have standardization
                                    172

-------
Bordner:


Lane:
or control over media, the results will
be  meaningless.  For  instance,  one
laboratory  in Hawaii  reported that
they couldn't get coliforms to ferment
brilliant  green when they  used it as
normally  prepared,   but   of  they
doubled  the  concentration  of  the
brilliant  green  they   could get gas
production.  These things do happen.
Through laboratory evaluation, EPA
will   encourage  the  people  in  the
laboratory to maintain a good quality
control program. That's their respon-
sibility.

However, there is  another form  of
quality control  where manufacturers
can  help. We  know  there  are  bad
batches of media.  This generally is
what happens: the laboratory calls the
representative of the scientific supply
house; the representative comes in and
replaced  the old batch  with  a new lot
of medium. However, others may have
been  using that bad lot of medium.
We don't know its bad unless we have
quality control  check it.  I  would
urge  the manufacturers who find a
bad  lot or batch of media to recall it
so that the  rest of us don't have to
find  this out  the hard way. In New
England where two states next to each
other  had the same  lot of medium
they  had the same  problems with it.
The  company recalled the  medium,
but I know of another state close by
that  used that medium. They didn't
quality control  it so they  consumed
the batch of medium. I am asking you
and  others who manufacture media
that  if once you find that you have a
bad  batch tell us you  will  replace it
or refund our money. Look  up in the
records and find where in the market
the product is being used and recall it.
Thank you.


Aaron, do you want to reply to this
comment?

I'm quite sure that our quality assur-
ance  laboratories do  not approve a
batch of  medium unless it performs in
the manner for which it was designed.
With the coliform MF media as I men-
tioned yesterday we use river water
           and ATCC cultures. I  am not saying
           that those state laboratories mishand-
           led the medium. Something  may have
           happened. The  problem  is that the
           manufacturers are not being informed.
           If we  were given the opportunity to
           test the sample that the state is using
           against our official sample, there may
           not have been a problem.  Our quality
           control  laboratory  must report  all
           problems and  complaints  they  are
           made aware of. Did  these state labor-
           atories  talk with our representatives?

Geldreich:  Yes, I  am told they did. I  was not
           there.  The man gave them a new lot
           and took the old lot away.

Lane:      I would just  like to say that many of
           the  products are  purchased through
           distributors and  in  the case of  per-
           formance complaints I  would ask that
           you come back to the manufacturer.
           You are probably  visited  by distribu-
           tor  representatives  far  more  often
           than you are by our own sales force.
           In  such  instances,  I  don't  know
           whether the  reports get back to us. I
           do get letters once in a while from one
           supply house in particular.

           They have a  form  letter which is very
           helpful. I think you should  talk with
           the  manufacturer. If you have other
           problems, deal with the distributor -
           he's the  one you  brought  it  from.
           When  you  have  performance  prob-
           lems which are serious, deal with the
           manufacturer.  If  you  want to call,
           that.s fine. We will take it from there.

           Without the  lot number, one doesn't
           know  how many  years that product
           has been sitting around. I  know some
           of  the distribution  houses  have  not
           rotated their stock. I  think that is
           another  advantage of  the expiration
           date. I  know that once in  Dallas, I
           started  to get reports of people  re-
           ceiving  material  that was far too old.
           It was traced back to one distributor
           who had  found  a  box that  he didn't
           realize he had and shipped it all out.
           In  the future  this  should  improve.
           If on a re-test we  find that the refer-
           ence  shelf  sample  has  deteriorated
                                              173

-------
            prior to the expiration date we would
            be obligated to do a withdrawal. We
            know to whom we shipped it. It's up
            to the distributors also to have records
            to whom  they shipped  the material.
            I  think  we have a fairly good chance
            of getting back through the distribu-
            tion  system. Anything that can hap-
            pen,  will  happen,  sometime,  some-
            where. We must  realize that we are
            dealing  with  variables  in  biological
            materials and  transportation system,
            laboratories, samples and technicians.
            We will  do the best that we can, but
            I  would encourage you to come back
            and talk with  us. At  least there is  a
            chance that it will be recalled.

Bordner:    The  key   word  is  communication,
            right?
Alico:       Its been a  fine  symposium and I want
            to offer my  gratitude  to  the two
            people who summarized the presen-
            tations.  I have  suggestions on the title
            of the  symposium.  It  reads:  "The
            Symposium on  the Recovery of Indi-
            cator  Organisms  in  Applying  Mem-
            brane Filters."  I  would  like to add
            "For Coliform and related Organisms"
            since  the  general papers,  the  stress
            papers, and comparative papers dealt
            with  coliform-related organisms. I sug-
            gest this so that when we get the pro-
            ceedings of the symposium  it will be
            clear that  it deals with this one aspect
            of membrane   filters  and  not  with
            others.

            One  other comment I have  is  about
            the  Nuclepore membrane  filters.  I
            believe about a year and half ago at
            the ASM meeting in Miami, I received
            information  from   the  people  at
            Nuclepore  on  using  these  filters for
            the enumeration  of microorganisms.
            It was  stated  yesterday  that  they
            should  not be used in enumeration
            of microorganisms. I think Nuclepore
            should notify  people that their prod-
            uct should not be used for this pur-
            pose.

 Harris:      I  think what I  have to say is germane
            to the  issue.  The last  speaker  from
            EPA  slipped  in  a  little  statement
            which  I think  is the whole reason for
our being here. He  made  the  rather
astounding  statement,  in  my  view,
that gastrointestinal  diseases were on
the wane.  This is  just not so!  The
incidence of salmonellosis in  North
America,  I'm  talking of the  whole
continent now,  is rapidly on the in-
crease.  We  are talking of notified
cases. Bear  in mind that the average
physician does not  notify.  Probably
only one in ten is notified. So  which
ever way you slice the cake the reason
for our  being here  is not simply to
design a system. Our basic reason for
being here is  the eventual reduction
in G. I.  diseases. If we take this  as our
measure, we can fail, ladies and gentle-
men, fail dismally.  I think that  it is
important that we do design a system,
whether it is the fecal coliform indica-
tor system  or not.  I'm  speaking  as a
microbiologist and as a physician. I'm
seeing both sides of the issue. I  know
what  happens  from  the  physician's
side of the fence. He is presented with
a lot of data. He is told thus and such
an incidence of such and such  an or-
ganism.  He has not been trained. Even
many of our physicians in the Public
Health field  have not been adequately
trained  to  interpret  the laboratory
data.

We have been very  smug as scientists
in designing  indicator systems, coming
up  with good systems for protecting
public health, but leaving it there. We
have not taken the trouble to bridge
the gap between the scientist and the
physician. I  think that we should not
just be satisfied in designing a system
but also make sure that there is some
carry-over from this type  of meeting
to the people who are going to go into
the field and implement our findings.
Unless we do that we can sit here for
the next 20 years designing better and
better  indicators  but not getting to
the root of  the problem which  is the
reduction of G. I.  diseases. I  would
like to correct the statement made by
a  gentleman from  EPA,  that  G.  I.
diseases are on the decrease.
I  wanted  to  speak  before the gentle-
man from BBL and  Difco  got their
                                               174

-------
           plug in.  Maddox Chemicals  supplies
           the Oxoid media in Canada exclusive-
           ly.  I go along with large measure with
           what Dr. Lane said. Many of faults are
           at the user level. I am not saying that
           the media manufacturers are faultless.
           But many of  you  who  use Oxoid
           media are aware that if you run into
           problems  we  send  people  down.  I
           would underline that we don't do any-
           thing about a bad product or bad use
           of  a product  unless we know some-
           thing about it. Thank you.


Bordner:   We are glad to have the opportunity
           to  exchange  ideas  - and  problems -
           with the media manufacturers at this
           symposium.

Hendricks:  First of  all,  let me  be the first  to
           applaud  the  physician  from Canada
           who took  exception to what I had
           said about the incidence of intestinal
           disease. I think it is true that there is
           an  increase  in  terms  of  "running
           rampant" is a  relative term. I consider
           venereal disease in this country to be
           rampant, but  whether this is due to
           reporting, or  increased actual number
           of  cases, I don't  know. I rather sus-
           pect that it is a combination of the
           two. I will  stand by what I said earlier
           that where the coliform tests are used
           to  monitor water quality, I think the
           rate of  serious intestinal  disease is
           much lower than those areas that do
           not employ   such  procedures.  Of
           course,  these  microbiological proce-
           dures have to be followed with ade-
           quate treatment of  some sort.  I do
           believe that they have served us well.
           The mechanisms by which  intestinal
           disease seems to  be  increasing  un-
           doubtedly  is  due  to  a  variety  of
           things. I would hate to see us abandon
           our  coliform  procedures, to say that
           they are no good and throw them out,
           because  I  think  the results of not
           monitoring water  quality  would be
           disastrous.  This is  really what  I was
           saying  about  the significance of any
           technique  that's  going to  have a
           tendency  to   increase  numbers.  We
           know that  our measurements of coli-
           form procedures may at times be low.
           This may very well be one of the rea-
           sons why we can recover pathogens,
           bacterial pathogens  including viruses
           in water that appear to be of excellent
           quality. Again, I  applaud the gentle-
           man, but I think there are two sides to
           the picture and we should not elimi-
           nate procedures that have served us
           well. I think  we should improve upon
           them and be well aware of what we
           are counting. Thank you.


              FINAL REMARKS

Frith:      I  think everyone  agrees that  this was
           a very timely conference and we had
           an opportunity to share some of  our
           general  knowledge as well as to open
           up  some  additional  avenues for in-
           depth  study. I don't think that any-
           one will leave thinking that they  got
           short-changed from this meeting. It is
           important to understand the function,
           and  I don't want to give you a long
           detailed  and  boring  explanation  of
           ASTM  but  as  your  co-chairman  for
           the  sub-committee  I  would like  to
           explain the real function of ASTM in
           this whole operation and ask you or
           invite  you to  become a participant.

           It is  obvious that the only way  to
           eliminate  confusion  is to  develop a
           test method that will help everyone to
           know  that they are  buying  a  stand-
           ardized product,  that will represent
           the state of the art.  For about three
           years,  ASTM  has  been trying to do
           this  with just  the membrane  filter,
           not the media  - nor  the various con-
           trols. We have written three  draft
           procedures: one on recovery that you
           have heard about today, one on in-
           hibitory effects and  one other being
           proposed  today  on  retention. This
           afternoon we will be working on two
           or three of these procedures that have
           been round  robin tested, from which
           you  have seen  some  of the data.  We
           are trying to eliminate some of  the
           confusion.

           As an  ASTM member, you become a
           voter, a person who  has a  chance to
           see the draft copies and  to submit
           negative ballots if you feel  that there
           is something  wrong with the  way  the
                                              175

-------
           technique is being run or if you have
           data that does  not support what is
           being presented. These negative bal-
           lots or the additional data supplied
           have to be overridden to either prove
           or disprove your point.                 Ginsberg:

           Active membership and voting mem-
           bership is available by  joining the   Frith:
           ASTM Society. You are not restricted
           from  attending  any of our meetings
           as  a  non-ASTM member.  You can
           come,  you can  participate, you can
           offer  data  and comments.  ASTM
           would like  to  see  you  become  an
           active  member and be involved with
           us. We hope not  only to develop   Ginsburg:
           specifications for recovery, inhibitory
           effects and retention, but also to con-
           sider  extractables,  surface   charac-   Frith:
           teristics and other general topics such
           as membrane composition and impuri-
           ties. You  can  see the charter that we
           have ahead of us is quite broad.

           We will  be  doing the same thing, I
           am sure, with various types of media.
           We will  be  determining  how you
           actually  monitor  a temperature  of
           44.5  + 0.2  C.  All of  these issues
           will be coming  out of ASTM  in the
           next  few years. Your active  partici-
           pation can help  that move faster. Dr.
           Litsky has said, "Let's either write the
           specs by the state of the art and make
           them  meaningful or  forget  it." I'm
           thankful that  each  of  you has taken
           time for a day and a half to attend. I   Ginsberg:
           think it  has been a fantastic  sympo-
           sium  and  because of your  participa-
           tion,  we  will gain a  lot  more and
           move a  lot quicker than we have done   Frith:
           in the last ten years. Again I say, you
           are invited this afternoon to come and
           listen and participate.  Are there any
           questions about ASTM?
           test methods on  how do you really
           monitor the variations in membranes
           or monitor  the  difference in water
           samples, etc?

           This same idea,  is it carried through
           the week?

           No. The  test method will only carry
           through  this afternoon. Tomorrow,
           there  is going to be an organizational
           meeting for indicator organisms.  We
           are concerned not only with coliforms
           and  fecal  coliforms  but also with
           different indicator organisms.

           What  is the  procedure for participat-
           ing in  any future  round  robins?

           Well,  you'll  hear about that  at  the
           meeting today. If you as a laboratory
           representative  would  like to partici-
           pate you are welcome. We want to get
           enough  different  labs   around  the
           country to  get a representative sam-
           ple.  If you  feel  that your lab  can
           spend the time (it  is a  very exicting
           program  but it takes an awful lot of
           money and  time to  do what is re-
           quired) you will be getting notes as a
           member  of  the committee,  and you
           will be notified that a round robin  will
           be taking place.  Margareta Jackson,
           who unfortunately could not be here,
           is the one actually in charge of  the
           afternoon session.

           Do I  understand  that ASTM supplies
           the materials or  do we have to  buy
           them?

           In the first round robin  we asked  the
           membrane  manufacturers  and   one
           media manufacturer to  supply mater-
           ials but  you will make the  biggest
           contribution which is  man  hours.
Ginsburg:  This meeting  that  is coming up this
           afternnon, will it deal with the same
           problems that the  symposium  cover-
           ed?

Frith:      That is correct. We hoped to learn a
           lot from the  symposium and to use
           the knowledge we have gained the last
           day  and a half to make meaningful
Vlassoff:   When the people fill out these forms,
           won't  they  have some idea what's
           happening?

Frith:      Yes. That  is an  ASTM form, if you
           have not filled out one of these, you
           should. We will be submitting to you
           everything  that has happened at this
           symposium. How fast several of us can
                                              176

-------
Question:
Frith:
get together and  either  get these
papers  in  some form to send out or
get into a  publication  with which
both ASTM and EPA are concerned
will depend on  some man hours. We
will keep  you informed if you leave
your name and address, whether  you
are an ASTM member or not, by  this
route.

If  this  bulletin  is  published, will  it
go automatically to ASTM members
or will this be a personal-type mailing?

We are not sure right now. I think the
question has yet to be resolved - how
and in  what form this symposium be
                                                 Bordner:
provided to  you.  It will not be pub-
lished until  some degree of satisfac-
tion has been reached between ASTM
and EPA, as they were the sponsors.
Again, many thanks for your time and
your talents  and we will look forward
to those of  you  who would  like to
join  us  at 1:30  back  in this  room.

I  want  to thank  all of you for par-
ticipating in this symposium - those of
you who made special  efforts in pre-
paring papers and  presenting  them,
who  participated by offering thought-
provoking comments,  and who sum-
marized  what was said. To  all of
you  we owe particular  appreciation.
                                             177

-------
                                          APPENDIX

                        The following paper intended for the Seminar was
                        not received  in time for presentation. However, it
                        is relevant and is presented here for the considera-
                        tion of the readers.
            COMPARISON OF MEMBRANE FILTER COUNTS AND PLATE COUNTS
                 ON HETEROTROPHIC AND OIL AGAR USED TO ESTIMATE
                      POPULATIONS OF YEAST, FUNGI AND BACTERIA
                  J. D. Walker, B. F. Conrad,  P. A. Seesman  and R. R. Colwell
                                  Department of Microbiology
                                    University of Maryland
                                 College Park,  Maryland 20742
                 ABSTRACT

     Comparison of  filter  and plate  counts of
yeasts and fungi on heterotropic and oil agar re-
vealed higher counts were  nearly always obtained
with filters. Comparison of filter and plate counts
of heterotrophic estuarine and marine bacteria re-
vealed that, on the  average, filter counts were 80%
lower than the  plate counts. These results should
be considered when evaluating methods for enumer-
ation of microorganisms in the marine environment.

               INTRODUCTION

     Microorganisms present in seawater are usually
enumerated  using  membrane  filters  because of
their low numbers. However, in the estuarine en-
vironment, generally characterized by larger micro-
bial  populations, microorganisms  are  frequently
enumerated using plate counts. In the study report-
ed here, Chesapeake Bay water was tested, permit-
ting comparison of plate and filter counts for yeast,
fungi and bacteria.

         MATERIALS  AND METHODS

     Heterotrophic  yeasts and fungi were enumer-
ated using yeast medium,  details  of  which have
been  published  elsewhere  (1). Yeasts and fungi
capable of growth on petroleum were enumerated
using oil  agar #2, pH adjusted to 5.5 and supple-
mented  with streptomycin  and  tetracycline (50
ug/ml of each), as described by Walker and Colwell
(2). Heterotrophic bacteria were enumerated using
the basal  medium  of Walker and  Colwell  (D.
Bacteria capable of growth  on petroleum were
counted using a silica gel-oil medium (3). All plates
were incubated at 15 C for two weeks.

         RESULTS AND DISCUSSION

     Heterotrophic yeasts and  fungi count plates
were incubated for a minimum  of one week to in-
sure  appearance  of the maximum  number of
colonies (Tables 1 and 2). The results of the hetero-
trophic  counts  compared with counts for yeasts
and fungi  on oil agar plates incubated for  at least
two  weeks are  shown in Tables 3  and 4. It  was
necessary to examine the plates periodically, after
the initial  three days of incubation,  to avoid over-
growth with fungi.

     Before plate  counts  were compared with the
filter counts, a percent comparison was calculated
between the plate and filter replica plate counts
(Tables  5  and  6). These calculations indicated at
least 60% comparison between replicate counts was
obtained.  Similar  comparability was observed for
results of the yeast and fungi replicate  counts.

     Comparison of the  plate and filter counts of
heterotrophic yeast showed that the  highest counts
were always obtained when membrane filters were
used.  The percent comparison  was  always indica-
                                             178

-------
tive  of  a  significantly lower count on  the plates
compared with the  filters (Table  7). In most in-
stances, higher counts of heterotrophic fungi were
obtained using membrane filters,  compared with
the spread plate technique (Table 8). Half of the
percent similarities were significant. The number of
yeasts growing on oil agar, using the filter method
of enumeration, was higher than  on the oil agar
plates.  In  only one case did the two methods give
a reasonably close count (Table 9). As in the case
of the  heterotrophic fungi, higher counts of fungi
on oil  agar were obtained using membrane filters,
than with plate counts, although the counts were
minimally comparable (Table 10).  Results obtained
when plate and filter counts of yeasts and fungi on
heterotrophic and oil agar were compared suggested
that quite different counts were obtained by these
methods and that filter counts yielded higher num-
bers of  yeasts and fungi  when estuarine and marine
water and sediment samples were examined.

     Comparison  of  plate and  filter counts of
heterotrophic bacteria indicated that plate counts
yielded higher counts in most  samples, and only a
few samples gave  similar results by plate and filter
counts (Table 11). Plate and filter counts of bacteria
on silica gel medium  were not high enough to com-
pare results (Table 12). The  filter procedure gave
higher  counts on  silica gel oil for bacteria from
sediment  and  from  Eastern  Bay  water,  whereas
plate counts provided the best estimation of the
bacterial populations in Colgate Creek water enu-
merated on silica gel-oil medium.

     The two areas in Chesapeake  Bay included in
this study, Colgate Creek  in Baltimore Harbor and
Eastern Bay, differ ecologically. Colgate Creek  is
continuously  contaminated   with  oil,  whereas
Eastern Bay is an  oil-free, commercially productive
shellfish area.  The higher  counts  of yeast, fungi
and bacteria on oil media for Colgate Creek samples
can be  explained.  By virtue of larger populations,
more comparisons of plate counts and filter counts
were possible using  inocula  from  Colgate Creek
than from Eastern Bay.

     Plate and filter counts of seawater collected at
stations along the U.S.  east coast were compared
and, generally  speaking,  plate counts  of micro-
organisms  in  seawater,  concentrated  using an
Aminco concentrator (American  Instrument Co.,
Silver Spring, Md.) yielded higher counts than the
filter procedure (Table 13). By dividing the average
filter count by the average plate count and multi-
plying by  100,. a comparison was possible for estua-
rine  and   concentrated  seawater  samples.  The
average filter count for estuarine samples was 19.7%
of the average plate count and 19.2% for the con-
centrated seawater. Thus, for the average of the 14
estuarine samples  and 10 marine samples studied,
filter counts were  about 80% lower than the plate
counts.

     Nuclepore (General Electric, Pleasanton, Calif.)
filters were compared  with Millipore  (Millipore
Filter  Corp., Bedford,  Mass.) filters  by scanning
electron  microscopy  (Todd and  Kerr (4)),  but
not  for  efficiency  in  enumeration  of bacteria.
Nuclepore filters are thin (10 urn) films of polycar-
bonate plastic,  with  an average  pore  diameter ap-
proximating the individual pore diameter. Millipore
filters are thick (150 um) films of cellulose, the
average pore diameter differing  significantly from
the  individual  pore  diameters.  Millipore  counts
were compared with  Nuclepore  counts.    Data
for the estuarine samples showed no trend toward
higher counts for Millipore  filters in comparison
with  Nuclepore filters (Table 14).  This was unlike
the  marine samples which  always  gave higher
counts with Millipore filters (Table 15). Two dis-
tinct disadvantages of  using Nuclepore filters in
laboratories aboard ocean research vessels are that
Nuclepore membranes are markedly thin, making
it  difficult  to  place  them  on sintered  glass filter
holders and on agar surfaces. Air bubbles trapped
between  the agar and  the  filter are  often indis-
tinguishable from translucent  bacterial colonies.
     As a general  statement, membrane filters are
more efficient for counting  yeasts and fungi but
plate counting  for estuarine  and marine  water
bacteria is recommended over the  membrane filter
method.  The  ease of use and their sturdier nature
make  Millipore filters preferable to  Nuclepore
filters for field work.

             ACKNOWLEDGMENT

     This work was supported by Contract No.
N00014-67-0239-0027 between the Off ice of Naval
Research and the University of Maryland.
                REFERENCES

1.    Walker,  J.D.  and R.R.  Colwell.  Microbial
     Degradation of Model Petroleum at  Low Tem-
     peratures. Microbiol. Ecol. 1:63-95, 1974.
2.    Walker,  J.D.  and  R.R.  Colwell.  Factors
     Affecting Enumeration and  Isolation of Acti-
     nomycetes  from Chesapeake Bay and  South-
                                               179

-------
    eastern Atlantic Ocean sediments. Mar. Biol.          Petroleum-Degrading  Microorganisms. Micro-
    In Press, 1975a.                                  bial Ecol. Submitted,  1975b.
3.   Walker, J.D. and R.R. Colwell. Microbial De-     4.   Todd, R.L. and T.J.  Kerr. Scanning Electron
    gradation of  Petroleum:  Enumeration  of          Microscopy and Microbial Cells on Membrane
                                                   Filters. Appl. Microbiol. 23:1160-1162, 1972.
 Table 1.    EFFECT OF INCUBATION TIME ON FILTER COUNTS OF HETEROTROPHIC YEASTS

Incubation
time
(days)

3





14


Dilution/Volume
filtered
(ml)
0/100
0/250
-2/10
0/100
0/250
-2/10
0/100
0/250
-2/10
Counts
Colgate
Creek
sediment


2


2


2
for
Colgate Eastern
Creek Bay
water sediment
1
3
<1
2
OQ3
<1
2
OG
<1

Eastern
Bay
water
1
3

2
3

2
3

 aOvergrown with fungi
 Table 2.    EFFECT OF INCUBATION TIME ON FILTER COUNTS OF HETEROTROPHIC FUNGI

Incubation
time
(days)

3


7


14


Dilution/volume
filtered
(ml)
0/100
0/250
-2/10
0/100
0/250
-2/10
0/100
0/250
-2/10
Counts
Colgate
Creek
sediment


2


11



for
Colgate Eastern
Creek Bay
water sediment
5
9
<1
10
OGa
3
10
OG


Eastern
Bay
water
2
12

5
19

5
19

 aOvergrown with fungi.

                                          180

-------
Table3.   EFFECTOR INCUBATION TIME ON FILTER COUNTS OF YEASTS ON OIL AGAR

Incubation
time
(days)

3


7


14


Dilution/volume
filtered
(ml)
0/100
0/250
-2/10
0/100
0/250
-2/10
0/100
0/250
-2/10
Counts for
Colgate Colgate Eastern
Creek Creek Bay
sediment water sediment
<1
<1
<1 <1
<1
OGa
<1 <1
5
OG
1 <1

Eastern
Bay
water
<1
<1

1
2

1
2

aOvergrown with fungi.
 Table 4.    EFFECT OF INCUBATION TIME ON FILTER COUNTS OF FUNGI ON OIL AGAR

Incubation
time
(days)

3


7


14


Dilution/volume
filtered
(ml)
0/100
0/250
-2/10
0/100
0/250
-2/10
0/100
0/250
-2/10
Counts
Colgate
Creek
sediment


<1


3


3
for
Colgate
Creek
water
0
2

2
OGa

OG
OG


Eastern Eastern
Bay Bay
sediment water
0
0
<1
3
6
<1
3
6
<1
 aOvergrown with fungi.
                                   181

-------
Table 5.   COMPARISON OF REPLICATES FOR DUPLICATE PLATE COUNTS OF CHESAPEAKE
         BAY BACTERIA

                     Bacteria/ml
Replicate 1
5.2 x 105
1.7 x 104
6.4 x 102
1.4 x 102
2.3 x 103
7.3 x 104
3.0 x 101
2.8 x 102
4.3 x 104
1.7 x 103
1.0 x 105
Replicate 2
4.7 x 105
2.7 x 104
1.0 x 103
1.2 x 102
1.9 x 103
7.1 x 104
4.0 x 101
2.0 x 102
4.2 x 104
1.5x 103
1.6 x 105
Percent comparison
90.4
62.9
64.0
85.7
82.6
97.3
75.0
71.4
97.7
88.2
62.5
Table 6.   COMPARISON OF REPLICATES FOR DUPLICATE FILTER COUNTS OF CHESAPEAKE
         BAY BACTERIA

                     Bacteria/ml
Replicate 1
30
15
33
44
51
45
50
Replicate 2
20
10
25
37
39
42
41
Percent comparison
66.6
66.6
75.6
84.1
76.5
93.3
82.0
                                     182

-------
Table 7.    COMPARISON  OF  PLATE AND  FILTER COUNTS OF HETEROTROPHIC  YEASTS
Inoculum
Sediment
Sediment
Sediment
Water
Water
Water
Sediment
Sediment
Sediment
Water
Water
Water

Source Plate
Colgate Creek 3.0 x 101
5.0 x 10°
5.0 x 101
Colgate Creek 1.0x10°
<10°
5.0 x 10°
Eastern Bay <101
<101
<101
Eastern Bay <10°
<10°
<10°
Yeasts/ml
Filter
1.7x 102a
1.0x 101
2.0 x 102
1.7x 101b
4.0 x 10'1b
1.0 x 10°b
5.0 x 10°
<^101
<^1Q1
1.0x 10'2C
5.0 x 10-3C
2.0 x 10-2C
Percent
comparison
17.6
50.0
25.0
5.8
N.D.d
20.0
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
aResults obtained by filtering 10 ml of a 10~2 dilution of sediment samples in this and subsequent tables
 for yeasts and fungi.
bResults obtained by filtering 100 ml of Colgate Creek water in this and subsequent tables for yeasts and
 fungi.
cResults obtained by filtering 1000 ml of Eastern  Bay water in this and subsequent tables for yeasts and
 fungi.
     determined.
TableS.    COMPARISON OF  PLATE AND FILTER  COUNTS OF HETEROTROPHIC FUNGI
Inoculum
Sediment
Sediment
Sediment
Water
Water
Water
Sediment
Sediment
Sediment
Water
Water
Water
Fungi/ml
Source Plate
Colgate Creek 1.0 x 103
7.0 x 102
1.0 x 103
Colgate Creek 1.5 x 10°
<1QO
5.0 x 101
Eastern Bay <101
6.0 x 101
<101
Eastern Bay <10°
<1QO
1.0 x 10°

Filter
2.0 x 103
9.0 x 101
1.0 x 103
1.7 x 10°
2.5 x 10°
1.0 x 10-1
<101
5.0 x 101
3.0 x 102
5.0 x 10-2
7.0 x 10~2
5.0 x 10'2
Percent
comparison
50.0
12.9
100.0
88.2
N.D.
0.2
N.D.
83.3
N.D.
N.D.
N.D.
5.0
                                          183

-------
Table 9.    COMPARISON OF PLATE AND FILTER COUNTS OF YEASTS ON OIL AGAR
Inoculum
Sediment
Sediment
Sediment
Water
Water
Water
Sediment
Sediment
Sediment
Water
Water
Water
Table 10.
Yeasts/ml
Source Plate
Colgate Creek 3.0 x 101
1.0x101
Colgate Creek 1.0x 10°
<1QO
<10^
Eastern Bay 5.0 x 10°
<10^

-------
Table 11.   COMPARISON OF PLATE AND FILTER COUNTS OF HETEROTROPHIC
         ESTUARINE BACTERIA
Inoculum
Sediment
Sediment
Sediment
Sediment
Sediment
Sediment
Water
Water
Water
Water
Sediment
Sediment
Sediment
Water

Source Plate
Colgate Creek 3.8 x 105
4.5 x 106
6.4 x 105
2.8 x 104
3.2 x 106
1.9 x 105
Colgate Creek 2.9 x 104
9.1 x 104
2.5 x 103
8.1 x 104
Eastern Bay 9.3 x 104
4.7 x 105
1.6 x 104
Eastern Bay 2.2 x 102
Bacteria/ml
Filter
2.6 x 105
4.0 x 105
1.0 x 105
3.2 x 104
7.0 x 105
2.5 x 104
1.2 x 105
6.7 x 104
4.8 x 103
6.8 x 104
5.6 x 104
2.0 x 104
1.0x 104
3.0 x 101
Percent
comparison
68.4
8.8
15.6
87.5
21.9
13.2
4.1
73.6
52.0
84.0
60.2
4.2
62.5
13.6
                                   185

-------
Table 12.  COMPARISON OF  PLATE AND FILTER COUNTS OF BACTERIA ON SILICA GEL-OIL
         MEDIUM
                                                   Bacteria/ml
Inoculum
Sediment
Sediment
Sediment
Sediment
Sediment
Water
Water
Water
Water
Water
Sediment
Sediment
Sediment
Sediment
Sediment
Water
Water
Water
Water
Water
Source Plate
Colgate Creek <102
<102
<102
<102
<102
Colgate Creek 5.0 x 101
3.5 x 102
1.5x 102
1.5 x 102
5.0 x 101
Eastern Bay <102
<102
<102
<102
<102
Eastern Bay <102
1.0 x 10°
<10°
<1QO
<1QO
Filter
3.0 x 101
3.0 x 101
1.7 x 102
2.0 x 101
1.0x 101
<1.0x 10-1
<1.0x 10-1
<1.0x 10-1
<1.0x 10-1
6.1 x 10°
1.8 x 102
5.0 x 10°
<1.0x 10-1
<1.0x 10'1
<1.0x 10-1
9.0 x 10-2
<1.0x 10"2
<1.0x 10'2
7.5 x 10-2
5.0 x 10"2
    Table 13.   COMPARISON OF FILTER AND PLATE COUNTS OF HETEROTROPHIC MARINE
             BACTERIA
                                     Bacteria/ml
Inoculum
Sea water concentrate
Sea water concentrate
Sea water concentrate
Sea water concentrate
Sea water concentrate
Sea water concentrate
Sea water concentrate
Sea water concentrate
Sea water concentrate
Sea water concentrate
Plate
3.4 x 105
1.5x 105
1.8x 106
5.0 x 105
9.0 x 104
2.0 x 106
3.0 x 105
2.6 x 103
6.2 x 102
1.1 x 104
Filter
4.8 x 105
5.0 x 104
1.8x 105
1.4 x 105
7.4 x 104
5.6 x 104
6.0 x 103
7.1 x 101
7.2 x 102
3.5 x104
Percent comparison
70.8
33.3
10.0
28.0
82.2
2.8
2.0
2.7
86.1
31.4
                                      186

-------
Table 14.  COMPARISON  OF FILTER  COUNTS OF ESTUARINE BACTERIA  OBTAINED ON
         MILLIPORE AND NUCLEPORE FILTERS
Inoculum
Sediment
Sediment
Sediment
Sediment
Sediment
Sediment
Sediment
Water
Water
Water
Water
Sediment
Sediment
Sediment
Sediment
Sediment
Sediment
Sediment
Water
Table 15.

Source Millipore
Colgate Creek 2.6 x105
4.0 x 105
1.0x 105
3.2 x 104
7.0 x 105
2.5 x 104
20
Colgate Creek 1.2x 105
6.7 x 104
4.8 x 103
6.8 x 104
Eastern Bay 5.6 x 104
30
2.0 x 104
16
1.0 x 104
6
3
Eastern Bay 3.0 x 101
Bacteria/ml
Nuclepore
2.8 x 105
6.0 x 105
1.0x 105
3.2 x 104
1.9 x 106
1.0 x 104
8
7.9 x 104
8.3 x 104
3.4 x 103
5.3 x 104
7.4 x 104
30
1.0 x 104
2
1.2x 104
12
3
1.0x 101
COMPARISON OF FILTER COUNTS OF HETEROTROPHIC MARINE
OBTAINED ON MILLIPORE AND NUCLEPORE FILTERS
Bacteria/ 100 ml on





Inoculum Millipore
Sea water 360
Sea water 162
Sea water 260
Sea water 268
Nuclepore
340
147
106
100
Percent
comparison
92.8
66.6
100.0
100.0
36.8
40.0
40.0
65.8
80.7
70.8
77.9
75.7
100.0
100.0
12.5
83.3
50.0
100.0
33.3
BACTERIA
Percent
comparison
94.4
90.7
40.8
37.3
                                    187

-------
                                  TECHNICAL REPORT DATA
                           (Please read Instructions on the reverse before completing)
1. REPORT NO.
  EPA-600/9-77-024
                                                          3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
  Proceedings of the Symposium on  the  Recovery of
  Indicator Organisms Employing Membrane Filters
                                  5. REPORT DATE
                                   September 1977 issuing date
                                  6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)

 Robert H.  Bordner, Clifford F. Frith and John A. Winter
                                                          8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
 Environmental Monitoring &  Support Lab.
 Office of Research and  Development
 U.S. Environmental Protection Agency
 Cincinnati, Ohio 45268
                  -  Cin.,  OH
                            10. PROGRAM ELEMENT NO.

                                     .1BD612A
                                  11. CONTRACT/GRANT NO.

                                        In-House
12. SPONSORING AGENCY NAME AND ADDRESS
         SAME AS ABOVE
                                                           13. TYPE OF REPORT AND PERIOD COVERED
                                                                     Final        	
                                  14. SPONSORING AGENCY CODE

                                     EPA/600/06
15. SUPPLEMENTARY NOTES
 Held at the meeting of  the  American Society for Testing and Materials,  Committee D-19
 on Water, Ft. Lauderdale, Florida.   January 20-21, 1975.
16. ABSTRACT
     The Symposium on the Recovery of Indicator Organisms Employing Membrane Filters
 brought together users, manufacturers, research scientists and  representatives of
 government agencies to exchange technical information and review the performance of
 membrane filters for water  and wastewater analyses.  Problems with the recovery of
 bacterial indicators had been reported;they were most pronounced in the fecal coliform
 test.  A key question was whether the cause was differences  in  sample types, membrane
 filters or the test method  employed.
     Professionals experienced in water analysis presented relevant field experiences,
 laboratory data and research findings and discussed problems concerning recovery of
 organisms stressed or injured by environmental factors.  Media,  transport phenomena,
 physical and chemical characteristics of membranes, membrane sterilization methods,
 incubation temperatures, techniques for comparison of methods,  data analysis, and the
 status of the proposed ASTM methods for evaluating membrane  filters were discussed.
     Solutions suggested at  the Symposium included use of two-step incubation, overlay
 and/or enrichment techniques and modification of membrane filter structures.  Recom-
 mendations were made to manufacturers and to users to develop and improve intralabora-
 tory quality control programs, to standardize interlaboratory testing procedures, to
 participate in these collaborative studies and to generally  improve communications
 among users, manufacturers  and standard-setting organizations.	
17.
                               KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                             b.lDENTIFIERS/OPEN ENDED TERMS
                                               c.  COSATI Field/Group
 Aquatic microbiology
 Coliform bacteria
 Indicator organisms
 Membranes
 Fluid Filters
 Water
 Water analysis
Water
Water
Tests
microbiology
pollution
Bacterial tests
Environmental  factors
Fecal coliforms
Fecal streptococci
Membrane characteristics
Recovery of bacteria
Stressed microorganisms
06L
06M
13B
18. DISTRIBUTION STATEMENT
     Release to Public
                                              19. SECURITY CLASS (ThisReport)
                                                UNCLASSIFIED
                                                21. NO. OF PAGES
                                                     200
                     20. SECURITY CLASS (Thispage)

                        UNCLASSIFIED
                                                                        22. PRICE
EPA Form 2220-1 (9-73)
                                            188
                                         U. S. GOVERNMENT PRINTING OFFICE: 1977-759-092

-------

-------
0)  to
^ CB
o  5
•?  CD'
  Go
   9- 3
           

si
o- CD  g.
5. ?• 3
   ft) 0]

   C "•*
   ""s ^D
   3 o
   o
   CD
CD  CD
   "S3
   o
          "
        §
           8-
           o





r
\
>j

•




z
rn
D
C
|—

0
PPORTU
z
H
m
•D
n
O
m

m
Z
p
H
"S
-n
O
"0
j>
H
m
C
CO
m
CO
o
0



0
Tl
Tl
ICIAL BU
0)
Z
m
0)
en


o

^
5'
a>
O
^"
o'
en
O)
00




— i
CD
O
IT

O
0)
o'
B
0>
o'
ZJ
CO
(-*•
QJ



g
CD
O
— *»
33
esearch
0)
Q.
Develoi
T3
3
CD
i-^-
C
m
55
O
z
<^
m
z
H
T)
3)
O
—1
ECTIOh
£—
O
m








CD
o
o
CO
-Q
CD
0
"
_

T]
0
C
-1
C
w
m
Z
^
o
z

m
Z

J >
i r
CJ Tl

Tl
O
(f)
-\
>
C)
m
>
O
                                                                                   o
                                                                  n


                                                                  (A
                                                                  01
                                                                         01
                                                                               O  m
                                                                               H  m
                                                                               rn  CO
                                                                               0  13
                                                                               H

                                                                               O
                                                                               Z
                                                                                >
                                                                                D
                                                                            Q
                                                                            m
                                                                            Z
                                                                            O

-------