EXCHANGE

)IO3KA-STATE COLLEGE
OF AGRldJLTURE AND MECHANIC ARTS
OFFICIAL PUBLICATION

Vol. XX ' December 28, 1921 No. 31

BACTERIA FERMENTING LACTOSE

and
THEIR SIGNIFICANCE IN WATER ANALYSIS

By
MAX LEVINE

62

ENGINEERING, EXPERIMENT STATION
AMES, IOWA

Published weekly by Iowa State College of Agriculture and Mechanic Arts, Ames, Iowa.
Entered as Second-class matter, and accepted for mailing at special rate of postage provided
for in Section 429, P. L. & R., Act August 24, 1912, authorized April 12, 1920.

PURPOSE OF THE STATION

The purpose of the Engineering Experiment
Station is to afford a service, through scientific
investigations, evolution of new devices and
methods, and tests and analyses of materials:

For the manufacturing and other engineering
population and industries of Iowa;

For the industries related to agriculture, in the
solution of their engineering problems;

For all people of the state in the solution of the
engineering problems of urban and rural life.

IOWA STATE COLLEGE

OF AGRICULTURE AND MECHANIC ARTS

OFFICIAL PUBLICATION

Vol. XX December 28, 1921 No. 31

BACTERIA FERMENTING LACTOSE

and
THEIR SIGNIFICANCE IN WATER ANALYSIS

By

MAX LEVINE

Bacteriologist
IOWA ENGINEERING EXPERIMENT STATION

and
Associate Professor of Bacteriology, .

IOWA STATE COLLEGE, AMES, IOWA

BULLETIN 62

ENGINEERING EXPERIMENT STATION
AMES, IOWA

Published weekly by Iowa State College of Agriculture and Mechanic Arts, Ames, Iowa.
Entered as Second-class matter, and accepted for mailing at special rate of postage provided
for in Section 429, P. L. & R,, Act August 24, 1912, authorized April 12, 1920.

STATE BOARD OF EDUCATION

Members

Hon. D. D. Murphy, President Elkader

Hon. George T. Baker Davenport

Hon. Chas. R. Brenton Dallas Center

Hon. P. K. Holbrook. Onawa

Hon. Edw. P. Schcentgen Council Bluffs

Hon. C. H. Thomas Creston

Hon. Pauline Lewelling Devitt Oskaloosa

Hon. W. C. Stuckslager Lisbon

Hon. Anna B. Lawther Dubuque

Finance Committee

Hon. W. R. Boyd, President Cedar Rapids

Hon. Thomas Lambert Sabula

Hon. W. H. Gemmill, Secretary Des Moines

ENGINEERING EXPERIMENT STATION

Station Council

(Appointed by the State Board of Education)

Raymond A. Pearson, LL D President

Anson Marston, C. E Professor

Louis Bevier Spinney, B. M. E Professor

Warren H. Meeker, M. E Professor

Fred Alan Fish, M. E. E. E Professor

Allen Holmes Kimball, M. S Professor

0. R. Sweeney, M. S., Ph. D Professor

Fred R. White, B. C. E Chief Engineer, Iowa Highway Commission

Station Staff

Raymond A. Pearson, LL. D President, Ex-officio

Anson Martson, C. E Director and Civil Engineer

John S. Dodds, C. E Bulletin Editor

Warren H. Meeker, M. E Mechanical Engineer

Fred Alan Fish, M. E. E. E Electrical Engineer

Allen Holmes Kimball, M. S Architectural Engineer

O. R. Sweeney, M. S., Ph. D Chemical Engineer

Charles S. Nichols, C. E Sanitary Engineer

Louis Bevier Spinney, B. M. E Illuminating Engineer and Physicist

William J. Schlick, C. E Drainage Egineer

T. R. Agg, C. E Highway Engineer

John Edwin Brindley, A. M., Ph. D Engineering Economist

Max Levine, S. B Bacteriologist

Lulu Soppeland, M. S Assistant Bacteriologist

J. H. Griffith, M. S Structural Materials Engineer

D. A. Moulton, B. S. in Cer. Eng Ceramic Engineer

A. K. Friedrich, E. M Mining Engineer

G. W. Burke, B. S. in Chem. Eng Chemist

C. H. Giester, B. S Assistant Chemist

Clyde Mason, B. S. in E. E., B. S. in C. E. Engineer

C. J. Myers, B. S. in M. E Mechanical and Electrical Engineer

Geo. W. Rodgers Mechanician

Table of Contents

Page

I. Characteristics of the Colon Group of Bacteria 5

II. Evidence of Two Subdivisions in the Colon Group and Tests for their Differen-

tiation 17

III. Classifications of the Colon Group of Bacteria 32

IV. The Detection of the Colon Group in Water 43

V. The Colon Group as an Index of Pollution 65

VI. The Spore Forming Lactose Fermenters, and their Significance in Water Analysis 91

Appendix A. Routine Methods of Water Analysis and the Colon Index 99

Appendix B. Culture Media , 108

References . -.119

I. CHARACTERISTICS OF THE COLON GROUP OF BACTERIA*

The bacteriological analysis of water is an indirect and quantitative
one. Specific pathogenic organisms are not sought nor are they likely to
be detected even in a dangerous water. It devolves upon the analyst to
interpret his findings and particular emphasis is placed on the determination
of the presence of the colon group. The investigator and analyst should
therefore be thoroughly acquainted with the characteristics, peculiarities,
and idiosyncrasies of the organisms in the group, particularly with refer-
ence to their distribution, viability, and differential reactions.

Bacterium coli was first isolated by Emmerich from the feces of a
cholera patient in 1884. It was soon recognized as a normal inhabitant
of the intestinal tract of man and of other animals. For the past three
decades the colon group of bacteria has been extensively studied by bacter-
iologists and sanitarians especially those interested in water supply and
purification. Probably as much work has been done on this as on any
other group of bacteria but there is not as yet an absolute agreement as to
the limitations of the group.

Limitations of the Group. In the 1905 report of the Committee
on Standard Methods for Water Analysis of the American Public Health
Association a series of tests were described which were supposedly char-
acteristic of the colon group. These tests included morphology, motility,
fermentation of glucose, coagulation of milk, production of indot, reduction
of nitrates, and gelatin liquefaction. Many of these reactions, as gelatin
liquefaction, nitrate reduction, and indol formation, require a long incuba-
tion period. If the recommendations of the committee were adhered to,
it would take at least nine days to justify the inclusion of an organism
in the colon group. This was found to be very impractical and the ten-
dency has been to simplify the preliminary tests necessary to place an
organism in this category.

The 1917 and the 1920 Standard Methods of Water Analysis define
the colon group as follows: "It is recommended that the Bact. coli group
be considered as including all non-spore-forming bacilli which ferment
lactose with gas formation and grow rerobically on standard solid media."

This characterization is concurred in by Winslow who defines the
colon group as including all aerobic non-spore forming bacilli which produce
acid and gas in glucose and lactose media.

Hauser modifies the definition somewhat by excluding gelatin lique-
fiers.

Other investigators are inclined to extend the colon group to include
spore forming organisms which are capable of fermenting lactose with
gas production and which grow aerobically on solid media. A few such
strains have been recently encountered in the routine examination of water.
Perry and Monfort include such spore forming types as members of the
colon group. The statement of Lohnis and Smith, who have made studies

*In conformity with the recommendations of the committee on nomenclature of the
Society of American Bacteriologists the colon group is considered in the genus Bacterium.

on the life cycle of bacteria, that a single species of Azotobacter may pass
through as many as twelve to fourteen morphological forms including
spores, is not particularly relevant with respect to Bad. coli, the life cycle
of which has been carefully studied by Kellerman and Scales who note
specifically that spores were not observed. It is conceivable that unfavor-
able environment may lead to spore formation by members of the colon-
jerogenes group but the writer has never encountered nor heard of such
a transformation, although he has observed a large number of cultures
kept under various unfavorable conditions, nor does he anticipate such a
fundamental and radical change in form.

These spore forming, aerobic, lactose fermenters confuse the ordin-
ary tests for Bact. coli and must be taken into consideration in interpreting
water analyses, just as it is essential to differentiate the anaerobic spore
forming lactose fermenters which confuse, the presumptive test for Bact.
coli, but there is no logical reason nor justification for placing them in the
colon group.

Clark and Lubs raise the question as to the reliability of lactose fer-
mentation as a primary criterion. They say, "If a fundamental cultural
requirement of the members of the colon-aerogenes family is that it shall
ferment lactose, there is imposed the same sort of requirement for the
characterization of a whole family as is imposed by the MacConkey scheme
when groups within the family are separated on the basis of the fermentation
of another sugar, sucrose."

They point out that the fermentation of sucrose, which was formerly
employed to subdivide the colon group into species and varieties (Mac-
Conkey's scheme), has been found less desirable than differentation on
more recently devised tests, such as the gas ratio, the methyl -red reaction,
and the Voges Proskauer test. They raise the question as to whether it is
not possible that in the near future a test may be discovered which will
supplant lactose fermentation as the salient and fundamental requirement
for the whole colon-aerogenes family. We may well agree with Clark and
Lubs that the future holds out to us promises of improved differential
tests but we do not feel that, in consequence, we shall not utilize such
means as are now available. Surely the classification of the organisms of
this group to the best of our abiltiy on tests which are now known and
used, would simplify a reclassification when these more fundamental, and
we hope more reliable reactions of the future are brought out. The fact
remains that the fermentation of lactose has been successfully employed
for the separation of the non-pathogenic colon group from the disease pro-
ducing para-typhoid and typhoid groups. Until a test is developed which
will adequately supplant this, the fermentation of lactose with acid and
gas production is considered a convenient and reasonably reliable criterion
for members of the colon group of bacteria.

The colon group will therefore be considered to include non-sporing,
Gram negative bacilli which ferment lactose with the production of acid
and gas and which are capable of growing aerobically. The statement

frequently made, that the group consists of aerobes which ferment lactose
is somewhat confusing, for the characteristic fermentation of lactose (gas
formation) is determined under anaerobic conditions.

DIFFERENTIAL TESTS

The colon group, as defined, is a large and complex one including a
number of closely related but distinct species and varieties. The question
arises whether all the members are of equal sanitary significance or whether
some may not be more intimately associated with animal or particularly with
human pollution thereby becoming of special significance in the interpreta-
tion of water or other examinations. It may be of practical value to dis-
tinguish and classify these different types and, if possible, to correlate them
with their habitat. For these purposes a large number of tests have been
employed.

The more important reactions to be considered are:

Coagulation of milk.

Gelatin liquefaction.

Production of indol.

Motility.

Fermentation of carbohydrates.

Production of acetyl-methyl-carbinol . (Voges Proskauer reac-
tion ) .

Uric acid test.

Methyl red test.

Coagulation of milk. The test for the determination of the co-
agulation of milk is made by inoculating litmus or brom cresol purple
milk which is then incubated at the body temperature for 48 hours. Acid
is formed, some gas develops, and coagulation usually takes place in this
time. If the milk has not been clotted the tube is immersed in boiling
water for a few minutes or brought to a boil over a flame and if this treat-
ment induces coagulation the reaction is considered as positive. If the
milk has been over sterilized in the autoclave, coagulation does not take
place as readily as when the medium is sterilized by the intermittent method
in the Arnold. There is no digestion of the curd and the whey, if present,
is clear. The litmus may be reduced by some strains. The milk reaction
has been found very valuable in identification of the colon group.

Gelatin liquefaction. The liquefaction of gelatin is an important
test for the differentiation of colon species. Studies by Gligler and by
Johnson and Levine indicate that this test is well correlated with motility
and fermentation of glycerol. Unfortunately the liquefaction of gelatin is
difficult to determine. The period of incubation usually employed is
fourteen days at 20 degrees C. Gage and Phelps, and later Johnson and
Levine pointed out that the proportion of liquefiers recognized varies with
the period of incubation as mav be seen from Table I.

Among 202 strains studied by Johnson and Levine, 106 (52%) were
gelatin liquefiers after 34 days incubation but only 31 (15.3%) showed
this reaction in 13 days.

8

TABLE I. LIQUEFACTION OF GELATIN BY 202 COLON STRAINS FROM SOIL.

Incubation days

Cultures liquefied

% of total liquefiers

2

17

16.0

7

17

16.0

13

31

29.2

20

38

35.8

27

61

57.5

34

106

100.0

The long incubation period necessary makes this test an inconvenient
one for practical work and in some laboratories a modification has been
introduced employing 37° C. for1 48 hours. This of course liquefies
the gelatin and so to determine whether a physiological decomposition
has taken place the tubes are immersed in ice water until control tubes
solidify. Inoculated tubes which remain liquid are regarded as having
been peptonized by the action of the organism in question.

Indol. The production of indol from peptone is very extensively
determined particularly in England where indol formers are regarded as
"typical" £act. coli. The test is usually carried out in the following man-

ner :

One percent peptone water cultures are incubated for four or five days
at body temperature. The culture is acidified with 1 c. c. of a ten percent
solution of sulphuric acid and then 1 c. c. of a 1-5,000 potassium nitrite
is added so as to form a layer on the surface. After a period varying from
a few minutes to an hour a red ring will develop at the junction of the
nitrite and acidified peptone culture if indol is present. This is known
as the Salkowfsky test.

A more delicate reaction is obtained by the Ehrlich test which is per-
formed thus:

To the culture add 3 c. c. para-dimethyl-amido-benzaldehyde and 3
c. c. of a saturated solution of potassium persulphate. Presence of indol
is indicated by the production of a red coloration.

The significance and value of the test has been much in dispute. Howe
found it to be but slightly correlated with other characteristics and con-
sequently regarded it to be of little diagnostic value. Castellani and
Chalmers on the other hand, consider indol of fundamental importance in
classification of colon-like forms, and Houston believes it to be of par-
ticular significance for distinguishing the "typical" or "excretal" from
the "atypical" Bact. coli. Levine found that among members of the colon
group which were of intestinal origin, 91 percent formed indol, whereas of
those obtained from soil only 37.3 percent were indol positive.

The constancy of the reaction has been questioned. Smirnow reported
that subjection of Bact. coli. to the action of carbolic acid induces a loss
of the ability to produce indol but that this character is regained after
several sub-cultures in nutrient broth.

9
The source of indol in culture media is the amino acid tryptophane,

NH
which is decomposed with the liberation of indol.

NH

Many of the irregularities reported are undoubtedly due to varying quan-
tities of tryptophane in the media employed and these may be eliminated
by use of tryptophane broth as suggested by Kligler.

Motility. Motility may be determined either by the hanging drop
method or by the use of a semi-solid medium such as Hesse agar. In the
latter non-motile organisms grow along the line of inoculation with very
little diffusion into the medium whereas the motile organisms grow
rapidly away from the line of inoculation producing a distinct turbid zone
of several millimeters, in 6 to 12 hours, which may easily be observed
with the naked eye.

There is considerable disagreement as to the value of motility as an
index and differential test for members of the colon group. There are
undoubtedly both motile and non-motile colon bacilli in the intestinal tract
of man. The character seems to be quite variable as a number of prelim-
inary cultures are sometimes required to make motility evident and
McWeeney has reported that some strains were motile at 20 degrees and
not at 37 degrees C.

The statement that motile forms are characteristic of the human in-
testine appears to be in error as Stocklin (quoted by McWeeney) observed
116 non-motile strains among 300 colon bacilli from feces. Levine found
only 32 percent of 25 cultures from man to be motile and only 20 percent
of 30 cultures from raw and septic sewage whereas colon strains obtained
from animals were almost always motile (sheep 77.3%, cow 80.0%, pig
93.7%, and horse 100%). It should be noted that these results were ob-
tained with the use of a semi-solid agar (nutrient agar containing 0.5%
agar).

The relation of motility as determined by the hanging drop and semi-
solid media has recently been studied by Chen and Rettger with results
indicated in Table II.

It appears that for the true Bact. coli there is excellent correlation in
the two methods of motility determination. Out of a total of 173 cul-
tures examined, 119 strains were found to be motile by the hanging drop
method and 121 with the semi-solid agar medium. With the Bact. aerogenes
strains, however, out of 477 cultures observed, 122 were motile by the hang-
ing drop and only 75 by the agar method. It would seem that for the
aerogenes types observation of motility in semi-solid media is undesire-
able.

10

TABLE II. RELATION OF HANGING DROP AND SEMI-SOLID AGAR FOR

DETERMINATION OF MOTILITY.

(After Chen and Rettger, 1920)

Type of
Organism

Hanging drop

Semi-solid agar (Hesse)

Number
Examined

Motile

Non Motile

Motile

Non Motile

Bact. coli

119

54

121

52

173

Bact. aerogenes

122

325

75

372

447

Levine, however, in a study of 151 strains of the aerogenes-cloacae
group, found an excellent correlation between motility, as determined in
semi-solid medium, gelatin liquefaction and starch fermentation. Thus
of 89 motile organisms 81 (91.0%) liquefied gelatin and only 4 (4.5%)
fermented starch; whereas among 62 non-motile organisms only 2 (3.2%)
liquefied gelatin while 61 (98.5%) fermented starch. A test which cor-
relates so well with other characters is probably of differential value, and,
although it is not recommended at present for routine work, it may be of
significance and should be included in investigational studies.

MacConkey recommends that motility be observed in six hour cul-
tures using dark field illumination. Castellani and Chalmers also employ
motility as an important differential criterion.

Fermentation of carbohydrates. The carbohydrates which have
been most commonly employed in the study of the colon group are listed
below:

Monosaccharids. Glucose, levulose, and galactose.

Disaccharids. Lactose, sucrose, and maltose.

Trisaccharid. Raffinose.

Polysaccharids. Starch, inulin, and glycogen.

Alcohols. Glycerol, mannitol, dulcitol, and adonitol.

Glucoside. Salicin.

The media for fermentation tests generally consist of peptone water or
broth containing one percent of the test substance. Incubation is at the
body temperature for 48 hours and a positive reaction is indicated by
gas production. If desired, litmus, brom cresol purple, neutral red, or
the Andrade indicator may be added to the medium to observe acid for-
mation.

Kligler suggests that quantitive acid-production be substituted for
gas-formation as an index of fermentation. He points out that in stand-
ard meat-infusion sugar-freed carbohydrate broth media there is a rather
sharp dividing line between acid-producers and nonacid-producers at 1.5
percent normal acid and that quantitive gas-production is variable and
unreliable. Although quantitive gas-formation as ordinarly determined in
the Smith or Durham tube is markedly inconstant and therefore of little

11

value, the fact that gas is produced at all may, nevertheless be of con-
siderable significance. If a culture is inoculated into sugar broth and
gas is formed, while no gas is produced in plain broth, the organism would
most certainly be regarded as a fermenter of the test sugar irrespective of
whether more or less than 1.5 percent acid is formed.

The low liter might be due to a secondary alkali-production which
masks the acid, as suggested by Rogers. It has been repeatedly observed
by the author that Bact. aerogenes in peptone dipotassium-phosphate so-
lution, containing one percent or two percent glucose, may be acid to
methyl red after 24 hours' incubation but alkaline after a period of 48 to
96 hours at 37 degrees C.

Rogers, Clark, and Evans also determined titratable acid and selected
one percent normal acid as the point of demarcation between fermenters
and non-fermenters but they point out the possible errors in acid-deter-
mination and give precedence to gas-formation, if positive.

The author's observations are that with peptone water as a base and
one percent of the test carbohvdrates, nonfermenters rarely produced as
much as 0.2 percent normal acid.

At what point on the acid scale are fermenters to be differentiated
from nonfermenters? There is considerable disagreement as to the max-
imal amount of acid formed by Bact. coli. Kligler, using meat infusion
media, often obtained titers of four percent normal acid or more and sim-
ilar results have been recorded by Rogers and others. Browne, however,
using Liebig's meat-extract media, states that the limiting acidity for Bact.
coli is 2.4% normal acid as determined by titration with phenolphthalein.
Winslow and Walker determined acid-production in 12 substances by Bact.
coli. The maximal acidity observed was 0.45 c. c. N/20 NaOH to the
cubic centimeter of culture medium, or 2.25 percent normal acid.

The writer's experience, with peptone water as the basic medium, is
in entire accord with Winslow and Walker, and with Browne. Of more
than 2500 titrations, none showed over 2.4 per cent normal acid.

The difference in acid-production observed by various investigators is
due to differences in the composition of the media employed. It is now
well established that more acid is formed in meat-infusion broth than in
beef-extract broth. In media containing much phosphates, as yeast water,
even more acid is formed than in meat infusion broth.

Acid-production should not be given precedence over gas-formation.
They may be independent characters. If, however, after careful studies,
it appears that there is a marked correlation between quantitative acid-pro-
duction and qualitative gas-formation, then it may be feasible to supplement,
if not substitute, the gas test by the acid test. In that event, the line of
demarkation between fermenters and nonfermenters would have
to be determined for the medium employed.

Table III. shows the relation of gas-production to the amount of acid
formed from sucrose, raffinose, ducitol, glycerol, and salicin in peptone
water. Other test substances were observed but are not indicated because
they were invariably fermented with production of gas.

12

TABLE III. RELATIONSHIP BETWEEN QAUNTITATIVE ACID-PRODUCTION
AND GAS-FORMATION BY COLON GROUP.

Test Substance

Gas

Strain

Percentage of normal acid

1
0-0.19 i 0.20-0.39

1

1
| 0.40-0.591 0.60-0.79j0.80ormore

| 1

Sucrose

i:

fNo.
1%
[No.
1%

0

79
98.8

0

!

1

1.2

8
10.6
0

48
64.0
0

19
25.4
0

Raffinose

i:

fNo.

\%
JNo.

[%

1
1.3
72
93.5

0

1
1.3

2
2.5
2
2.6

18
22.8

0

58

73.4
2
2.6

Dulcitol

•

" +

JNo.

1%

JNo.

1%

0

86
97.8

2

3.0

'
1

1.1

5
7.5
1

1.1

23
34.3
0

37
55.2
0

Salicin

+

fNo.
[%
JNo.

1%

0

43

79.7

1

0

1
'
1.8

1

10

•
3

5.6

19
18.6
6
11.1

82
80.4
1
1.8

Glycerol

•

+

fNo.

1%
fNo.

1%

0

4

10.5

0 16
13.6
5 23
13.2 60.5

61
51.7
5
13.2

41
34.7

1
2.6

It will be noted that acid-production in sucrose, dulcitol, and raffinose
is well correlated with the presence or absence of gas. With salicin the
correlation is not so marked, while with glycerol the line of demarcation
between gas-formers and non-gas-formers, as indicated by the quantity of
acid produced, is very indistinct. The substitution of quantitative acid-
production for gas-formation would therefore be particularly undesirable
when working with glycerol.

13

These results are well in accord with those of Winslow and Walker,
who observe: "Gas-formation coincided with acidity except in the case of
dextrin."

All investigators are agreed that members of the colon group normally
ferment the monosaccharids with production of both acid and gas. Very
detailed studies have been carried out on the products of the fermenattion
of glucose, particularly the gas ratio and the H+ion concentration. These
careful observations have yielded very fruitful results.

The disaccharid lactose is of course fermented by all members of
the group and maltose is also attacked. Sucrose is practically always de-
composed by strains obtained from the soil, from grains, or from animal
feces, but less frequently by strains isolated from human dejecta or sewage.
The fermentation of sucrose has been recognized by many investigators as a
convenient and important character for subdivision. It is the primary
character in the MacConkey classification, it is employed by Jackson, and
has been recognized by all the more recent investigators of the colon group
as a most important and convenient differential characteristic.

The trisaccharid raffinose is fermented by practically all strains
which ferment sucrose. This marked correlation between sucrose and raf-
finose fermentation was emphasized by Howe, who observed that dextrose,
lactose, sucrose, and raffinose constitute a metabolic gradient, noting that
fermentation of any of these carbohydrates was always accompanied by
fermentation of the less complex sugar in the series.

TABLE IV. CORRELATION OF SUCROSE AND RAFFINOSE
FERMENTATION IN COLON GROUP.

Acid and gas
in
raffinose

Acid and gas
in sucrose

+

—

+

233

8

—

8

84

In a study of 333 strains obtained from soil, sewage, and feces of
various animals and man, Levine found only 8 sucrose nonfermenters
among 241 strains which attacked raffinose, whereas of 92 raffinose non-
fermenters, 84 failed to ferment sucrose.

Both sucrose and raffinose need hardly be employed simultaneously
in a study of the colon group. The correlation between fermentability of
these carbohydrates has also been observed by Winslow and Walker; Birk;
Rogers, Clark and Davis; Kligler; Murray; Rogers, Clark and Lubs; and
others. The fact that these two sugars are similar in chemical construc-
tion (neither possesses a reacting aldehyde group), may explain the similar-
ity in the behavior of colon bacilli towards them.

The polysaccharids are fermented by relatively few of the1 species
or varieties in the colon group. Ford pointed out that the Bact. aerogenes
was a starch fermenter and that a few strains also fermented inulin. Gly-
cogen is very rarely attacked. Laybourn reports that Bact. aerogenes
usually attacks starches from many different sources.

14

The alcohols have been very frequenlty utilized in investigational
studies. Thus dulcitol was employed by MacConkey and Jackson in their
classifications, but it is now being generally supplanted by other sugars.
Mannitol is fermented by practically all members of the colon group
except a small group observed by Rogers in his grain series. The al-
cohol adonitol was recently suggested as a means of differentiating the
fecal from the non-fecal Bact. aerongenes and this was adopted by the
Committee of Standard Methods of Water Analysis in the 1917 report.

Glycerol has been found by Kligler and later by Levine to correlate
well with gelatin liquefaction and they both distinguish Bact. aerogenes
from Bact. cloacae on the basis of fermentation of this material. The alco-
hols are thus an important group of carbohydrates for studies of the colon
group of bacteria.

The glucoside salicin has been recently suggested to supplant dulcitol
for classification purposes by Kligler and by Levine and Castellani and
Chalmers employ it for primary subdivision of their sucrose negative strains.
Salicin fermentation is an important differential test.

Voges Proskauer reaction. (Acetyl methyl carbinol test). By
the Voges Proskauer reaction is meant the production of an eosin-like col-
oration in dextrose broth cultures by some members of the colon group,
if made strongly alkaline with potassium hydroxide. It takes its name
from the fact that it was first observed by Voges and Proskauer in 1898.
The coloration develops slowly from the surface of the medium gradually
extending throughout the culture. The test is ordinarily carried out by
adding two or three c. c. of 10 percent potassium hydroxide to an equal
volume of a 48 to 96 hour dextrose broth culture and after thoro shaking
the mixture is allowed to stand exposed to the air. The characteristic
eosin-like color will develop in a few hours but it is well to record after
24 hours exposure if negative in a shorter time.

It is suggested that the term Voges Proskauer reaction be restricted
to designate the formation of acetyl methyl carbinol from glucose but when
referring to its production from other carbohydrates or alcohols, the
term acetyl-methyl-carbinol test be applied. The nature of the sub-
stance being tested for is thus indicated just as is the case with indol. The
Voges Proskauer reaction has been found very valuable and many investi-
gators have observed that it is characteristic of the colon-like organisms
of the soil and grains while it is very rare to encounter intestinal members
of the colon group which give this test.

Methyl Red Test. Clark and Lubs in 1915 devised the socalled
methyl red test which serves to split the colon group into a methyl-red-
positive subgroup, found to be characteristic of the organisms obtained
from cow feces and other intestinal sources, and the methyl-red-negative
subgroup, which is the predominating type in the soil and on grains.

The test is made by adding a few drops of methyl red indicator to a
0.5 percent dextrose-peptone- (Witte)-dipotassium phosphate culture and
noting the reaction; a yellow coloration indicates alkalinity or a negative
test and a red coloration denotes acidity or a positive test. The reaction

15

correlates very well with the Voges Proskauer test and will be considered
more in detail in the following pages.

Uric Acid Test. Koser, in studying the utilization of nitrogen from
various sources, observed that some members of the colon group (Bact. aer-
ogenes) can utilize nitrogen from uric acid, whereas others (Bact. coli]
can not. He also showed this characteristic to be correlated with the
methyl red and Voges Proskauer reaction.

Resume. The colon group obviously includes many varieties and
species of bacteria. The monosaccharids and the disaccharid lactose are
always fermented with acid and gas formation; nitrates are reduced and
milk is acidified and coagulated; but with respect to other tests there is
considerable variation within the group.

TABLE V. CHARACTERISTICS OF 333 STRAINS OF COLON GROUP FROM
SOIL, FECES AND SEWAGE.

SOURCE

Soil

Feces and sewage

All strains

Number Studied

177

\
156

333

Character

No.

%

No.

%

No. .

%

pos.

pos.

pos.

pos.

pos. | pos.

Voges Proskauer Test

142

80.3

9

5.8

151

45.4

Motility

123

69.6

96 | 61.5

219

65.8

Gelatin

83

46.8

0

0.0

83

29.9

Indol

66

37.3

142

91.1

208

62.4

Sucrose*

165

93.3

76

48.7

241

72.5

Raffinose*

162 91.6

79

50.7

241

72.5

Dulcitol*

74

41.8

68

43.6

142

42.7

Glycerol*

78

43.1

118

76.2

196

58.8

Salicin*

159

89.9

102

66.1

261

78.4

Dextrin*

82

46.4

8

5.1

90

27.0

Inulin*

21 11.9

0

0.0

21

6.3

Starch*

57 | 32.2

7 '

4.5

64

19.2

'Acid and gas formation observed.

In general the reactions of the organisms isolated from the soil are
quite different from those obtained from various animal feces (horse, sheep,
pig, cow, and man) and sewage. In Table V. and figure 1 are shown the
frequency of the various reactions.

~ -
1 -8 I

S5

i

(0

16

1 £ I

DC

•H E

f t ."> : S

<o & 1 i

43.6

•7fc.7

B

n

43.1

Fig. 1. Percent of Positive Reactions of Coli-like Bacteria from (A) Feces and Sewage,

and (B) Soil

17

II. EVIDENCE OF TWO SUBDIVISIONS IN THE COLON GROUP
AND TESTS FOR THEIR DIFFERENTIATION.

The mass of recent work clearly demonstrates that the colon group
includes two quite distinct subgroups which differ culturally but partic-
ularly with respect to carbohydrate and nitrogen metabolism. These sub-
divisions, which will be referred to hereafter as the coli and aerogenes
sections, are also quite strikingly correlated with habitat, the former pre-
dominating in feces and sewage, the latter in the soil and on grains. The
evidence for this subdivision together with a detailed consideration of the
differential reactions employed is presented here.

General Differential Characteristics. Escherich, in 1885, distin-
guished Bact. coli commune from Bact. (lactis) aerogenes by the greater
plumpness of the latter, its lack of motility, and its more rapid coagulation
of milk. Later Smith, in 1893 and 1895, indicated that the Bact. aerogenes
produced a heavier growth and showed a tendency toward capsule forma-
tion. He also remarked that gas production was more rapid from glucose
and that the proportion of carbon dioxide to hydrogen was greater. Chen
and Rettger observe that in glucose broth the volume of gas is seldom
greater than 40 percent with the coli section, whereas the aerogense sub-
group frequently produced much more gas.

TABLE VI. GAS PRODUCTION IN 1% GLUCOSE BROTH
(After Chen and Rettger, 1920)

Number of strains

Less than 40%

40% or more

No/ | %

No. %

Coli Subgroup
Aerogenes Subgroup

173

447

165 95.4
51 11.4

8 4.6
396 88.6

Burton and Rettger report also that in a modified Uschinsky medium
(glucose substituted for glycerol) the coli subgroup grows very poorly, if
at all, whereas the aerogenes strains show vigorous growth with almost
complete utilization of the sugar.

Temperature Relationships. In some very careful studies on the
rate of multiplication of Bact. coli, Barber found that the rate increased
with increase in temperature showing a maximum at 44 degrees to 45 de-
grees C. A temperature of 40 degrees C. has often been recommended for
the isolation of the colon group from water and in the Eijkman test 46
degrees C. is employed.

With reference to the aerogenes section, we find that Rogers and his
associates often mention the necessity for using a relatively low temperature
(30 degrees C.) for growth of some strains isolated from grains. This
observation is concurred in by Winslow and Cohen, and more recently
Chen and Rettger report that in studies of soil organism a large number

18

refused to grow at 37 degrees C., at least for a while. In some unpub-
lished work the writer has observed that in peptone lactose media at 43
degrees C. (in a water bath) all coli culture studied (16) grew luxuriantly,
as evidenced by strong turbidity in 24 hours, but 69 percent of these strains
showed no gas or only a bubble in this time. Of 20 aerogenes cultures
only 2 showed luxuriant growth, 2 a slight growth, while 16 did not grow
at all.

That these two subgroups should show this temperature relationship
might be anticipated from a consideration of their respective habitats. The
coli section, being most frequently encountered in the animal intestinal
tract would naturally have a higher optimum growth temperature than the
aerogenes section, which is more characteristic of non-fecal origin. It would
be interesting to know whether the members of the aerogenes section, isolated
from the intestinal contents of man can be differentiated from the soil
and grain strains on the basis of temperature relationship.

EVIDENCE FROM CARBOHYDRATE METABOLISM

The metabolism of carbohydrates, particularly glucose, has been care-
fully studied by a number of investigators. The work of Harden and Wai-
pole on the products of fermentation of glucose; that of Keys and Gillespie,
and of Rogers and his associates on the gas ratio; the observations on the
Voges Proskauer test and on acid production particularly the recent studies
on the limiting H+ion concentration have served to clarify the entire group.

The Products of Fermentation of Glucose. Harden has prob-
ably done the most important and extensive work on this problem. Among
the products of glucose fermentation, Harden and Walpole list alcohol,
acetic acid, lactic acid, succinic acid, formic acid, carbon dioxide, and
hydrogen as indicated in the following table:

TABLE VII. PRODUCTS OF DECOMPOSITION OF GLUCOSE
(After Harden and Walpole 1905-06)

Products of fermentation

Per cent by weight of sugar fermented by

Bact. aerogenes

Bact. coli

Alcohol
Acetic acid
Lactic acid
Succinic acid
Formic acid
Carbon dioxide

17.1
5.1
5.5
2.4
1.0
38.0

12.85
18.84
31.90
5.20
0.
18.1

Carbon dioxide c. c.
per gram glucose
Hydrogen c. c.
per gram glucose

C°2/H2
H2/C02

198.3

82.4
2.4

.42

91.8

110.0
.83
1.19

19

A perusal of the table will show that Bact. coli produces a large
amount of acid whereas the acidity produced by Bact. aerogenes is very
much less. This is particularly noticeable with respect to acetic and
lactic acid. It will also be noted that the Bact. aerogenes produces about
twice as much carbon dioxide as Bact. coli and that the volume of hydro-
gen gas formed is more nearly the same. They observed further that Bact.
coli utilized only a part of the available carbohydrate whereas the Bact.
aerogenes strains completely exhausted the sugar. They conclude from
these observations that the two organisms act upon glucose in a totally
different manner and must therefore be regarded as separate and distinct.

The Voges Proskauer Reaction. If the products of glucose de-
composition enumerated above are summed up, it will be found that
only 69 percent of the carbon is accounted for in case of Bact. aerogenes
and 87 percent with Bact. coli. This led Harden to search for the discrep-
ancy which he accounted for by the presence of a crude glycol. This con-
sists for the most part of 2:3 butyleneglycol (CH3CHOH-CHOH-CH3),
On oxidation it yields acetyl-methyl-carbinol (CH3, CHOH, CO. CH3), a
volatile reducing substance, which, when mixed with potassium hydroxide
in the presence of peptone, imparts an eosine-like coloration to the mixture
on standing. Butyleneglycol is oxidized to acetyl-methyl-carbinol by Bact.
aerogenes but not by Bact. coli.

Neither acetyl-methyl-carbinol nor butyleneglycol give the eosin-like
coloration when mixed with potassium hydroxide. In the presence of
peptone, however, the coloration develops on standing in the case of the
carbinol but not with the glycol. According to Harden the reaction is
due to further oxidation of the carbinol (CHSCO.CHOH.CH3) to diacetyl
(CH3CO.CO.CH3) which reacts with some constituent of the peptone. In
a later study Harden and IN orris report that in the presence of strong potas-
sium hydroxide solution diacetyl reacts with proteins to give a pink color-
ation together with a green fluorescence. With arginine, creatine, dicyan-
amide and guanidine acetic acid, the pink coloration is also obtained but
the fluorescence is absent. The reaction depends on the presence of the
group NH:C (NH2) N:HR. The exact significance of R. has not been de-
termined. Harden ascribed the Voges-Proskauer reaction to the production
of acetyl methyl carbinol.

The reaction takes its name from the fact that it was first observed
by Voges and Proskauer in 1898, in their studies on the "Bacteria of Haemor-
rhagic Septicaemia." They describe this observation as follows:

"On addition of caustic potash, we observed a new interesting color
reaction. If the tube be allowed to stand 24 hours and longer at room
temperature, after the addition of the potash, a beautiful fluorescent color
somewhat similar to that of a dilute alcoholic solution of eosin forms in the
culture fluid particularly at the open end of the tube exposed to the air.
We have investigated a few of the properties of this coloring substance,
which is not produced by the action of the alkali on the sugar, and have
found that it is fairly resistant to the action of the external air. After a

20

time however, it becomes paler, and finally gives place to a dirty greenish
brown."

Considerable work has been carried out on the Voges Proskauer test
in the last five years particularly with reference to its constancy, reliability,
and methods of determination.

It was customary to allow the potassium hydroxide-culture-mixture to
stand 24 hours, and some investigators did not record until after
48 hours. This is extremely unfortunate as it results in unnecessary
loss of time. Levine, Weldin, and Johnson found that of 140 strains which
gave the Voges Proskauer reaction from glucose, 130 (92.9%) were posi-
tive after 5 hours. A similar result was observed with sucrose, where of
134 positive carbinol tests, 127 (94% ) were obtained in 5 hours. The
same was true of other substances from which acetyl-methyl-carbinol
was produced as is shown in the accompanying table. They conclude
from this that a period of 5 hours after the addition of the alkali is suf-
ficient as a presumptive test for the Voges Proskauer reaction.

TABLE VIII. COMPARISON OF FIVE HOUR AND TWENTY-FOUR HOUR
RECORDS OF TESTS FOR ACETYL-METHYL-CARBINOL.

Test Substances

Total
positive
reactions

Positive
in 5 Hr.

Positive in 24 Hr.
Negative in 5 Hr.

No. %

No. o/e

Glucose
Sucrose
Raffinose

140
134
114
131
104
14
12
18

130
127
114
116
86
11
12
17

92.9
94.8
100.0
88.5
82.7
78.6
100.0
94.4

10
7
0
15
18
3
0
1

7.1
5.2
0.0
11.5
12.3
21.4
0.0
5.6

Mannitol

Salicin

Dulcitol

Dextrin

Starch :....

West suggested that the reaction could be hastened by heating the
mixture and blowing air through it. Levine, Weldin and Johnson employed
various oxidizing agents (potassium dichromate, potassium perchlorate,
bleaching powder, barium peroxide, and hydrogen peroxide) all of which
were capable of accelerating the reaction but best results were obtained
with hydrogen peroxide. To 3 c. c. of a 48 hour culture was added an
equal volume of 10 percent potassium hydroxide; the mixture was heated in
a boiling water bath for 3 minutes and then 2 or 3 drops of hydrogen
peroxide were added. The pink coloration appeared in one or two min-
utes and persisted for several hours. An excess of hydrogen peroxide or
any other oxidizing agent is to be avoided as the coloration will disappear
in a very few minutes, or even instantly, if the excess is very great.

Chen and Rettger suggest the following technique for the Voges Pros-
kauer test: Five or six c. c. of the culture are added to an equal volume
of 10 percent potassium hydroxide in a test tube, well shaken, and incu-
bated at 30 degrees C. for one to three hours, after which the tube is again
vigorously shaken until the liquid becomes foamy. A decided eosin-like
coloration will develop in an hour or two.

21

The test for acetyl-methyl-carbinol is but little affected by the period
or temperature of incubation of the culture. Positive reactions have been
obtained after one, three, or five days at 30 to 37 degrees C. and Chen and
Rettger obtained positive results in 10 to 14 hours at 30 degrees C. The
kind of peptone does not influence the test, but brighter reactions are ob-
tained in peptone than in synthetic media. As neither the character of
the medium nor the period of incubation of culture interferes seriously
with the test, the Voges Proskauer reaction should serve as a convenient
and valuable index for the differentiation of aerogenes from the more
objectionable coli section.

Gas Production and Gas Ratio. When the presumptive test for the
colon group was first suggested, it was pointed out that a volume of 25 to 75
percent gas was particularly likely to be due to colon bacilli. Escherich
in 1885 determined gas ratios for Bact. aerogenes, but Theobcjd Smith
in 1895 first called attention to the significance of the ratio of the gases
evolved in the decomposition of glucose, pointing out that, whereas Bact.
coli produced twice as much hydrogen as carbon dioxide, the Bact. aerogenes
differed in that it produced equal volumes of these two gases. These ratios
were obtained with the Smith fermentation tube and may be referred to
as the crude gas ratio. The determination of the composition of the gases
in the Smith tube is very inaccurate and unreliable due in part to the ab-
sorption and solution of carbon dioxide and to neutralization by amphoteric
substances in the culture medium which would tend to reduce the amount
of carbon dioxide observed. Referring to Table VII., it will be noted
that in the careful quantitative studies of glucose fermentation by Harden
and Walpole, Bact. coli evolved carbon dioxide and hydrogen in ap-
proximately equal volumes and not in the ratio of one to two as had
been observed by Smith. On the other hand, Bact. aerogenes forms twice
as much carbon dioxide as hydrogen instead of equal volumes observed
with the Smith tube. These differences are easily accounted for, as has
been stated above, by the loss of carbon dioxide in the Smith tube.

Keyes and Gillespie carried out a series of very careful experiments
on the gas ratio and their work has since been confirmed and amplified by
Rogers and his associates. They conclude that the accurately determined
gas ratio, obtained by growing the test organism in glucose medium in
a vacuum and measuring all gas formed, is of fundamental significance
and importance in studies on the colon group.

The real significance of this accurately determined gas ratio was not
fully appreciated until 1914 when Rogers called attention to the striking
correlation between this ratio and the source of the organisms. In
three papers by Rogers, Clark, and Davis (1914) and Rogers, Clark, and
Evans (1914 and 1915), it is demonstrated very conclusively that the
colon strains obtained from bovine feces decompose glucose with the
liberation of carbon dioxide and hydrogen in about equal volume, while
strains isolated from grains formed two 'or more times as much carbon
dioxide as hydrogen. The former group (C02/H2=1:1) seemed in
their other characteristics to resemble Bact. coli; the latter (C02y/H2=2:l)

22

appeared to be the Bact. aerogenes. This work of Rogers and his asso-
ciates aroused considerable interest in the possible different sanitary sig-
nificance of these bacteria and has stimulated considerable investigational
studies.

Acid Production. Studies on acid production have been carried
out from two points of view. The earlier observations were restricted to
qualitative tests or to the determination of total titratable acid with phen-
olphthalein as an indicator but the more recent studies have concerned
themselves with actual or effective acidity (i. e. the H+ion concentration) ;
Evidence as to the differentiations of Bact. coli and Bact. aerogenes by
both of these methods will be considered briefly.

Total or Titratable Acid. In considering total acidity, we are
struck with the fact that the titer is affected by the composition of the
medium as was indicated in the preceding chapter. Observations of
different investigators are therefore extremely difficult to compare as com-
parable data can be obtained only with the same medium. Furthermore
in order to avoid fallacies, it is necessary to observe considerable numbers
of strains and to treat the results statistically. A few extremely high or
low results will influence considerably the average acid production of a
collection of organisms. The use of unqualified averages may therefore
lead to a misconception of the acid producing properties of a group. To
supplement the arithmetic mean or numerical average some statement should
be made as to the distribution of the individual strains (variates) about
the average. This may be indicated by the probable error or by the
standard deviation. The coefficient of variability (the ratio of the stand-
ard deviation to the mean) is an excellent abstract measure of variability.
The modal acid production (the amount of acid most frequently formed)
is usually of greater significance than the average amount of acid formed.

The standard deviation is the measure of variability most commonly
employed, particularly by mathematicians. It may be expressed mathe-
matically as

where "n" is the number of variates or observations, "d" the deviation of
the individual variates from the mean, and "f" the frequency of a devia-
tion "d".The standard deviation serves to indicate whether the departures
from the mean are small or great. The closer the individual organisms
group themselves about the mean, or average, the smaller the standard
deviation.

An example may make clear the meaning and significance of the
standard deviation. Suppose that the amounts of acid formed by a group
(A) of 4 organisms in glucose broth are 2.1, 2.2, 2.2, and 2.3 percent normal
acid, and that those formed by another group (B) of 4 organisms are 1.9,
2, 2.4, and 2.5 percent normal acid. The average for each group is 2.2,
but mere inspection shows that the organisms in Group A and those in

23

Group B are quite differently distributed with respect to this average. In
large collection of data inspection is impracticable but the standard devi-
ation serves well in its place. The standard deviation in Group A is
±0.07 while for Group B it is ±0.25. The larger deviation in B denotes
that the individuals in the group wander farther away from the average than
do those in Group A.

Acid Production from Glucose. Considering 156 strains, isplated
from various sources (including sheep, horse, cow, pig, man, and sewage),
the V. P. negative (coli section) gave an average of 1.82 percent normal
acid with an empirical mode at 1.9 percent while the V. P. positive (aero-
genes section) produced an average of 1.46 percent with a mode at 1.5
percent normal acid. Although the difference (0.36 ±0.04) is not
very large, it appears significant for rather striking differences in
acid formation between the V. P. positive and V. P. negative strains are
observed with many other test substances, as maltose, sucrose, glycerol,
and dulcitol. This observation is also in line with the work of Harden
and Walpole, already referred to that Bact. aerogenes produces less acid than
Bact. coli from glucose.

TABLE IX. ACID PRODUCTION BY COLON GROUP FROM GLUCOSE AND

SUCROSE.

Percentage

Glucose

Sucrose*

of

All

Voges-

All

Voges-

normal

strains

Proskauer

strains

Proskauer

acid

Negative | Positive

Negative | Positive

0.00-0.19

79

79

0.20-0.39

1

1

0.40-0.59

8

8

0.60-0.79

1

1

48

48

0.80-0.99

3

3

6

6

1.00-1.19

2

1

1

1.20-1.39

11

9

2

3

1

2

1.40-1.59

19

15

4

7

3

4

1.60-1.79

22

20

2

2

2

1.80-1.99

54

54

2.00-2.19

41

41

1

1

2.20-2.39

3

3

Total acid-

formers

156

147

9

75

66

9

Mode

1.90

1.90 | 1.50

0.70

0.70

1.50

Mean

1.80

1.82

1.46

0.84

0.74

1.57

Probable error....

-+-.20

±.20 ! ±.12

±.23

±.14

±.16

Standard devi-

ation

-+-.30

±.30

±.18

±.34

±.20

-+-.23

"Only fermenters included in calculations.

Acid Production from Sucrose. Sucrose is not attacked by all
members of the colon group. It is practically always fermented, how-
ever, with acid and gas production by the aerogenes section and many of
the strains of the coli section. In the foregoing table are given the fre-

24

quency of acid production from sucrose by the V. P. negative (coli) and
V. P. positive (aerogenes) strains isolated from feces and sewage. It
will be observed that those V. P. negative forms which are capable of

TABLE X. ACID-PRODUCTION IN FERMENTABLE SUBSTANCES BY VOGES-
PROSKAUER-POSITIVE AND NEGATIVE BACILLI OF COLON GROUP

Test Substance

Percent of Normal Acid

Excess of acid
(in percent of Normal)
by the V.P.+ strains

American-Museum
strains

Levine's
strains

V.P.-

V.P.+

V.P.-

V.P.+

American-
Museum
strains

Levine's
strains

Glucose

1.82
1.31
0.96
1.37
0.66
0.71
0.79
0.58
0.83
1.00

1.52
1.28
0.85
1.41
1.01
1.52
1.22
1.27
1.15
1.38

1.82
1.36
1.31
1.31
0.75
0.74
0.96
0.70
0.81
0.94

1.46
1.21
1.26
1.48
1.12
1.57
1.32
1.28
1.30
1.28

-.30
-.03
-.11
+ .04
+ .35
+ .81
+ .43
+.69
+.32
+.38

-.36
-.15
-.05

+.17
+.37
+ .83
-h.36
+.58
+.49
+.34

Galactose

Lactose

Mannitol

Maltose

Sucrose

Raffinose

Glycerol

Dulcitol

Salicin

flc/ct- Production by Voy&s PrasAauer Pos/fre
k^ Bacteria

25

attacking sucrose produce less acid than the V. P. positives. The means
for the two groups are .74 and 1.57 percent respectively and the empirical
modes 0.7 and 1.5 percent normal acid.

Acid Production from Various Other Carbohydrates. Obser-
vations on the average quantities of acid formed from various substances
by members of the coli (V. P. — ) and aerogenes (V. P. + ) subgroups,
obtained from the American Museum collection and those isolated by the
author from feces and sewage gave the results shown in Table X.

Inspection of Table X. and figure 2 indicates that considering all of
the 167 strains studied, the aerogenes strains form less acid from glucose
than does the coli section and about equal quantities form galactose, manni-
tol, and lactose. In all other test substances, maltose, salicin, raffinose,
dulcitol, glycerol, and sucrose, the V. P. positive strains (aerogenes) give
rise to considerably more acid, the excess increasing in the order named.
Although the differences obtained in salicin, raffinose, and possibly glucose,
may not be so significant on account of the variation observed among in-
dividual strains, it is nevertheless quite striking that a small number of
strains taken at random, from the American Museum collection and a few
freshly isolated cultures should show such a marked parallelism as is in-
dicated in Table X. and figure 2 with respect to the amount of acid formed
when they decompose such a variety of carbohydrates.

The H+ion Concentration. In 1915 Clark and Lubs first pointed
out that in a medium consisting of 0.5 percent anhydrous glucose, dipotas-
sium phosphate, and Witte's peptone, the low ratio cultures (coli section)
produced a high acidity (H+ion) which remained permament; the high
ratio strains (aerogenes section) were much less acid and became progress-
ively more alkaline. This difference in acidity was easily recognized by
the use of the indicator methyl red which gave an acid reaction with the
low ratio and an alkaline reaction with the high ratio group. This test
has become known as the methyl red reaction.

Principle of the Methyl Red Reaction. Michaelis and Morcora*
observed that cultures of Bact. coli fermented lactose until a hydrogen ion
concentration of IXlO5 was reached and then ceased their activity. They
considered this point a physiological constant for the organism in question.
Clark, in 1915, obtained similar results with a culture of Bact. coli in glu-
cose peptone water. He found that, irrespective of the initial acidity, the
final hydrogen ion concentration varied but slightly (PH 4.37 to 4.55).
If we again consider the products of decomposition of glucose by Bact. coli
and Bact. aerogenes (see Table VII.), it will be noted that whereas Bact.
aerogenes decomposes 14 percent of glucose into acids (acetic lactic, suc-
cinic, and formic) Bact. coli produces acids from 56 percent of the glu-
cose. Thus, from a given quantity of sugar, the amount of acid evolved
and consequently the H+ion reached will be greater for Bact. coli than for
Bact. aerogenes. If now the amount of glucose in a medium is restricted
to that quantity necessary to yield the limiting (inhibiting) H+ion con-
centration for Bact. coli, the resulting reaction with Bact. aerogenes will

*Cited by Clark and Lubs

26

necessarily be less acid. On further incubation the Bact. coli would die off,
whereas the more alkaline Bact. aerogenes culture would continue to grow
exhausting the available sugar, after which the reaction would become
progressively more alkaline. This reversion is due partly to the decompo-
sition of the peptones with the liberation of alkali but particularly to the
conversion of the organic acids originally produced from the glucose into
carbonates and bicarbonates, as shown by Avers and Roup. As incuba-
tion is prolonged, the difference in reaction between a Bact. aerogenes cul-
ture and one of Bact. coli may be considerably increased.

Clark and Lubs, in some very careful work, found that a concentra-
tion of 0.5 percent anhydrous glucose, dipotassium phosphate, and Witte's
peptone affords the proper combination of glucose and buffer substances
for the differentiation of the coli and aerogenes sections. They recommend
an incubation temperature of 30 degrees C. for 5 days, after which period,
Bact. coli is acid and E act. aerogenes will be found to be alkaline to the
indicator methyl red.

TABLE XI. EFFECT OF THE TEMPERATURE AND THE PERIOD OF INCUBA-
TION ON THE REACTION WITH METHYL RED.

Reaction

Incubation at 37 C. in days

Incubation at 30 C. in davs

2nd

3rd | 4th

6th

2nd

3rd

5th | 7th

Acid ..

143
12
12

146
9
12

146
9

12

146
9

12

139
18
10

149
6
12

148
7
12

148
6
13

Neutral
Alkaline

As to the effect of temperature and period of incubation on the re-
action with methyl red, the indications are that the final reaction is reached
more quickly at body temperature than at 30 degrees C. (Table XL).
With an incubation period of 5 days there is little choice between 30 de-
grees C. and 37 degrees C., but if the time of incubation is reduced to two
or three days, which would be very desirable for routine water work, the
body temperature seems preferable. It is of course recognized that some
strains of the aerogenes section will not grow well at body temperature,
but from the point of view of the bacterial analysis of water it is felt that
there is little value in detecting such strains. The extra expense of main-
taining a 30 degree incubator for their isolation is consequently not war-
ranted.

Too much emphasis can not be placed on the necessity of employing
Witte's peptone in preparation of the medium mentioned above. The in-
discriminate substitution of other peptones is bound to lead to confusion
and error. Owing to the great difficulty in obtaining Witte's peptone,
Clark and Lubs were led to devise a synthetic medium which consists of
the following:

0.7% anhydrous Na2HP04.

0.2% acid potassium phthalate.

0.1% aspartic acid.

0.4% anhvdrous dextrose.

27

Incubation at 30 degrees C. for three to five days is recommended.
This synthetic medium may also be used for the Voges Proskauer test by
mixing with it (at the time of addition of the alkali) 0.1 percent pure
casein.

Evidence from Nitrogen Metabolism. In the foregoing paragraphs
it was pointed out that there were two main divisions in the colon group
which could be easily distinguished by differences in carbohydrate meta-
bolism. It is becoming apparent that the differentiation may also be
made upon nitrogen metabolism.

In 1903, Rettger noted that the products of protein metabolism (mer-
captan, scatol, phenols aromatic axyacids, etc.) are formed much more
slowly by Bact. aerogenes than by Bact. coli.

Koser (1918) in a study of 124 colon cultures found that in a medium
containing uric acid as the sole source of nitrogen the V. P.-)- strains
(aerogenes section) grew luxuriantly whereas the V. P. — strains (coli
section) failed to grow.

TABLE XII. DIFFERENTIATION OF COLI AND AEROGENES SECTIONS

(After Koser, 1918)

Growth in
Uric Acid medium

Voges Proskauer
reaction

Methyl Red
Test

Cultuies employed

Pos. | Neg.

Pos. | Neg.

Pos. | Neg.

Coli subgroup 74

0 | 74

0 | 74

72 | 2

Aerogenes subgroup 50

50 | 0

50 | 0

0 | 50

5 grams.
0.2 grams.
0.1 grams.
1.0 grams.
30.0 grams.
0.5 grams.

The medium employed by Koser consisted of the following:
Distilled ammonia free water 1000 c. c.
NaCl

Magnesium sulphate
CaCl2

Dipotassium phosphate
Clycerol
Uric acid

Sterilization is in the autoclave at 13 to 15 pounds for 15 minutes.
A slight turbidity, presumably due to the finely divided precipitate of cal-
cium sulphate may be evident immediately after autoclaving but disap-
pears on cooling.

A solid uric acid medium may be prepared by the addition of 1.5
percent washed agar.

He suggests the differentiation is due to the inability of the strains in
the Coli Section to break down the purin ring and thus obtain nitrogen
for growth from the uric acid.

NH—C '-O

I I

o~c c

I

NH

*
— C—NH/

28

Evidently scrupulous care must be employed in the preparation of
the medium to avoid introduction of any nitrogenous compounds which
might be present in imperfectly cleaned glass-ware or ordinary tap water
or even distilled water, for the nitrogen thus introduced might serve to pro-
mote growth of organisms of the coli subgroup and thus obscure the
differential reaction.

The reliability of this uric acid differential test has been confirmed by
Chen and Rettger in a study of 640 strains isolated from soil and from
intestinal tracts of man and various animals. They note, however, that
half the coli strains isolated from the soil grew in the uric acid medium.

TABLE XIII. GROWTH OF THE COLON GROUP IN URIC ACID MEDIUM
(After Chen and Retteer, 1920)

Growth

Pos.

Neg.

Total

Aerogenes strains from soil (V. P. -(-)

447

0

447

Coli strains from feces (V. P. — )

0

173

173

Coli strains from soil (V. P. — )

10

10

20

They showed further that xanthine may be substituted for uric acid and
that all the coli strains from the soil failed to grow in the xanthine medium.
It is probable that xanthine will be found preferable to uric acid for dif-
ferentiation of the coli from the aerogenes types but there may be an
intermediate group resembling the former with respect to the V. P. and
methyl red reactions but which differs from it in its ability to attack uric
acid.

The temperature of incubation employed by Koser was 37 degrees
C. for four days. Chen and Rettger employed a temperature of 30 degrees
C. for three to five days.

Differentiation on Solid Media. The coli and aerogenes sections
also present several cultural differences particularly with respect to the
appearance of colonies on some solid media. The aerogenes colonies are
generally larger more opaque and more convex than those of the coli sub-
group. On litmus lactose agar the former sometimes revert to an alkaline
reaction. These differences, however, are very difficult to detect on litmus
lactose agar but may be readily observed in the Endo and particularly on
eosin-methylene-blue agar.

Ferreira, Horta and Paredes noted that Bact. aerogenes and Bact. clo-
acae produced a rose color on Endo agar but that the metallic luster, so
characteristic of Bact. coli, was absent. Levine employing a simplified
fuchsin sulphite (Endo) agar notes striking differences between the col-
onies of the coli and aerogenes subgroups. The coli strains formed deep red
button-like colonies with a greenish metallic sheen (by reflected light), and
were usually three or four m. m. in diameter when well isolated. The
aerogenes colonies are lighter colored, markedly convex, and do not show
the metallic luster.

29

Wood records that on neutral-red-bile salt lactose agar, the V. P.-|-M.
R. — strains may develop mucoid colonies which are generally paler than
the V. P.— M. R.-f- strains.

Another distinct differentiation may be obtained with the modification
of the (Holt-Harris) eosin-methylene-blue agar (see appendix). On this
medium the well isolated coli colonies are about three m. m. in diameter
appearing very dark, almost black by transmitted light and by reflected
light they seem to be button-like, often concentrically ringed with a dis-
tinct greenish metallic sheen. Those of aerogenes, on the other hand, are
larger, tend to run together, are markedly convex, very much lighter in
color, (when viewed by transmitted light), and a metallic sheen is rarely
observed. That these cultural differences are quite reliable, at least for
routine work, is indicated by a report of Levine, who found that 96.9 per-
cent of 122 colonies picked as of the coli section from their appearance
on eosin methylene blue agar and 82.4 percent of 102 colonies supposedly
of the aerogenes section were confirmed by subsequent tests.

Correlation of Reactions. The studies of carbohydrate and nitrogen
metabolism and cultural characters, all indicate that the colon group em-
braces two distinct sections which may be readily distinguished by a num-
ber of tests including (1) the gas ratio, (2) acidity to methyl red, (3) the
Voges Proskauer reaction, (4) growth in uric acid medium, and (5) ap-
pearance of colonies on agar (Endo and eosin-methylene-blue).

Although no investigator has employed all these reactions simultan-
eously, they are known to be very strikingly correlated. Thus Clark and
Lubs in 1915 observed a perfect correlation between the gas ratio and the
methyl red test. Levine pointed out that the strains which were positive
for methyl red were negative for the V. P. test and vice versa. The cor-
relation between these two reactions had been extensively confirmed.
Johnson; Burton and Rettger; Chen and Rettger; Hulton; Greenfield; and
many others have observed an almost perfect correlation. Clark and
Lubs (1917) correlated the gas ratio w^ith the V. P. and M. R. reactions,
while Koser and Chen and Rettger have shown that the uric acid test,
the V. P. and methyl red reaction were strikingly confirmatory of each
other.

Choice of a Routine Differential Test. The question naturally
arises as to which test shall be employed or at least given preference in
future studies of this group. A choice of a test must of course be de-
pendent upon the nature of the work at hand. For investigational studies
it should be emphasized that the carefully determined gas ratio, as is urged
by Rogers, Clark, etc., is of fundamental importance and that in order
to throw further light upon the reliability of the other reactions men-
tioned it is desirable that they should all be considered and observed. The
problem, however, is quite different when applied to routine water an-
alysis. There the time available is limited, the apparatus and other lab-
oratory facilities are at a minimum, and the skill of the analyst, we must
regrettably admit, is too frequently not comparable with that of the chem-
ist.

30

For routine water work the selection of a differential test must be
governed to a great extent by the following considerations:

1. The medium should be simple, easy to prepare, and permit of a
considerable degree of variation.

2. The reaction should be distinct and constant.

3. The test must be one which can be completed in a short time,
preferably not more than 24 hours.

In the light of these considerations it becomes evident that the gas
ratio is out of the question. That it is permanent and constant has been
well demonstrated hy the investigations of Rogers and his associates, Hut
the skill and apparatus necessary together with the time required removes
it from the possibility of a routine water test.

The uric acid reaction is a convenient and reliable test but the scrup-
ulous care required in preparation of the medium can not be obtained at
the present time at least, in laboratories which are concerned with routine
water analysis. The period of incubation, three to five days, is also a
disadvantage.

The choice thus becomes limited to the Voges-Proskauer and methyl
red tests. Opinions are quite at variance as to which is preferable. The
methyl red reaction has been urged by Clark and Lubs and by Winslow.
The necessity for employing Witte's peptone together with the fact that the
reaction is based on a delicate adjustment of the source of carbon and
buffer substance and the insatiable desire among bacterialogists to deviate
from the media recommended and to make individual substitutions has
lead to numerous difficulties in the application of this test in practice.
The synthetic medium of Clark would of course eliminate some of these
difficulties. The methyl red reaction is simple, reliable, and when care-
fully performed, constant, but the period of incubation for accurate dif-
ferentiation is quite long (3 to 5 days), too long, in fact, to be conven-
iently employed as a routine test.

The tendency recently has been toward the V. P. reaction. Thus Chen
and Rettger conclude from a study of 640 strains that the V. P. reaction
is even more satisfactory than the methyl red test in that it is simple in
operation, and when correctly carried out, is thoroughly constant in its
results.

Clark and Lubs, on the other hand, point out that the production of
acetyl methyl carbonol, being the result of secondary reaction, and pos-
sibly synthetic to some extent, may not be intimately connected with the
main course of the fermentation and the quantity produced may be very
slight, thus giving a faint V. P. test even though the fermentation may be
very vigorous. Nevertheless, the reaction has been found to be very con-
stant and it has proven very satisfactory in the hands of practically all
who have tried it. The advantages of this reaction are:

1. Any peptone medium in which the organisms will grow and
which contains glucose (in a wide range of concentration) is suitable.
It is preferable, however, to have the medium as free from color as pos-
sible.

31

2. The reaction may be obtained after 14 to 24 hours incubation at
30 degrees or 37 degrees C.

3. The brand of peptone employed does not affect the intensity of
the reaction.

It is interesting to note that in case of a mixture of coli and aerogenes
types, the methyl red test will be acid and the V. P. reaction will be
positive. Undoubtedly many instances are recorded as examples of inter-
mediate strains when they were really due to impure cultures. This was
the experience of Johnson and Levine who found that by repeated puri-
fication, the proportion of strains which would not show perfect correlation
was considerably reduced. Chen and Rettger record that of 18 strains
which persisted in giving methyl red positive and V. P. positive reactions
in all media, which they employed, only four continued to give non-corre-
lating reactions after purification, although contaminating organisms could
not be demonstrated.

Resume. The gas ratio, Voges Proskauer, methyl red, and uric
acid tests are strikingly correlated. The members of the colon group
which produce acetyl methyl carbinol, are capable of using the nitrogen
from the purin ring of uric acid, give an alkaline reaction with the methyl
red test, and in the decomposition of glucose, yield a relatively small
quantity of acid and two or more times as much C02 as H2. On the
other hand, the organisms, which do not produce- acetyl methyl carbinol,
can not utilize the nitrogen from the purin ring, give an acid reaction with
methyl red, break down glucose with the production of a relatively large
amount of acid and liberate C02 and H2 in approximately equal volumes.
The colon group therefore includes two distinct subdivisions which are
characteristically of different sources. These have been designated the
coli and aerogenes sections. Their characteristics are tabulated below:

TABLE XIV. DIFFERENTIATION OF THE MAIN SUBDIVISIONS OF THE

COLON GROUP.

Section

Gas Ratio
C02/H2

M. R.

Test

V. P.

Test

Growth
Uric Acid
medium

Habitat

Coli

1.0
(low ratio)

acid

neg.

Negative
(no growth)

Predominates
in feces and sewage

Aerogenes

1.5 or more
(high ratio)

alk.

pos.

Positive
(good growth)

Predominates
in soil and on grains

32

III. CLASSIFICATIONS OF THE COLON GROUP
OF BACTERIA.

Several attempts have been made to classify the numerous organisms
of the colon group for the most part on the basis of acid and gas produc-
tion from various carbohydrates. The reliability of such studies is some-
times questioned on the ground that fermentation reactions are inconstant
and may be easily acquired -or lost.

Smirnow grew colon strains in media containing various chemicals
(3.0% glucose, 4.0% sodium chloride, 0.5% sodium sulphate, and 0.25 to
0.75% phenol) and found that after successive transfers, for periods of one
to three months, indol formation and later fermentation of carbohydrates
were suppressed. Reversion to the original characteristics took place rapidly,
however, when grown on ordinary media.

Bronfenbrenner and Davis found colon organisms in food which fer-
mented lactose slowly when first isolated but after cultivation on lactose
media the rate of decomposition of this substance became normal.

Twort, Penfold, and others, observed spontaneous mutations such as
the loss or acquistion of the ability to ferment various carbohydrates.

After a careful survey of the data on biological variations and muta-
tions, Winslow concludes that "Taking the great mass of colon typhoid
strains, as they are isolated from the bodies or intestines of man and
animals, and cultivated under standard conditions, fermentative character-
istics exhibit a high degree of constancy and what is even more important
a higher degree of correlation with other biochemical and serological and
pathogenic properties."

From the practical water analysis view point, we are especially in-
terested in the effect of a long sojourn in water or soil on the biological
activities of Bact. coli.

Houston repeatedly found that the proportion of strains not forming
indol was much greater among those isolated from purified than from
raw waters. He thought this was due to a loss of indol producing power
as a result of unfavorable conditions encountered in water. He speaks
of such indol negative forms or strains, which differ from the supposedly
original Bact. coli commune, as "atypical" Bact. coli.

Horrocks in 1903 exposed Bact. coli in various types of soils and
waters for two to three months, and Savage in 1905 observed the effect
of tidal mud. They came to the conclusion that alterations in character-
istics were not induced and that there was no evidence that Bact. coli ever
becomes "atypical."

MacConkey placed a broth culture of Bact. coli in a sterile Pasteur
candle which was then suspended in tap water. The water was changed
occasionally. Fact, coli was isolated from the candle at irregular inter-
vals up to 358 days and examined biologically and biochemically. In
no instance was a loss or gain of a character detected.

It seems that the remarkable facts are not that an organism may oc-
casionlly show a biological variation, but that, considering the simplicity
of the bacterial cell, such variations are so infrequent.

33

Theobald Smith, Kligler, Winslow and others have reported that with
cultures of Bact. cloacae motility and fermentation of carbohydrates per-
sisted after the power to liquefy gelatin had disappeared. Digestion of
gelatin is generally considered a reliable differential test. We therefore
need have no compunction about utilizing the apparently more persistent
fermentation reactions in studies on classification of the colon group.

Theobald Smith, 1893, suggested that sucrose fermentation may be
employed to subdivide the colon group into two subgroups and in 1901
Durham suggested the name Bact. coli communior for the sucrose positive
variety and noted that the fermentation of starch was distinctive of Bact.
lactis aerogenes.

MacConkey's Classification. In 1905 MacConkey divided lactose
fermenting (acid and gas) bacilli into four groups on the basis of fer-
mentation of sucrose and dulcitol.

TABLE XV.

MacConkey
group

Acid and gas

Type species

Sucrose

Dulcitol

I

—

Bact. acidi lactici

II

—

+

Bact. coli (commune}

III

+

+

Bact. neapolitanum

IV

+

—

Bact. (lactis} aerogenes

In 1909 he recognized in each group a number of varieties which were
distinguished on the basis of their reactions to the Voges Proskauer test,
motility, indol production, gelatin liquefaction, fermentation of inulin
adonitol, and inosite. The probable existence of 128 different strains or
32 varieties in each of his four subdivisions was suggested. Many of these
have been isolated, described, and given specific names. Others have been
merely indicated by a number in his classification as shown in Table XVI.

Bergey and Deehan's Varieties. Very similar to MacConkey's
classification is that of Bergey and Deehan (1908). They employed eight
characters — fermentation of sucrose, dulcitol, adonitol, and inulin; gelatin
liquefaction, indol production, motility, and the Vogese Proskauer reaction —
and from a consideration of all possible combinations recognized the pos-
sible existence of 256 varieties.

The Jackson Classification. In 1911, Jackson proposed a classi-
fication resembling that of MacConkey but preference is given to dulcitol
over sucrose for the primary division. Each of the four groups thus
formed, which are regarded as species, are then further divided on raf-
finose and mannitol into four varieties designated (A, B. C, and D), and
further differentiation may then be made on motility, indol, gelatin lique-
faction, fermentation of other carbohydrates, etc., giving subvarieties, in-
dicated by numerical suffixes (A1? A2, B2, etc.). This scheme which was
included in the Standard Method for a Water Analysis for 1912 is detailed
in Table XVII.

34

TABLE XVI. MACCONKEY'S CLASSIFICATION OF LACTOSE FERMENTING

BACILLI.

6

g

MacConkey
Group

£

Variety (Name)

o
5-8

1-

-

'S

J3 §.

o

'o
-a

o

'a

0

PN

03

03

<

c

I-H

^

^

1

_j_

_^-

-j.

—

—

2

B. acidi lacti (Huppe)

—

—

+

-f.

—

—

—

Group I.

3

B. levans

-f-

-}-

—

-\-

—

-f-

(Nos. 1-32)

4

B. Grunthal, B. sulcatus gasoformans,

—

-f-

-j-

—

—

—

—

Sucrose—

B. castellus

Dulcilol—

5

B. vesiculosus

—

_j_

—

—

—

6

—

_j_

—

-(-

7

_j_

—

—

—

8

—

33

4.

^_

-J-

—

—

Group II.

34

B. coli communis

—

4.

^-

—

—

—

(Nos. 33-64)

B. cavicida

Sucrose —

35

B. Schafferi*

—

—

-j-

—

—

—

—

Dulcitol-f-

36

-f

+

—

—

—

—

—

65

B. oxytocus perniciosus

_l_

•+

+

+

+

_|_

66

—

—

^_

_j_

—

—

67

_J_

_i_

_j_

Group III.

68

B. rhinoscleroma, B. Friedlander

—

—

—

-j-

_

-f

—

Nos. 65-96)

69

_}_

4.

—

—

-\-

—

-|-

Sucrose +

70

_j-

_j_

—

—

-J-

—

—

Dulcitol-t-

71
72

B. neapolitanus

—

±

+

—

—

I

73

-j-

4.

—

—

—

4-

74

—

4.

—

-

—

75

—

—

—

-

—

+

+

97

_l_

—

_|_

+

.4.

—

4.

98

_

+

_!_

_l_

_j_

99

—

+

_|-

-f.

—

100

—

-|-

-f

+

—

—

101

—

—

-j-

+

—

-}-

—

Group IV.

102

-f

4-

—

+

—

+

(Nos. 97-128)

103

B. lactis aerogenes,

—

—

—

-j-

—

-f

+

B. dysenteriae vitulorum,

Sucrose+

104

B. capsulatus (Pfeiffer)
B. gasoformans non-liquefaciens*

—

—

—

+

—

+

±

Dulcitol—

105

_l_

4.

—

—

-1-

—

+

106

—

-|-

-f

—

—

—

—

107

B. coscoroba

—

—

-j-

—

—

—

—

108

B. cloacae

_j_

-|-

—

—

—

-i-

-\-

1

[109

—

4-

—

-

-

±

—

*Usually produce only a small amount of gas.

35

++ 1 +

;. •

1 'fi

i §

co ~-5

PH
+ + + 1 1

o .^

d

cd -H "

— '.;;;

r | a

£

CD ttj

rH CM

i

§§§§§

CD

CU

'3

p

1-jjH

i i i _i_ i _i_ i

0

Q

T"~r i ~r~r~r 1

PH

M-H

0

+ ^

PH"

25

O K

H — h+++ 1 1

0

0 h*

_]

L S o ^

d

O

CJ

r_

ja I

8 e

eflS 5 5 S S 2

i

0

CD ft5

i— 1 (N CO r-H (N

^r*

^^* *

§

<J<3<Jpq<lcjQ

H + +

^ CO CD

a s s ^

1 IJ

+ 1 + 1

CJ

i i

PH" " "
+ + 1 1

S5

— c

^^

rt B -S

cd ** ** ^*

S

" -8 1

s"

<

1 cq

w

^^^—,

£

-)-

^.?3^£S

H

OJ

w. - i

^3

d

§

1

++I + 1

>

•

x

03 2 i S 5

3

+ -2

+ + + 1 1

< ;;,;.,.

1 i"

^ » , »

i „•

es ,

<PHpHCJQ

j

i

8

'S

1

36

Castellani and Chalmers' Grouping. These authors have recently
suggested a rather unique classification of bacteria in which the organisms
generally considered as members of the colon group are distributed among
three Tribes and three Genra as follows:

Tribe Encapsulateae. (Castellani and Chalmers, 1918. Bacillaceae
growing well on ordinary media, without endospores; neither fluorescent
nor chromdgenic; aerobes; facultative anaerobes; not liquefying gelatin;
possessing capsules in animal tissue.

The genus Encapsulates includes two lactose fermenting (acid and gas)
species differentiated on fermentation of inosite.

Inosite not fermented. 1. Encapsulates acidi lactici

Inosite fermented with acid and gas. 2. Encapsculatus lactis aerogenes

Tribe Proteae. (Castellani and Chalmers, 1918) Bacillaceae grow-
ing well on ordinary media; without endospores; aerobes; neither fluorescent
nor chromogenic; aerobes; facultative anaerobes; not 1 iquefying gelatin,

They recognize two lactose fermenting species in the genus Cloaca
which are different on sucrose fermentation.

Sucrose fermented with acid and gas. 1. Cloaca cloacae.

Sucrose not fermented 2. Cloaca levcns.

All other members of the colon group are placed in their Genus Escher-
ichia of the Tribe Ebertheae.

Tribe Ebertheae. (Castellani and Chalmers, 1918). Bacillaceae
growing well on ordinary laboratory media; not forming endospores; aer-
obes, and often facultative anaerobes; without fluorescence pigment forma-
tion or gelatin liquefaction; without polar staining; Gram negative; with-
out a capsule.

Genus Escherichia. (Castellani and Chalmers, 1918.) Ebertheae
which ferment glucose and lactose completely with acid and gas; milk
clotted.
I. Indol positive division. (Smith)

A. Sucrose fermented — Communior section of Durham.

1. Dulcitol fermented with acid and gas.

a. Adonitol fermented with acid and gas.

(1) Motile 1. Escherichia oxytocus

(2) Non motile. 2. Escherichia metacoli

b. Adonitol no change.

(1) Motile.

(a) Agglutinated by pseudocoli serum

3. Escherichia pseudo-coli

(b) Not agglutinated (late fermentation of sucrose)

4. Escherichia coliformis

(2) Non motile. 5. Escherichia neapolitanus

2. Dulcitol not fermented,
a. Motile

(1) Inosite fermented with acid and gas.

6. Escherichia pseudocoloidella

(2) Inosite no change 7. Escherichia pseudocoloides

37

b. Non motile. 8. Escherichia pseudocoscoroba

B. Sucrose not fermented — Communis section of Durham.

1. Dulcitol fermented with acid and gas.

a. Salicin fermented with acid production.

9. Escherichia cavicida

b. Salicin fermented with acid and gas.

(1) Motile. 10. Escherichia coli

(2) Non motile.

(a) Inosite not fermented.

11. Escherichia coloidella

(b) Inosite fermented with acid and gas.

12. Escherichia coloides

q. Salicin no change. 13. Escherichia metacoloides

2. Dulcitol — no change.

a. Motile.

(1) Maltose — acid and gas.

14. Escherichia paragrunthali

(2) Maltose — no change.

15. Escherichia griinthali

b. Non motile.

(1) Maltose — acid and gas.

16. Escherichia colitropicalis

(2) Maltose — no change.

17. Escherichia vesiculosus
II. Indol negative division (Smith). 18. Escherichia coli mutabilis

The prominence given to the indol reaction and the fermentation of
dulcitol, adonitol, and maltose by Castellani and Chalmers is quite at var-
iance with the practices of American investigators. From the extensive
recent studies on carbohydrate and nitrogen metabolism it appears that
the acidi-lactici types are more closely affiliated with the strains in their
Genus Escherichia, than with the lactis-aerogenes forms and that the Genus
Cloaca of Castellani and Chalmers should be associated with the lactis-
aerogenes forms rather than with the proteus group.

A very serious objection to such classifications as those of MacConkey,
Bergey and Deehan, and Jackson is their extreme flexibility and complexity;
for, as the number of fermentable substances or other characters observed
increases, the number of "varieties" increases geometrically (approaching
infinity) and soon produces a most unwieldy scheme. The number of
"varieties" is given by the formula 2n where "n" is the number of char-
acters studied. Thus with eight characters there are 256 possible combi-
nations or "varieties." This number rises to 1,024 with 10 characters and
to 65,536 when 16 tests are considered. It is aparent therefore that to
regard each character as of similar and equal differential value will quick-
ly result in an unwieldy grouping.

Another objection to these classifications is the arbitrary manner of
selecting the order in which reactions are to be employed for division.
Organisms which are very closely related may be far separated in two

38

schemes of classification if the authors happen to select different characters
for the initial subdivision.

Statistical Classification of the Colon Group. To offset these
objections, Winslow suggested the utilization of the statistical method first
employed by Andrews and Horder and Winslow and Winslow in studies on
Coccaceae. Individual characters are not considered paramount and inde-
pendently but only in relation to each other.

Howe first attempted the statistical method in 1912. This investigator
made a detailed study of acid and gas production and various other tests
on 630 strains freshly isolated from human intestinal contents. He con-
cluded that indol, nitrate reduction, motility, fermentation of dulcitol and
mannitol, and starch were not correlated with other characters and were
consequently not of classificatory value. He recognized only two groups
the sucrose positive, Bad. communior, and the sucrose negative, Bact. com-
munis.

Rogers and his associates in 1914 and 1916 studied a large number of
colon strains from milk,, grains and bovine feces and on the basis of ac-
curately determined gas ratio from dextrose concluded that two distinct
groups may be distinguished. One, referred to as the low ratio group,
produced carbon dioxide and hydrogen in equal volumes and includes
about 52 percent of the strains from milk, 99.6 percent of the strains from
bovine feces, and only 4.8 percent of their grain strains; whereas the high
ratio group, which was characterized by the production of two or more
times as much carbon dioxide as hydrogen, included 47.5 percent of the
milk strains, only 0.5 percent of the bovine fecal strains, and about 95
percent of the strains obtained from grains. There is thus a very strong
correlation between these subdivisions and the source.

Kligler (1915) suggested that salicin be substituted for dulcitol, in
subdividing coli-like bacteria, pointing out that salicin fermentation cor-
relates better with the Voges-Proskauer reaction than does dulcitol de-
composition. He recognizes a sucrose negative-salicin negative group (B.
acidi-lactici) ; sucrose negative-salicin positive group (B. communis) ; su-
crose positive-salicin negative group (B. communior) and sucrose positive-
salicin positive (B. aerogenes) . B. cloacae is differentiated from B. aero-
genes by its inability to ferment glycerol.

The characterization of B. communior as salicin negative is probably
untenable. The term B. coli- communior was first employed by Durham
to describe members of the colon group which fermented sucrose and which
were motile. Later Ford recognized it as a species B. communior.

Of 77 motile sucrose fermenting bacilli of the coli section, 56 (73%)
were found by the writer to be salicin fermenters. It is felt therefore that
Bact. communior should not be described as a salicin non-fermenter.

Where the principle of correlation has been employed the best cor-
related character has apparently been picked out by inspection of the data.
Inspection is a tedious and difficult procedure, entirely inapplicable where
the number of characters considered is large, and it does not permit of a
concise statement of the degree of correlation which exists between differ-

39

ent reactions. The author feels that considerable information in an ab-
stract, concise, and workable form may be obtained from a study of the
coefficients of correlation.

The Coefficient of Correlation. Where we are concerned merely
with the presence or absence of characters the coefficient of correlation be-
tween any two characters may easily be determined. Suppose that it is
desired to know if the characters X and Y are correlated and that a study
of a number of organisms showed that 'a' cultures are positive for both
X and Y; 'b' organisms positive for X but negative for Y; V cultures are
negative for X and positive for Y; and 'd' strains are negative for both X
and Y. The distribution of the organisms is first tabulated thus:

Y

The degree of association, or the coefficient of correlation, is then
expressed, according to Yule, by the formula

' ad-bc (1)

ad+bc
If 'ad' is equal to 'be' the coefficient becomes

or 0; which

ad _|_ be

indicates that there is no correlation whatever. If either 'b' or 'c' is zero

ad

the formula becomes~j~-=l; indicating a perfect positive correlation. If

—be.

ca' or 'd' is zero then we have ^~~= — 1; showing a perfect negative cor-
relation. It should be observed that an absolute positive correlation exists
in reality only if both 'b' and 'c' are zero and an absolute negative cor-
relation when both 'a' and 'd' are zero. In order to avoid coefficients of
1 or — 1 where only one group — 'a', 'b', 'c', or 'd' — is zero. Yule gives
the formula

a (a + b + c + d) - (a + c) (a + b) (2)

\/ (a + c) (b + d) (a + b) (c + d)

In practice, however, a few strains are almost always found in each
of the four groups and Yule suggests the use of the simpler formula (1).
Some caution should therefore be employed in interpreting coefficients
of 1 or — 1.

It was assumed that if the coefficient between two characters is numeric-
ally greater than 0.5 they may be regarded as correlated, but if less than

•/-

—

+ 7

43

- /32

2

39

-?-

—

+ &7

SO

- 76

9

/^es/ft

Corre/'

40

0.3 there is probably no association. A few examples of correlation co-
efficients actually obtained in the course of this study are given to illus-
trate the method of calculation.

The principle of correlation should not be applied indiscriminately
to collections of data for systematic purposes. Certain characters and
properties have been universally accepted as reliable and appropriate for
bacterial differentiation; thus, staining reactions such as the Gram and
acid fast stains, spore formation, and aerobiosis and anaerobiosis, hardly
need to be bolstered up by correlation with other characters to justify their
taxonomic value. On the other hand the significance of such characters
as motility, indol production, and fermentation of certain substances, is
still debatable.

Motility is regarded by many as a highly variable property. Perhaps
it is in reality a reliable morphological difference. Certainly if it could
be shown that this character goes hand in hand with several others, more
reliance and attention should and would be given to motility. The same
is true of the indol test. In dealing with gas formation from carbohydrates,
alcohols, or polysaccharids, the question naturally arises as to which
substance should be given preference for subdivision, or whether all are
to be considered of equal taxonomic value. The lack of a criterion for
determining the most significant fermentable substances has led to con-
siderable confusion. It has already been pointed out how subdivision
on every character studied results in an infinite number of varieties. Where
we are dealing with a number of characters each of which is assumed to
be of equal taxonomic significance, it would certainly be desirable and
advantageous to subdivide on that character which gives the greatest amount
of information as to the manner in which the resulting subgroups react
with respect to other characters. It is under such circumstances that the
principle of correlation of characters may be legitimately, conveniently,
and advantageously employed.

The Method of Selecting the Best Correlating Character. The
following example will illustrate the method of selecting the best correlated
character for the purpose of subdivision of a group of organisms. Let
us take for instance a group of 89 strains of the coli section, which were
found to be non-fermenters of sucrose, and which it is required to further
divide on one of the following characters: — motility, indol production,
dulcitol, glycerol, or salicin fermentation. Tabulation is first made, as
indicated below, so as to show the relation of each character to every
other character and also to facilitate the calculation of correlation coefficient
which are then determined for each pair of characters and recorded as in-
dicated in Table XVIII.

For subdivision that character is selected which gives the highest co-
efficient of correlation with the greatest number of other characters. Thus,
in the group under consideration, motility is not well correlated with any
other character. Dulcitol and glycerol each have a high correlation co-
efficient with salicin but not with any other character. Salicin fermenta-
tion, on the other hand, is well correlated with three characters — glycerol,

41

TABLE XVIII. (a) SHOWING CORRELATION OF CHARACTERS AMONG 89
SUCROSE NEGATIVE STRAINS OF THE COLI SECTION.

Moli/ity

Indol

Dulcitol

6/ycerol

5alicif}

y-

—

-+•

—

-f-

—

y-

+•

I

+

53

45

8

22

31

43

/O

33

20

36

34

2

II

25

26

/O

/8

18

1

+

45

34

79

29

50

6/

/S

5/

25

8

S

ID

4

6

S

2

10

1

+

Z2

//

29

4

33

27

6

28

5

31

25

50

6

56

42

/4

23

33

^

/-

43

26

61

8

27

42

69

48

21

/O

/O

f8

2

6

14

20

3

17

1

/

33

/8

J/

28

23

48

3

&

-

20

18

26

/O

5

33

21

/7

38

TABLE XVIII. (b) COEFFICIENTS OF CORRELATION FOR EACH PAIR OF
CHARACTERS IN TABLE XVIII. (a).

tfoMty

Zvdo/

0i//cjfoi

(j/ycero/

Sa//c/v

WoM/ty

-.50

+.2Z

+ .25

+ .25

Indol

-.50

-.07

-08

+/.00

0ufoto/

+.22

+.07

+.20

+.78

(7/ycera/

+.£5

-.08

+.20

+.86

Sa/icio

+.25

+/.00

+.78

+.86

dulcitol, and indol, showing coefficients of +0.86, +0.78, and +1.0 re-
spectively. Subdivision is therefore made upon salicin.

For each of the resulting subgroups new correlation tables are made
and further subdivision again carried out on the best correlated char-
acter. A point is very quickly reached where further subdivision, upon
correlated characters, is no longer feasible. These groups are regarded
as species and to each was assigned, as far as possible, the name of the
MacConkey variety which it most resembled.

Levine's Classification. The following classification is suggested
by the author, based upon a study of 333 strains obtained from soil,
sewage, and the feces of man, horse, sheep, pig, and cow.

The characters employed are the methyl-red and Voges-Proskauer re-
actions, indol production, motility, gelatin liquefaction and gas formation
from sucrose, raffinose, dulcitol, glycerol, salicin, dextrin, inulin and corn

42

starch. Other fermentable substances — lactose, maltose, galactose and
mannitol — were also observed but as these substances were all attacked with
gas formation they need not be considered.

As was discussed in detail in Chapter II, the investigations of Harden,
Smith, Rogers and others on carbohydrate and nitrogen metabolism have
demonstrated adequately and conclusively that the colon group includes
two distinct subgroups, the methyl -red positive- Voges Proskauer negative-
uric acid negative group designated herein as the coli section, and methyl
red negative Voges Proskauer positive-uric acid positive group or aerogenes
section. These two main sections are recognized and each is subdivided
into species on the basis of correlated characters as described above.

KEY TO THE MORE IMPORTANT SPECIES OF THE COLON GROUP.

The colon group includes all non sporing Gram negative short rods,
fermenting glucose and lactose with acid and gas production and which
grow aerobically.

I. Not producing acetyl methyl carbinol (Voges Proskauer reaction
negative) ; acid to methyl red, and carbon dioxide and hydrogen
produced in approximately equal volumes from glucose. Can-
not utilize uric acid as a source of nitrogen.

COLI SECTION

A. Sucrose not attacked.

1. Salicin fermented with acid and gas.

1. Boot, coli

2. Salicin not attacked. 2. Bact. acidi-lactici

B. Sucrose fermented with acid and gas.

1. Motile. 3. Bact. communior

2. Non-motile.

a. Salicin fermented with acid and gas.

4. Bact. neapolitanum

b. Salicin not fermented. 5. Bact. coscoroba

II. Producing acetyl methyl carbinol (Voges Proskauer reaction posi-
tive) ; alkaline to methyl red, and forms two or more times as
much carbon dioxide as hydrogen from glucose. Capable of
utilizing uric acid as a source of nitrogen.

AEROGENES SECTION

A. Glycerol and starch fermented with acid and gas formation; non

motile, gelatin not liquefied. 6. Bact. aerogenes

B. Glycerol and starch not fermented; motile, gelatin liquefied.

7. Bact. cloacae

In the 1917 and 1920 Standard Methods of Water Analysis the coli
section is not differentiated into species. The aerogenes section is sub-
divided on the fermentation of adonitol into a fermenting variety sup-
posedly of fecal origin and a non-fermenting variety which is regarded
as of non-fecal origin. The value of adonitol for this distinction has
not as yet been adequately investigated and is not generally accepted.

43
IV. THE DETECTION OF THE COLON GROUP IN WATER.

Various methods have been suggested for the isolation of members
of the colon group. They are all based on the principle that all members
of this group of organisms are capable of decomposing lactose with acid
and gas production. Media are employed which contain lactose and some
indicator to show whether fermentation has taken place.

Isolation may be accomplished (1) directly by plating on solid dif-
ferential media or (2) indirectly on solid differential media, after pre-
liminary enrichment in some fluid medium.

The mediums most commonlly employed for direct isolation are:

Litmus lactose agar.

Phenol ated litmus lactose agar.

Endo agar.

Conradi Dragalski agar.

MacConkey agar.

Litmus lactose agar consists of nutriment agar, neutral to phenolph-
thalein, which contains sufficient litmus or azolitmin to give a distinct blue
color. Fermentation of lactose with the production of acid is indicated
by the formation of red colonies.

The technique for direct isolation is merely to place definite quan-
tities of the test sample (1.0 c. c., 0.1 c. c., 0.01 c. c., etc) into petri dishes
to which are added about 10 c. c. of litmus lactose agar. (In some labor-
atories the litmus is added to the petri dish and then lactose agar poured
in.) The plates are incubated at 37 degrees to 40 degrees C. for 24 hours
after which members of the colon group will appear as red colonies on a
blue background. As there are other organisms, particularly the strep-
tococci, which are capable of fermenting lactose) with acid production,
the mere appearance of acid colonies is not absolute proof of the presence
of the colon group. In the process of sterilization, lactose is sometimes
broken down so that organisms other than the colon group may then
produce slightly acid reactions. They must therefore be examined fur-
ther. Suspicious colonies are fished to agar and a Gram stain is made
to insure that the organism is a Gram negative non-spore-forming rod.
From the agar slant various other tests (fermentation of carbohydrates,
gelatin liquefaction, milk coagulation, the V. P., methyl red, etc.) may be
carried out.

Acid produced by the growth of fermenting forms diffuses through
the medium, often coloring considerable portions of the plate, making it
extremely difficult to detect the acid producing colony. The medium
exerts very little inhibitory action and overgrowths of non-colon forms
are not infrequent. These disadvantages may be overcome to some ex-
tent by (1) the use of porous tops to prevent spreaders, (2) by increasing
the concentration of agar to 3 percent thereby diminishing considerably
the diffusion of acid, and (3) by the addition of selective antiseptics to
inhibit the growth of forms other than the colon group. The use of the
higher temperature (37 degrees C.) and preferably 40 degrees C., exerts

44

some inhibitory action on the ordinary water forms and many soil forms.
This may be further increased by the addition of small quantities of phenol
to the agar medium. Chick found that 0.1 percent (1 part per thousand)
gave very excellent results. In using phenolated media, it must always
be borne in mind that there is danger of inhibiting some members of the
colon group as well.

Endo agar consists of nutrient lactose agar containing 3 percent agar
with basic fuchsin decolorized by sodium sulphite as an indicator. The
lactose fermenting organisms produce red colonies often with a distinct
metallic luster, the medium itself being colorless or slightly pink. The
reaction is presumably due to the production of acid and aldehyde by the
lactose fermenting organisms. These products react with the fuchsin
sulphite combination liberating the fuchsin. The high concentration of
agar serves to check diffusion of acid and also to eliminate many of the
water forms. It was thought at one time that the indicator itself exerted
inhibitory action but this does not seem to have been adequately proven.

Conradi-Drigalski agar consists of litmus lactose agar to which has
been added some nutrose and crystal violet. The crystal violet checks
the growth of many forms particularly the streptococci but is supposedly
non-inhibitory for the colon group in the concentration employed. Both
the Endo and the Conradi Drigalski have been found very valuable for
the isolation of the para typhoid and typhoid bacilli from stools but they
have never found favor in the United States although they are the media
of preference in Germany for direct isolation of the colon group from water.

MacConkey agar (also known under the name Rebipelagar) consists
of peptone lactose agar with 0.5 percent bile salts (sodium glycocholate
and taurocholate), the function of the latter being to check saprophytic
forms. The inhibitory action of bile is well illustrated in Table XX.

TABLE XX. SELECTIVE ACTION OF BILE SALTS.
(After Jackson 1906)

Bacteria

per. c. c.

Uncontaminated
well

Contaminated
ponH

Suspension
of feces

Suspension
of feces

Gelatin 200

920

2700

350.000

900,000

Agar 370

25

170

450.000

900,000

Bile Agar *3?0

14

43

300,000

900,000

Bile Agar 370

0

16

60.000

900,000

Lactose Bile
Agar *370

0

25

250,000

675,000

Lactose Bile
Agar *370

0

17

250,000

600,000

*Bile diluted, 1:1.

where it will be noticed that the bile salts exerted a much greater re-
straining action on forms generally found in water than on the organisms
(almost exclusively Bact. coli) present in a suspension of feces. The

45

indicator in the Rebipelagar is neutral red. Acid producing colonies are
distinctly red. This medium is very extensively used in England but has
not found favor in the United States for, as will be shown later in the
discussion of the presumptive test for Bact. coli, there is considerable evi-
dence that the bile salts may inhibit some strains of Bact. coli. Reference
to the foregoing table shows that in one of the fecal suspensions a count
of 450,000 on agar was reduced to 60,000 by bile.

The isolation of the colon group by direct plating is recommended for
highly polluted waters and sewage where the number of coli strains are
quite large, being present in one c. c. or less. If, however, it is necessary
to employ large samples (5 c. c. or 10 c. c.) as is essential in routine
water work, it becomes extremely inconvenient to plate out. Furthermore,
if the incidence of the colon group is small compared to other forms, they
may be missed. Although theoretically the colon colonies are clearly dif-
ferentiated from others, it is the practical experience of all investigators
that when growing on a plate with a large number of alkali formers, the
acid of Bad. coli and related forms may be neutralized, and the characteris-
tic reaction thereby masked, so that a colon colony may thus escape de-
tection. On the other hand, in the process of sterilization of medium the
lactose may be decomposed in part and it will then be found that some
soil forms will produce an acid reaction on solid media. Acid production
on any of the solid media described above can therefore not be relied upon
to always indicate the colon group, and so further tests must be resorted
to as has been mentioned in discussing litmus lactose agar.

For the isolation from relatively large quantities of water, or for an
admixture with numerous other forms, the best results are obtained by the
indirect or so-called preliminary enrichment method. The procedure con-
sists of three steps:

1. Growth in a favorable liquid medium containing some constituent
which will indicate the probable presence of the colon group. This is
spoken of as a preliminary enrichment tube and is often employed as a pre-
sumptive test.

2. Isolation from this preliminary enrichment tube on some one of
the solid media described above. If typical colonies are formed the pres-
ence of members of the colon group is considered partially confirmed.

3. Suspicious colonies on (2) are examined, by fishing to agar and
then testing for coagulation of litmus milk, gas in lactose broth, V. P., etc.
This would constitute a completely confirmed test.

The media most commonly employed for preliminary enrichment are:
Glucose broth
Phenol broth
Liver broth
Lactose broth
Lactose bile salts broth

Eijkman test (glucose peptone water at 46° C.)
Lactose peptone bile

46

Glucose broth consists of ordinary broth, neutral to phenolphthalein,
with 1 percent glucose. This medium was the one most commonly em-
ployed in the United States prior to 1906. The probable presence of the
colon group is indicated by the production of gas after 24 to 48 hours
incubation at body temperature.

Phenol broth consists of ordinary broth containing 0.1 percent car-
bolic acid which exerts a selective antiseptic action on the normal water
bacteria thereby supposedly favoring the growth of the colon types. This
medium is employed by the French, but it has been observed that the phenol
exerts an antiseptic action on some members of the colon group and that
its employment is particularly objectionable with water of fairly good
quality.

Liver broth has been found particularly efficacious for isolation of
colon types and in 1912 the Standard Methods for Water Analysis of the
A. P. H. A. recommended that, "If a further study of all gas forming
bacteria, including attentuated forms, is desirable, then liver broth should
be employed in preference to the usual dextrose broth, as it gives a larger
number of attentuated forms, has better rejuvenating powers, and gives
fewer anomalies, and greater and more rapid gas production." For routine
water work, however, liver broth has not found favor.

Since 1906, there has been a tendency to substitute lactose for dextrose
in the preliminary enrichment mediums. It is quite surprising that this
was not done sooner considering that the essential characteristic of the
colon group is lactose fermentation.

Lactose broth, which is nutrient broth containing 0.5 to 1.0 percent
lactose, is most commonly employed in the United States at the present
time and is the medium recommended in the Standard Methods of Water
Analysis of the A. P. H. A. of 1917 and 1920. It is supplanting the lac-
tose peptone bile medium of Jackson, which was extensively used between
1912 and 1917. In both of these mediums the probable presence of the
colon group is indicated by gas production from the lactose.

Lactose bile salt broth of MacConkey consists of 0.5 percent lactose,
2% peptone, 0.5 percent sodium taurocholate with neutral red as an in-
dicator. The colon group will produce gas and acid indicated by red
coloration of the liquid after 24 to 48 hours at 37° C.

Technique for Isolation. The steps in the isolation of the colon
group from water may be summarized as follows:

(1) Place various quantities of the water under examination (10 c. c.,
1 c. c., 0.1 c. c., etc.) into Smith or Durham fermentation tubes containing
the preliminary enrichment medium.

(2) Incubate at the body temperature (37° C.) for 24 hours. If
gas is formed, it is taken as presumptive evidence that Bact. coli or its close
allies are present.

(3) The next step is then to isolate these organisms on some one of
the solid media enumerated above. This may be done in one of the fol-
lowing ways:

47

Method 1. Barely touch the liquid in the preliminary enrichment
tube, showing gas, with the point of a platinum needle, wash off this
needle in a tube of melted litmus lactose or Endo agar and pour into a
petri dish. Incubate at the body temperature for 24 hours.

Method 2. Pour L. L. A., Endo or other differential agar into petri
dishes and allow them to solidify and dry in the incubator. Dip a small
loop into a positive preliminary enrichment tube and wash it off in a
tube of sterile salt solution. Place a loop of this wash water in the
center of a plate containing the differential medium and spread it over
the surface with the aid of a sterile glass rod bent at right angles. Incubate
20 to 24 hours at body temperature.

Method 3. Dip a platinum needle, whose end is bent at an angle
of about 120°, into the positive preliminary enrichment tube, stab into a
plate containing the differential medium (to remove the excess organisms)
and then make a series of streaks on this plate about a quarter of an inch
apart, taking care to always streak in the same direction and to lift the
needle at the end of each stroke. With a little practice, very excellent
isolations may be obtained by this simple method. This was first demon-
strated to the author by Mr. Greenfield of the Illinois Water Survey and
it has proved very satisfactory. Incubation is, as before, at body tem-
perature for 24 hours.

TABLE XXI. RELATIVE GROWTH OF BACT. COLI AND SEWAGE STREPTO-
COCCI FROM POLLUTED WATER IN GLUCOSE BROTH.
(After Prescott and Baker 1904)

Sample

No.

Initial
number
of red
colonies
per c. c.
on L. L. A.

Type

Number found in million per c. c.
after various periods of growth

First
gas
noticed
after

11 hrs.

16 hrs.

23 hrs.

39 hrs.

63 hrs.

2

10

Bact.
coli

20

76

150

0

0

11
hrs.

Strep.

0

25

85

420

0

7

35

Bact.
coli

130

140

200

20

8

10
hrs.

Strep.

0

20

110

400

45

8

460

Bact.
coli

332

420

405

24

0

6

hrs.

Strep.

0

210

350

105

150

If characteristic colonies are present on these agar plates the test is
regarded as partially confirmed for the colon group. For further con-
firmation, it would be necessary merely to fish a well isolated colony and

48

to determine its Gram characteristics and other reactions, particularly fer-
mentation of lactose broth, as previously explained.

The objections to the preliminary enrichment method are:

(1) It is difficult to estimate the incidence of the colon forms for
by employing Ic. c., 0.1 c. c., etc., there is considerable error introduced
in calculating the number of organisms present in a unit volume of water.

(2) There is a danger of loss of the colon forms which may be killed
off before isolation is attempted either by (a) products of their own meta-
bolism or (b) by over-growths of other organisms.

In glucose broth inoculated with a mixture of Fact, coli and Strepto-
cocci, the former usually predominated for the first 24 hours, but thereafter
the number of Streptococci was considerably greater, in some instances
having completely choked out or killed the colon types in 39 hours.

The danger of losing colon strains due to overgrowth and death may
be minimized considerably, if not completely eliminated, by (1) reducing
the concentration of fermentable substances to 0.5 percent (as recommended
by Burling and Levine) and (2) attempting isolation at the earliest pos-
sible time (after 12 to 24 hours or at the first appearance of gas). It will
be noticed from Table XXI. that at the first appearance of gas the colon
forms were far in excess of the Streptococci, in fact practically in pure
culture. With these precautions the preliminary enrichment method will
be found very reliable and convenient.

THE PRESUMPTIVE TESTS

The method for isolation of members of the colon group described
above, comprising as it does the reaction of the isolated organism in milk,
production of indol, reduction of nitrates, gelatin liquefaction, and mo-
tility was in general use between 1905 and 1910. The method, however,
was soon found to be tedious and inconvenient for routine work. The
time required for such a complete analysis was at least nine days, whereas
for practical purposes it is desirable to have a report available within
24 to 48 hours. Furthermore, there is no evidence that the determination
of the characters enumerated shed any light upon the sanitary significance
of the bacilli isolated. For routine water work, it is essential to have some
test which is simple and rapid and which can be relied upon as a reason-
able index of the probable presence of the colon group in a large pro-
portion of instances. Such a test is known as a presumptive test for
the colon group. From his wide experience with the glucose broth pre-
liminary enrichment method, Whipple in 1903 observed that a consider-
able portion of tubes showing characteristic fermentation were found, on
further examination, to contain colon-like organisms. He therefore sug-
gested that dextrose broth alone might be employed as a rapid presumptive
test of the probable presence of Bact. coli. The media which have been
and are most commonly employed for the presumptive test are (1) glucose
broth, (2) lactose peptone bile, and (3) lactose broth. Other media have
also been suggested, as for example the Eijkman test (glucose peptone

49

water at 46° C.) and neutral red broth, but these are not employed to any
extent in the United States and will not be considered further.

The Glucose Broth Presumptive Test. The test consists of the
determination of the amount of gas and the ratio of hydrogen to carbon
dioxide in glucose broth after 24 hours incubation at the body temperature.

Production of 25 to 70 percent gas, one-third of which is carbon
dioxide and the remainder hydrogen (that is H/C02=2 to 1) is con-
sidered a "typical" test and is regarded as indicating the presence of
Bact. coli or closely related bacilli. Thus Irons stated that when the pro-
portion of carbon dioxide is approximately 33 percent Bact. coli commune
is almost invariably present.

If the amount of gas is (a) between 10 and 25 percent, (b) more than
70 per cent or (c) if the proportion of carbon dioxide is more than 40
percent of the total volume of gas formed, the reaction is considered
"atypical" and the probable presence of the colon group is doubtful.

If there is less than 10 percent gas, or if no gas is formed at all, the
probability of isolation of the colon group is very slight and the test is
considered negative.

The early studies indicated that the colon group was successfully iso-
lated from about 70 percent of positive presumptive tests. More recent
and detailed studies have shown, however, that the proportion of presump-
tive tests confirmed by further observation is not constant but is effected by
the character, treatment, etc. of the water under examination and the season
of the year.

Objections to the Glucose Broth Presumptive Test. The ob-
jections to the glucose broth presumptive test may be summarized as fol-
lows:

1. The volume of gas produced is not a significant criterion of the
probable presence of the Bact. coli or its close allies.

2. The gas ratio, as determined in the Smith tube, is unreliable.

3. There are many glucose fermenting species which are incapable of
attacking lactose.

4. The reliability of the glucose presumptive test varies with (a)
the season of the year, and (b) differs for raw and treated waters.

5. The glucose broth medium is particularly subject to overgrowths.
That 25 to 75 per cent gas is not necessarily characteristic of the colon

group and that even pure cultures produce as little as 10 percent gas has
been demonstrated by Fuller and Ferguson, Longley and Batton, Hale and
Melia, and others. They also have shown that the so-called 'typical' gas
ratio is not at all characteristic of the colon group. Thus of 818 tubes of
sterilized water, inoculated with Bact. coli but containing no other gas
former, only 474 or 58 percent gave the typical gas volume and gas ratio.
The errors inherent in the gas ratio as determined in the Smith tube have
already been discussed in detail in Chapter II.

Fermentation of glucose with gas production is not a very reliable in-
dication that the organism will also ferment lactose. There are many
species of bacteria capable of attacking the monosaccharids, but which

50

are inert as respects the more complex disaccarid lactose. Thus Clark
and Gage in 1903 pointed out that there were 58 well described species
which gave the presumptive test in dextrose broth of which 23 are widely
separated from the colon group and since then many other glucose-positive-
lactose-negative forms have been described.

Prescott and Winslow call attention to the seasonal variation in the
reliability of this presumptive test noting that Fromme (1910) isolated
colon bacilli from 87 percent of positive glucose broth tubes in the fall
and winter (Oct. to Mar.) but only 66 percent from similar tubes in the
spring and summer (Apr. to Sept.)

In stored or filtered waters the incidence of glucose fermenters which
do not attack lactose is considerably greater than in raw waters. The
reports of Houston for 1909-11, which are summarized below, show that
only 9.5 percent of coli-like bacteria isolated from raw river water after
preliminary enrichment in glucose broth failed to ferment lactose but that
with filtered and ground waters 38.3 percent of the isolated bacilli did not
attack this disaccharid.

TABLE XXII. GLUCOSE AND LACTOSE FERMENTATION BY COLI-LIKE

BACTERIA FROM RAW AND FILTERED WATERS.

(After Houston 1909 to 1911)

Raw River waters | Filtered waters*

Number of Strains

7,657 | 10,620

Glucose+; Lactose —
Glucose-f ; Lactose +

9.5%
90.5%

38.3%
61.7%

*Includes ground waters.

TABLE XXIII. EFFECT OF NATURAL AND ARTIFICIAL PURIFICATION ON

INCIDENCE OF GLUCOSE AND LACTOSE FERMENTING

BACILLI IN PULTA WATER.

(After Clemesha 1912)

Intake

Settled

Filtered

"No. of
strains

% Lactose
negative

*No. of
strains

% Lactose
negative

*No. of|% Lactose
strains I negative
)

October
(Heavy rain)

40

40.0

40

50.0

40

50

November
(River muddy but no rain
locally)

80

31.2

80

16.2

80

48.7

December
(River clearing)

60

70.0

60

85.0

60

91.6

January
(River clear)

100 | 61.0

100

71.0

60

100.0

February
(River clear)

80

93.7

80

98.7

60

98.2

rAll glucose fermenters.

51

Clemesha (1912) has made numerous observations on the relative
incidence of glucose and lactose fermenting bacilli in waters of India
subjected to various periods of storage, exposure to sunlight, and filtration.
In Table XXIII. taken from his book on the bacteriology of Surface Waters
in the Tropics are shown the effects of treatment and the influence of self
purification on the relative incidence of glucose fermenting bacilli which
are not lactose fr actors.

The increased incidence of glucose positive-lactose negative organisms
in waters during dry weather is particularly striking. Thus during the
rainy month of October 40 to 50 percent of the strains studied were of this
type, the number gradually increasing as the river cleared and the forces
of natural purification (sedimentation, sunlight, etc.) manifested them-
selves until in February, 93.7 to 98.2 percent of the glucose fermenting
strains were non-lactose fractors. The glucose broth presumptive test is
therefore particularly unreliable as an index of the colon group when
dealing with treated and stored waters.

Clemesha attributes the marked increase of glucose -(-, lactose — ,
bacilli in India stored waters to the presence of an organism which he
designates Bacillus P. This bacillus is widely distributed in nature, ex-
tremely resistant and capable of multiplying in water.

The important characteristics of the organism are given below:

Morphology, etc. Gram positive, motile bacillus.

Litmus Milk. Slight acidity on top (occasionally). No coagulation.

Indol. Negative.

Gelatin. Not liquefied in ten days.

Fermentation. Glucose acid and gas.
Sucrose acid and gas.
Lactose. Faint trace of acid on long cultivation in

laboratory.
Dulcitol

Mannitol
Adonitol

Neither acid nor gas

Inulin

Voges Proskauer Reaction. Strongly positive.
Bacillus P. of Clemesha resembles the B. proteus but differs from it
in being Gram -|-» and giving a positive Voges Proskauer reaction.

Not only is gas production in glucose broth a poor index of probable
presence of Bact. coli but the organism may be present even in the absence
of gas formation, particularly if the water has a high bacterial count or
contains streptococci. It is not infrequently observed with this medium
that small samples of water (1 c. c. or 0.1 c. c.) may be positive whereas
larger portions (10 c. c.) may be negative. Upon considering that there
are numerous organisms capable of attacking glucose with acid produc-
tion, this choking off of the colon bacilli may be readily explained by the
inhibiting action of the acids formed.

52

This lack of constant relation between the dextrose broth fermentation
and the actual presence of the colon group has resulted in a gradual abandon-
ment of the glucose broth presumptive test in favor of media contain-
ing lactose.

The Lactose Peptone Bile Presumptive Test. The use of bile
containing lactose was first suggested in this country by Jackson in 1906.
He recommended fresh ox bile containing 1 percent lactose but the in-
clusion of 1 percent peptone was soon found to be desirable. The medium
is prepared by adding 1 percent peptone and 1 percent lactose to fresh
undiluted ox gall (or a 10% solution of dried ox gall) which is placed
in fermentation tubes and sterilized in the usual manner. He found, how-
ever, that a longer period of incubation than is employed in dextrose
broth presumptive test is necessary and recommended 72 hours. For or-
dinary work a period of 48 hours is ample. In this medium, the pro-
duction of 10 percent or more gas is considered a positive presumptive
test. The proportion of positive presumptive tests found on further
study to contain the colon group is very high, Stokes and Stoner reporting
95 and Gumming 87 percent. Its superiority to dextrose broth is quite
marked. Thus Prescott and Winslow, from a study of 176 surface waters
in eastern Massachusetts, obtained 120 positive fermentations with dextrose
broth and 78 with lactose bile; but Bact. coli was isolated from only 70
(58%) of the dextrose as compared with 64 (82%) of the positive bile
tubes. Reliance on the dextrose broth presumptive test would have intro-
duced an error of 70 percent as compared to an error of 11 percent for the
lactose bile medium.

The superiority of lactose bile, -as a presumptive test, is evident. This
is due partly to the selective antiseptic action of the bile salts but also
to a considerable extent to the use of lactose which automatically elimin-
ates the dextrose positive-lactose negative organisms. It should also be
noted that the presence of lactose may indirectly inhibit non-lactose fer-
menters for by stimulating the growth of the colon group, the resultant
acid production through fermentation of the disaccharid may in turn in-
hibit markedly the growth of water forms. In lactose media the probabiK
ity of the colon group being overgrown is considerably reduced as com-
pared with glucose media where acid may be produced by many other
bacilli resulting in the retardation of the colon group itself.

Consideration of the example cited above indicates, however, that
although lactose bile is superior as a presumptive test, the actual number
of positive isolations with this medium (64) was about 8.5 percent less
than was obtained with dextrose broth (70). Apparently therefore the
inhibitory action of the bile is not restricted to water forms, but is exerted
to some extent, on the colon group as well.

The Committee on Standard Methods (1912) suggested that this in-
hibitory action is exerted on the weaker members of the colon group,
strains which have become attenuated, so to speak, and which are indicative
only of remote pollution and therefore of apparently little significance.
Thus they say, "In the interpretation of the sanitary quality of a water, it

53

is best to discount the presence 'of attenuated Fact, coli and to be sure to
obtain all vigorous types. The lactose bile medium accomplishes both these
objects."

This contention, that bile affects the weaker strains, must be regarded
merely as an assumption, for it is not substantiated by actual observations.
Jordan, in a careful study of the subject, concludes that there is no re-
lation between attenuation and the antiseptic action of the bile. He found
that freshly isolated strains were inhibited to as great or even greater
degree than old strains. Gumming, in a comparison of lactose bile and
lactose broth, found that with sewage preliminary enrichment in lactose
bile yielded only 25 percent as many colon forms as were obtained with
lactose broth. In a similar study with river water, the bile method gave
50 to 70 percent as many colon organisms as the broth. These results
would indicate quite the converse of the contention of the 1912 report of
the Committee on Standard Methods that "Attenuated B. coli does not re-
present recent contamination, and all B. coli not attenuated grow readily
in lactose bile."

The 1917 and 1920 reports of the Committee on Standard Methods of
Water Analysis recommend the use of lactose broth for preliminary en-
richment and the presumptive test on the ground tha it gives a higher total
number of positive isolations. The concensus of opinion at present seems
to be in favor of the substitution of broth for bile.

Lactose Broth vs. Lactose Bile. Jordan in 1913 called attention
to the greater proportion of successful isolations of T act. coli with lactose
broth. Of forty — 5 c. c. portions of Lake Michigan water inoculated
into lactose broth and lactose bile media, colon bacilli were isolated from
42 percent of the former and only 30 percent of the latter. Similarly one
hundred and fifty — 1 c. c. samples yielded 31 percent positive isolations
with lactose broth compared to 22 percent when lactose bile was em-
ployed for preliminary enrichment.

Creel in a study of drinking waters on railroad trains reports the
following frequencies of the colon group in comparative tests:

Water samples containing colon group 91

Positive from lactose bile only 18

Positive from lactose broth only 45

Positive from both media 27

The inhibitory action of the bile is very marked. About 50 percent
of the colon bacilli isolation would have been lost if reliance had been
placed on lactose peptone bile alone.

Similarly Dr. Gumming in the Potomac River studies calculates the
number of colon bacilli to be 84 per c. c. by lactose broth enrichment
followed by confirmatory tests as compared to 47 when lactose bile was
employed and followed by confirmatory tests, again indicating that about
half of the colon bacilli were lost when preliminary enrichment was car-
ried out in the bile medium.

Obst comes to the same conclusion. Houser of the Cincinnati Water
Supply prefers the broth medium. Ritter in a detailed comparison of

54

both media concludes that if both lactose broth and lactose bile are positive
in 24 hours, the probability of the presence of the colon group is very
great (about 98% confirmed by subsequent tests) and that, in general,
if there is fermentation in both bile and broth media the presumptive test
is reliable in about 75 percent, but if either medium is negative, then the
proportion of positively confirmed tubes is very much less, even as low
as 25 to 45%. This high proportion of positively confirmed presumptive
tests, when both broth and bile are positive, may find a ready explanation
in the observations of Creel. Most of the spurious presumptive tests are
due to the presence of spore forming anaerobic lactose fermenters. Creel
found that some of these anaerobes are capable of growing in broth
but not in bile. He calls this "Group A" whereas another "Group B"
will grow readily in bile but not in broth. Thus a positive test in both
lactose broth and lactose bile automatically eliminates both of these anaer-
obic spore formers and thereby accounts for the high proportion of pos-
itively confirmed tests.

Opposed to the views expressed above is that of Hale who recently
(1917) stated that from a careful detailed, and extensive comparative
study on the two media at the Mount Prospect Laboratory, he considers
the results were all in favor of bile, that gas formation was more rapid,
produced in larger amounts, and that the B. clostridium welchii types (an
aerobic spore formers) were less frequent. Probably his experiences
may be explained by the fact that, in these experiments, he employed a 5 per-
cent dried bile, whereas the observations of other investigators were con-
cerned with the original medium, containing either undiluted ox gall or
10 percent of the dried bile.

The work of Salter and the author in the Laboratory of the Iowa En-
gineering Experiment Station indicated quite clearly that bile may inhibit
or stimulate multiplication of Bact. coli, depending upon the concentration
of bile salts. With less than 0.5 percent bile salts a stimulating action
on the rate of multiplication was observed, whereas higher concentrations
were markedly inhibitory. The effect was studied by observing changes in
the generation time (See Table XXIV.).

Since bacteria divide by simple fission, the number of organisms pres-
ent at any time in an actively growing culture may be expressed thus:
b = B X 2n

Where "B" is the initial number of bacteria
"n" the number of generations
"b" the number of bacteria after "n" generations

If the time elapsed is "t" then "g" the generation time — equals
t t

— or "n" equals — .

n g

Thus b = B X 2n = B X 2%

t Iog2
° ~ logb — logB

55

TABLE XXIV. EFFECT OF CONCENTRATION OF COMMERCIAL BILE SALTS

ON GROWTH OF BACT. COLl IN 0.5 PERCENT PEPTONE WATER.

(Temp, of incubation 37° C.)

Concentration
of bile salts

Bacteria per c. c. After

Average
generation
time in minutes

0 hrs.

2 hrs.

4 hrs.

8 hrs.

Control
0.0%

124

410

12,800

14.600.000

28.7

0.1%

124

595

17,900

23,700,000

27.5

0.2%

124

515

17,900

36,000,000

26.6

0.3%

124

450

17,000

24,700,000

27.7

0.5%

124

520

20,200

21,500,000

27.8

0.7%

124

425

13,400

10,800.000

29.5

1.0%

124

435

7,150

1,650,000

35.4

The experience of Hale, that Cl. Welchii forms were less frequently
encountered in bile than in broth, may merely mean that in the waters he
worked with Creel's anaerobe of "Group A" were more frequent than
"Group B". In other laboratories, the reverse might be true.

The relative merits of lactose broth and lactose bile may be summed
up as follows:

1. Lactose bile is a more reliable presumptive test but a greater
proportion of the colon group may be detected by preliminary enrichment
in lactose broth.

2. If the proper concentration of bile salts could be determined the
bile medium would probably be preferable. For the present, considering
the difficulty of obtaining bile of constant composition or the chemically
pure salts, and in view of our insufficient knowledge as to the optimum
concentration of bile salts, it seems best to employ lactose broth as a more
uniform medium may thus be obtained in different laboratories. It is
very probable that if a standardized evaporated bile were available a con-
concentration of 1 to 2 percent in lactose peptone water would be superior
to lactose broth.

THE LACTOSE BROTH PRESUMPTIVE TEST.

The use of lactose broth for preliminary enrichment and as a pre-
sumptive test has received considerable impetus through the investigations
of the Public Health Service and has been accepted by the Committee
on Standard Methods of Water Analysis of the A. P. H. A. (1917).

Factors Affecting the Preparation of Lactose Broth. In 1917,
the Standard lactose broth consisted of 0.3 percent beef extract and 0.5
percent peptone with 1.0 percent lactose, the reaction being neutral to
phenolphthalein. The medium was tubed and sterilized in the autoclave
at 15 pounds (120° C.) for 15 minutes.

Hasseltine pointed out that sterilization in the autoclave was objection-
able as it brought about a breaking down of the lactose resulting in a con-
siderable increase in the number of unconfirmable presumptive tests as com-

56

pared with the lactose broth medium employed in the Public Health Ser-
vice.

The Public Health Service lactose broth is made as follows: Nutrient
broth neutral to phenophthalein is prepared in bulk and sterilized in the
autoclave. A sufficient quantity of a 20 percent solution of lactose in
distilled water, previously sterilized in the Arnold for an hour and a half,
is then added to make a concentration of 1 percent. The medium is then
distributed aseptically into sterile Smith fermentation tubes, which are
then heated in the Arnold for 30 minutes.

Hasseltine found that the "Standard" lactose broth gave 73 percent more
positive fermentations and yielded 10 percent fewer confirmations than the
Public Health Service medium. With reference to these observations, it
should be pointed out, however, that in the preparation of the Standard
broth, the period of sterilization was quite prolonged. He states that the
broth was in the autoclave for about an hour (25 minutes to raise the
pressure to 15 pounds, 15 minutes at that pressure, and about 20 minutes
to allow the pressure to fall sufficiently to permit opening the autoclave
without blowing out or wetting the stoppers). In the autoclave in which
steam must be generated this prolonged period is of course probably neces-
sary but in autoclaves provided with pressure steam, the entire period of
sterilization may be reduced to 25 or at most 30 minutes; with such an
instrument, the very significant objections raised by Hasseltine do not
apply. In fact, Mudge, in a careful study of the relative effects of tem-
perature and pressure as compared with time of exposure, concludes that
the time factor is the more vital one in the decomposition of lactose. He
maintains that heating in streaming steam for three successive days, as is
ordinarily done in the Arnold, will cause a greater hydrolysis of lactose and
maltose than sterilization in the autoclave at 15 pounds for 15 minutes.
These conclusions are based upon the relative increase in monosaccharids
as indicated in Table XXV. By use of bacterial cultures confirmatory re-
sults were obtained.

TABLE XXV. EFFECT OF METHOD OF STERILIZATION ON DECOMPOSITION

OF CARBOHYDRATES.
(After Mudge 1917)

Method of
sterilization

Lactose

Maltose

Sucrose

Raffinose

Autoclave 15 minutes

Trace*

1

0

0

Autoclave 30 minutes

1.5*

3

0

0

Autoclave 60 minutes

2.0

Very great

Slight

0

Arnold 1st day

0

4

0

0

Arnold 2nd day

1

Very great

0

0

Arnold 3rd day

2

Very ^reat

0

o-

*Figures indicate relative degree of reduction as determined by Barfoed's method
• for monosac,charids. .

Mudge also makes an interesting observation as to the cause of in-
creased acidity in sterilized culture media. As is well known, a neutral

57

solution of amino acids become acid on the addition of neutral formalde-
hyde. Sugars may be considered as polymers of formaldehyde, and, as some
sugars have aldehyde groups capable of reacting, it is suggested that the
increased acidity in sterilized sugar media may be due to the interaction
of the amino acids in the medium with the aldehyde group of the sugar.
This reaction takes place very slowly and does not become evident before
30 minutes autoclaving. In Table XXVI. this contention is very clearly
demonstrated. Neither asparagine, maltose, lactose, nor rafinose become
acid on autoclaving, even for an hour,, but if maltose or lactose are
heated in the presence of asparagine a distinct acidity develops in 30 min-
utes, and a much more marked acidity after one hour. With raifinose, on
the other hand, there is no increase in acidity. The disaccharids (maltose
and lactose) contain reactable aldehyde groups whereas the trisaccrid,
raffinose, does not. The contention of Mudge that it is the presence of
an unstable sugar molecule together with an ammo acid which gives rise
to the acidity on sterilization appears very plausible.

TABLE XXVI. EFFECT OF AN AMINO ACID (ASPARAGINE) ON PRODUCTION
OF ACID FROM SUGARS, BY AUTOCLAVING.

(After Mudge 1917)

Medium

Acidity ( % normal) after autoclaving for

15 min.

30 min.

60 min.

Asparagine

0.0

0.0

0.0

Maltose

0.0

0.0

0.0

Lactose

0.0

0.0

0.0

Raifinose

0.0

0.0

0.0

Asparagine -j- Maltose

0.0

0.44

0.58

Asparagine -f- Lactose

0.0

0.34

0.64

Asparagine + Raffinose

0.0

0.0

0.0

In the revised Standard Methods for 1920, two important modifications
in the preparation of lactose broth were recognized:

(1) The method of sterilization recommended for autoclaving is 15
pounds for 15 minutes provided the total time of exposure to heat is not
more than a half-hour. "Otherwise a 10 percent solution of the required
carbohydrate shall be made in distilled water and sterilized at 100° (for
one and a half hours), and this solution shall be added to sterile nutri-
ment broth in amounts sufficient to make a 0.5 percent solution of the
carbohydrates and the mixture shall then be tubed and sterilized at 100° C.
for 30 minutes, or it is permissable to add by means of a sterile pipette
directly to a tube of sterile neutral broth enough of the carbohydrates to
make the required 0.5 percent. The tubes so made shall be incubated at
37° C. for 24 hours, as a test for sterility."

(2) The concentration of lactose is reduced to 0.5 percent. This
seems to be desirable as it tends to eliminate the danger of loss of the
colon group through destruction by acid formed in fermentation.

58

Clemesha, in 1912, found that in peptone bile salt, neutral red broth
containing 1 percent lactose, pure cultures of Bact. coli and Bact. aerogenes
begin to fall off in numbers after 24 hours. Burling and Levine observed
that the concentration of carbohydrates influences markedly the viability
of the colon group in culture media. With 0.3 percent glucose or lactose,
the number of viable organisms reached a maximum of 100 to 1,000 mil-
lion in about 10 hours and remained constant for about 7 days. Increasing
the amount of carbohydrates to about 0.5 percent resulted in death of
over 99 percent of the organisms in a week, whereas with a concentration of
1 percent of the carbohydrates a count of 100 million, after 24 hours, was
reduced to less than 10,000 (a reduction of 99.9 percent in 72 hours.)

In preparing carbohydrate media the author has used the following
method with excellent results:

The broth containing 0.5 percent of the carbohydrate is distributed in
sterile Durham fermentation tubes, using ordinary precautions but not
necessarily strict asepsis, and autoclaved at 10 pounds for 10 to 15 minutes
in a pressure steam sterilizer. Immediately after removing from the auto-
clave, the tubes are cooled by placing in running tap water after which
they are incubated at 37° C. for 24 to 48 hours to eliminate unsterile tubes.
Several tubes are inoculated with Bact. enteritidis or Bact. paratyphosum
and incubated for 24 hours to detect hydrolysis as shown by gas formation.

Wagner and Monfort have recently reported that if gentian violet
(1-20,000 to 1-100,000) be added to lactose broth, the medium may be
prepared by a single heating in the Arnold or simply pasteurization. If
further experience proves gentian violet non-inhibitory to the growth of
Bact. coli., such a medium would be practically ideal, for not only is the
danger of inversion of the lactose eliminated, but growth of the spore
forming lactose fermenters is also reported to be checked.

In recording the presumptive test with lactose broth (this applies to
lactose bile also), the production of 10 percent or more gas is regarded
as a positive presumptive test. If the amount of gas is less than
10 percent the test is regarded as questionable. If there is no gas it is of
course negative. Experience indicates that where the quantity of gas is
more than 70 percent the fermentation is very likely due to anaerobic spore
formers rather than to the colon group.

Reliability and Value of the Lactose Broth and Lactose Bile
Presumptive Tests. In studies on the Ohio and Mississippi Rivers,
Fuller noticed that the results with these presumptive tests were far be-
yond what could be explained by ordinary pollution and much higher
than results obtained by a complete isolation test. Similarly in studies
on the Potomac River, Cummings found these tests were reliable when pol-
lution was excessive, but as purification progresses, the correlation be-
tween the presumptive and confirmatory tests become less marked. Using
Endo agar for confirmation 92 percent of positive presumptive tests were
confirmed in regions close to the source of pollution but in the vicinity of
oyster beds, which were further removed from pollution, only 47.5 per-
cent of the positive presumptive tests were found to contain members of

59

the colon group. The reliability of these presumptive tests therefore varies
directly with the degree of contamination or inversely with the remoteness
in time and distance from the source of pollution. The non-confirmed
presumptive tests are due, for the most part, to lactose fermenting spore
producing anaerobes which persist for a long time in water and which
are probably of very little sanitary significance.

Hall and Ellefson have suggested the addition of gentian violet to
lactose broth to restrict the growth of these anaerobes, and recently Meur
and Harris indicate that brilliant green may be employed in lactose bile to
eliminate these spurious presumptive tests. This will be considered more
in detail in a subsequent chapter.

The reliability of the presumptive test also varies with the treatment
which the water has received. Hauser, using lactose broth, obtained con-
firmations in 97.4 percent of raw water samples as compared with 86.1
percent of samples taken at the outlet from the clear water reservoir.

The effect of chlorination on the reliability of the presumptive test
is particularly marked and is shown in the following able:

TABLE XXVI. SHOWING CORRELATION OF RATE OF GAS PRODUCTION

WITH CONFIRMATION OF THE PRESUMPTIVE TEST FOR COLON

GROUP IN LACTOSE BROTH.

Rate of Gas Production

Untreated supplies
\

Chlorinated supplies

Number
of tubes
showing
gas

Percentage
of gas
tubes con-
firmed

Number
of tubes
showing
gas

Percentage
of gfts
tubes con-
firmed

Rapid (10 percent or more
in twenty-four hours)

684

97.7

57

44.0

Moderate (less than 10 per-
cent in twenty-four hours)

193

91.2

15

20.0

Slow (no gas in 24 hours;
10 percent or more in
forty-eight hours)

276

73.2

156

6.4

Very slow (no gas in
twenty-four hours ; less
than 10 percent in forty-
eight hours)

33

45.5

26

0.0

The results indicated are based on a study of 1559 water analyses
in the Advanced Sector of the American Expeditionary Forces at Dijon,
France, comprising waters from both treated and raw supplies and from
various sources. An interesting correlation is noted between the rate of
gas production and the probability of confirmation. It appears evident
that

1. The positive presumptive test (10 percent or more gas in
24 hours) is a very reliable index of the probable presence of Bact. coli

60

when dealing with untreated waters but it is not to depended upon when
testing chlorinated waters; 97.7 per cent of the former and only 44 per cent
of the chlorinated samples showing gas, were confirmed for Bact. coli
or its close allies.

2. The doubtful presumptive test (less than 10 percent gas in
24 hours, or more than this quantity in 48 hours where 24 hours was nega-
tive) is only a fair index of the* probable presence of Bact. coli in un-
treated waters, while for chlorinated specimens it is practically negligible,
as 75.9 percent of gas tubes from untreated, and only 7.6 percent of those
from chlorinated samples were successfully confirmed.

3. A small amount of gas in 24 hours is a more reliable index for
Bact. coli than 10 percent or more gas in 48 hours. The rate of gas pro-
duction seeems more significant than the total volume of gas formed as a
presumptive test.

Similar results were obtained by Graff and Mote. Of 966 instances
of gas production in lactose peptone bile from settling basins (unchlorin-
ated) 41 percent were positive. Of 844 instances from reservoirs (after
chlorination) 12 percent were confirmed, and 1063 gas tubes from taps
only 5 percent were confirmed. They also noted that vigorous gas pro-
duction in 24 hours yielded positive tests for Bact. coli in a greater percent
of cases, from all sources, than was obtained if gas was not noted previous-
ly to the 48 hour period.

It is thus apparent that gas production in either lactose broth or
lactose bile, particularly with treated, partially purified, or chlorinated
waters, cannot be relied upon as an index of the presence of the colon
groujj. Confirmation must be restored to, except possibly for water
known to be polluted.

CONFIRMATORY TESTS

it is merely necessary to streak out or smear on the surface of a differential
medium (litmus lactose agar, Endo agar, or Eosin-Methylene-Blue agar)
which is then included at 37° C. for 24 hours. If typical colonies develop,
the presence of the colon group in the presumptive test is considered con-
firmed. This is known as the partially confirmed test and for routine
work is sufficient, although it must be borne in mind that there are a few
spore forming lactose fermenters which are capable of growing aerobically.

If the colonies developing on this differential medium are not typical
it becomes necessary to further study these colonies to determine that they
are Gram negative non-spore formers and that they are capable of pro-
ducing gas from lactose. The latter may be determined by merely fish-
ing a colony into lactose broth which is incubated for 24 hours. If gas
is formed and if the Gram stain showed the colony to be a non-spore form-
ing Gram negative short rod, the test constitutes what is spoken of in the
Standard Methods as the completed test for the colon group.

If no growth at all develops on the confirmatory differential medium
the assumption is that the gas in the presumptive test tube was due to
strict anaerobes and the colon group is recorded as absent. It is con-

61

ceivable, of course, that such a test may be due to death of the colon
types. A transfer from this unconfirmed presumptive test to another tube
of lactose broth, which on incubation shows gas production, would be
excellent evidence that the gas forming organism in question was an an-
aerobe and that the negative confirmatory test was not due to death or over-
growth of the gas producer.

The media for confirmatory tests which have found most favor among
American bacteriologists are litmus lactose agar, and Endo agar; and
more recently the simplified Eosin-Methylene-Blue agar was suggested by
the author. The litmus lactose agar has been discussed in detail in con-
nection with the isolation of the colon group. It will merely be added
here that Meyer suggests that by the addition of 3 percent agar, giving a
stiffer and dryer medium, more favorable results are obtained than with the
ordinary medium.

The Endo Agar. This is perhaps the most extensively employed
confirmatory medium at the present time. It is the medium of choice of
the Public Health Service and is recommended in the Standard Methods
(1917 and 1920). The colon colonies produce a distinct red coloration,
often with a metallic sheen, and are quite easy to detect. Unfortunately
there has been, in the past, considerable lack of uniformity in the prepara-
tion of Endo. The method of adjustment of the reaction is very crude
and there was lack of agreement as to the quantity of indicator to be em-
ployed. Thus Kendall, and Prescott and Winslow recommended 0.1 c. c.
of a saturated basic fuchsin and 1 c. c. of a 10 percent sodium sulphite
per 100 c. c. of agar. The Army Laboratory Manual, employed exten-
sively during the war, recommends 0.18 c. c. of fuchsin and 2.5 c. c.
of a 10 percent sodium sulphite. In the Hygienic Laboratory Endo medium
the indicator consists of 0.5 c. c. of basic fuchsin and the equivalent of 2.5
c. c. of a 10 percent sodium sulphite solution. Some investigators, Rob-
inson and Rettger, have recommended decolorization with sodium acid
sulphite in place of sodium sulphite.

The appearance of the colon colonies will depend to a considerable
extent on the concentration of the indicator, being very intense red and
showing the characteristic metallic sheen only with the higher concentra-
tions. The description of a typical colony of Pact, coli on one of these
mediums may therefore not tally at all with its appearance on supposedly
the same medium prepared in some other laboratory. It thus becomes
essential to detail, in each instance, the method of preparation of the
medium employed. We may well exclaim with Morse and Wolman, "What
shall the standard test for Bact. coli include if a mere difference in the
proportion of fuchsin and sulphite in two Endo media results in the one
showing typical colonies in 75 percent of the tubes, in which Bact. coli is
present, while the other shows the same colonies in only 14 percent?"
The apparent difference is due to a confusion of the term "typical" Bact.
coli colony. What is considered typical for one medium with one con-
centration of dye will of course not be typical for another.

62

One of the great difficulties in the preparation of Endo medium is
the adjustment of the reaction. In most of the methods the reaction is
adjusted on the phenol phthalein scale to some definite point and then a
quantity of sodium carbonate is added to make it more alkaline. This
is a rather inaccurate method and results in some batches being very ser-
viceable while others are almost worthless.

Levine suggested a modified and simplified Endo medium which re-
quires no adjustment of reaction and which need not be filtered. The
medium consists of 1 percent Difco peptone, 0.3 percent dipotassium phos-
phate, 0.5 percent agar, and 1 percent lactose. One-half c. c. of a satur-
ated basic fuchsin, decolorized by 2.5 c. c. of a 10 percent sodium sulphite,
as recommended by the Hygienic Laboratory, is employed as an indicator
for each 100 c. c. of the medium. Aside from the simplicity of prepar-
ation, an advantage claimed is that Bact. coli may be differentiated from
Bact. aerogenes. The former possess a distinct metallic sheen, the colonies
are flat and button-like, and about two or three m. m. in diameter; whereas
the latter usually produces considerably larger colonies which are convex
and a metallic sheen is rarely observed. The disadvantage of the medium
is that diffusion of color, due to acid production, is very rapid. This
may be reduced by increasing the content of agar but when that is done
the differentiation between Bact. coli and Bact. aerogenes becomes less dis-
tinct.

All of the Endo mediums above have the disadvantage of instability.
Exposure to light or air induces a deep red coloration which interferes
seriously with the detection of acid formers thus making it necessary to
prepare the medium fresh and at frequent intervals.

Kahn suggests the use of Endo medium in tubes instead of plates
inoculating from the presumtive test onto the slant and also into the butt
in a manner analagous to the Russel double sugar medium. He states that
Endo agar prepared and sterilized in tubes may be kept from three to four
weeks without deterioration and that anaerobes will not develop in the
butt of the Endo agar tube. Thus the test for acid production under aero-
bic and gas formation under anaerobic conditions by members of the colon
group is determined simultaneously thereby performing both partially
confirmed and a complete test, described above in a single tube. This
use of the Endo medium is very interesting and suggestive and worth trial
and consideration.

The Eosin Methylene Blue Agar Confirmatory Medium. Le-
vine, in 1917, suggested a modification of the eosin-methylene blue agar
(first described by Holt, Harris and Teague) for the confirmation of the
presumptive test. The medium is prepared in the following manner:

Distilled water 1000 c. c.

Peptone (Difco) 10 gm.

Dipotassium phosphate 2 gm.

Agar 15 gm.

63

Boil ingredients until dissolved and make up any loss due to evapora-
tion. Place measured quantities in flasks and sterilize at 15 pounds for
15 minutes.

Just prior to use, add to each 100 c. c. of the melted agar, prepared
as above, the following constituents:

Sterile (20%) lactose solution 1 gm. or 5 c. c.

Aqueous (2.0%) eosin (yellowish) solution.. ..2 c. c.
Aqueous (0.5%) methylene blue solution 2 c. c.

Pour medium into petri dishes, allow, it to harden in incubator and
inoculate in the ordinary way. Smearing the surface with a glass rod
seems preferable to the streaking method sometimes employed.

There is no adjustment of reaction and nitration of medium is not
necessary.

Test tubes may be substituted for petri dishes if desired. The value
of such a change is (1) the reduction of expense, as only 3 or 4 c. c. of
medium is needed for every test tube while about 15 c. c. is usually em-
ployed with petri dish, and (2) test tubes may be stored for long periods
whereas the medium in petri dishes would have to be prepared at intervals
of a week or less.

TABLE XXVII. DIFFERENTIATION OF BACT. COLI AND BACT. AEROGENES
ON EOSIN-METHYLENE BLUE AGAR.

Size

Bact. coli (1)

Well isolated colonies are 2 — 3
m.m. in diameter.

Bact. aerogeneS (2)

Well isolated colonies are larger
than coli; usually 4-6 m.m. in
diameter or more.

Confluence

Neighboring colonies show little
tendency to run together.

Neighboring colonies (2) run to-
gether quickly.

Elevation

Colonies slightly raised; surface
flat or slightly concave, rarely
convex.

Colonies considerably raised and
markedly convex; occasionally
the center drops precipitately.

Appearance by trans-
mitted light

Dark almost black centers which
extend more than % across
the diameter of colony; inter-
nal structure of central dark
portion difficult to discern.

Centers deep brown; not as dark
as Bact. coli and smaller in
proportion to the rest of the
colony. Striated internal
structure often observed in
young colonies.

Appearance by re-
flected light

Colonies dark, button-like, often
concentrically ringed with a
greenish metallic sheen.

Much lighter than Bact. coli.
Metallic sheen not observed
except occasionally in de-
pressed center when such is
present.

(1) Two other types have been occasionally encountered: One resembles the type de-
scribed, except that there is no metallic sheen, the colonies being wine colored. The other
type of colony is somewhat larger (4 m. m.) grows effusly, and has a marked crenated or
irregular edge, the central portion showing a very distinct metallic sheen. These two
varieties constitute about 2 or 3 percent of the colonies observed.

(2) A small type of aerogenes colony about the size of the colon colonies, which
shows no tendency to coalesce, has been occasionally encountered.

64

The advantages claimed for the eosine-methylene-blue medium are (1)
ease of preparation, as neither filtration nor adjustment of the reaction
is required, and (2) relative permanency. (Petri dishes poured have
been used after a week or longer, if stored in the ice box) and (3) the
medium affords an excellent differentiation between Bact. coli and Eact.
aerogenes, as described in Table XXVII above.

The method of preparation of the test tubes is briefly as follows :
The medium prepared as above is poured aseptically into sterile test
tubes and allowed to solidify so as to give a long slant without a butt.
To confirm the presumptive test, inoculation is made on the surface of
these slants with a straight platinum needle. If desired the medium may
be placed in test tubes and sterilized in the Arnold or in the auoctlave
in the ordinary manner. In that case it will be found that the agar, when
first removed from the sterilizer, is fluorescent like eosin but as it cools
the typical wine color returns.

The reliability of the differentiation of Bact. coli and Bact. aerogenes
is indicated in the following table:

TABLE XXVIII. RELIABILITY OF PRESUMPTVE DIFFERENTIATION OF
BACT. COLI AND BACT. AEROGENES ON EOSIN-METHYLENE BLUE AGAR.

Designation
from appearance
on agar

1918-1919, France

1917, Iowa

Number of
colonies tested

Percentage
correct

Number of
colonies tested

Percentage
correct

Bact. coli

87

94.2

102

96.9

Bact. aerogenes

55

85.5

122

82.4

Thus in 1917, of 122 colonies fished, as probably Bact. coli, 96.9 per-
cent were proved to be such and in 1918 and 1919, 94.2 percent of 87
colonies fished were correctly designated from their appearance on eosin-
methylene blue agar. The proportion of Bact. aerogenes, correctly indenti-
fied was 82.4 percent and 85.5 percent in the series in Iowa and France
respectively. The medium therefore appears to be quite reliable for the
routine presumptive differentiation of these two groups of colon bacteria
and may therefore be of considerable value if in the future it is consid-
ered desirable to distinguish these forms because of their probable differ-
ent sanitary significance.

65
V. THE COLON GROUP AS AN INDEX OF POLLUTION

Safe Water. A safe water for human consumption may be defined
as one which is free from harmful constituents important among which
are disease producing microorganisms. The logical and most direct pro-
cedure to determine the potability and safety of a water would be to
determine the presence or absence of pathogenic bacteria but unfortun-
ately this task is an impossible one for routine and recourse must there-
fore be taken to an indirect index of the probable presence of harmful
germs. Since the diseases transmissable through water are primarily of
intestinal origin the detection of the presence of intestinal material natur-
ally leads to the presumption that a potential danger exists, for if such
material is present it is very probable and certainly possible that intes-
tinal disease germs are also present.

A number of tests both chemical and bacteriological have been sug-
gested as indicators of intestinal pollution. The bacterial examination,
by reason of the large number of bacteria present in feces and sewage and
the ease with which they may be detected in water, is a particularly delicate
test. Three groups of bacteria have been regarded as indicators of pol-
lution:

The colon group.

Sewage streptococci.

Spore forming anaerobes.

An organism to be considered an ideab index of fecal pollution should
have the following characteristics:

1. It should be distinctively and characteristically of human or
animal intestinal origin.

2. It should be absent or extremely rare in nature outside of the
intestinal tract.

3. It must be capable of easy and rapid detection.

4. Its incidence in water should bear some constant relation to the
sanitary survey or our knowledge as to the probability of pollution, par-
ticularly with sewage.

5. It should be distinctly more viable and more resistant in water
and to treatment than are the intestinal pathogens (Bact. typhi, Bact. dysen-
teriae, etc.), but not excessively so.

Such an ideal index is not available but the general concensus of opin-
ion among English and American bacteriologists is favorable to the em-
ployment of the Colon group for this purpose. Although bacteria of this
group are not restricted in habitat to the intestinal tract of man being
characteristic also of the intestinal tract of the lower animals, it is never-
theless true that there is a correlation between the quantitative incidence
of a least the coli section and known pollution. The whole group is easy
of detection as will be seen from the following considerations. It is more
viable than Bact. typhi but yet dies off relatively quickly; colon bacilli
are present in relatively large numbers in water known to be polluted but
only infrequently in natural supplies. A correlation has been established

66

between the incidence of the colon group in drinking water and the ty-
phoid fever rate in the community.

Viability of Bact. coli and Bad. typhi in Water. These organ-
isms do not find favorable conditions for growth in natural waters. The
food supply is apparently insufficient or unavailable, the temperature is
unfavorable, other organisms exert a deleterious effect, and in consequence
a diminution in number rapidly ensues. It is conceivable that under cer-
tain conditions of extremely high temperature, as may occur in summer,
there may be an initial increase but death soon results. A convenient
method for appraising the relative viability of two organisms is to com-
pare the rates at which they die off under a given set of conditions.

The rate of death of bacteria may be expressed by

1 ' N

K z 77^0 lo§ T where;

"N" is the number of organisms at the beginning, i. e. time "To".

"n" is the number of organisms at the end of the time, "T^'.

"K" is a measure of the rate of death.

If the logarithms of the number of surviving bacteria at different
periods are plotted against the elapsed time it will be observed that the
points fall along a straight line intersecting the "x" axis at an angle whose
tangent is numerically equivalent to "K". The greater the rate of death
the greater will be "K". In the following tables are given the number
of surviving Bact. coli and Bact. typhi when stored in water and exposed
to the air. It will be observed that the rate of death of Bact. typhi is
considerably higher than Bact. coli, though the latter organism also dies
off rather rapidly.

TABLE XXIX. VIABILITY OF BACT.

COLI IN WATER AT 2QO C.

(After Hinds 1916)

Elapsed time
in hours

Count

Rate of
death "K"

0

690,000

2

600,000

.069

6

420,000

.083

13

244,000

.079

24

123,000

.071

36

10,000

.1181

48

3,400

.111

72

120

.120

TABLE XXX. VIABILITY OF BAC1
TYPHI IN WATER AT 2QO C.
(After Hinds 1916)

Elapsed time
in hours

Count

Rate of
death "K"

0

118,000

1.38

2

7,500

1.38

4

500

1.37

6

280

1.01

8

230

.78

10

150

.67

12

30

.69

14

30

.59

24

2

.46

Average "K" 0.094

Average "K" 0.87

The writer made some observations at the Massachusetts Institute of
Technology in 1913 on the effect of temperature on the viability of Bact.
coli in distilled and conductivity water containing C. P. salt. The results
are shown in figures 3 and 4 where it is apparent that the higher the tem-
perature the greater the rate of death, and that the presence of gases in
solution also affects the death rate.

67

Houston artificially infected Thames River water with pure cultures
of Bact. typhi and noted the proportion of survivors are infleuced by the
temperature of storage. The results, based on ten experiments at each
temperature, are summarized below:

TABLE XXXI. EFFECT OF TEMPERATURE ON VIABILITY OF BACT. TYPHI

IN THAMES RIVER.
(After Houston 1913)

Temperature of storage

Surviving Bact. typhi per c. c.
after 1 week*

Bact. typhi absent from 1
c. c. after

GO C.

47,766

9 weeks

5o C.

14,894

7 weeks

10o C.

69

5 weeks

18o C.

39

4 weeks

2?o C.

19

3 weeks

37o C.

5

2 weeks

'Initial number 103,328 per c. c.

$ •'

c

V

j__

c

^"~X° *

~~~ — -

o

— ?

Vj

\

• ~~-~

--7-

^_

X

\

SB

~~ ^

fc

\^ **

>

N?

\

i

'/ability

of dac

T eo/t

1

in D/s.

<-il led W

vfer

s

cont

ct//7Jntj ,

1.85 ft A

ad (C

#)

/CO ZOO

7~/me E/apsed /n

68

"••**.

Viability of fiact.coh
in Conductivity Water

(///£ a/**<T<3 free)
containing 0.05% CP NaCI

/oo

2OC
7/mc. e/a/osed. /n

3OO

+00

That there is a marked correlation between the incidence of the colon
group and the probability of pollution may be clearly demonstrated from
Houston's observations on brooks and rivers which are more or less sub-
ject to pollution, and on two lakes, one of which, Loch Loggan, was re-
ceiving some drainage while the other, Loch Ericht, was free from con-
tamination of human or domestic animals. It will be observed that in
brooks and rivers, colon bacilli were usually present (61.5%) in small
portions of water (1 c. c. or less); whereas in the unpolluted, Loch Ericht,
the colon group was only occasionally encountered even in 100 c. c. samples.

TABLE XXXII. ON THE INCIDENCE OF COLON GROUP AND PROBABILITY
OF POLLUTION IN SURFACE WATERS.

(After Houston 1906)

Source

Contamination

% of Samples showing Bact. coli in

+0.1 c. c.

+ 1.0 c. c.
—0.1 c. c.

+10 c. c.
-1.0 c. c.

+ 100 c.c.
—10 c.c.

Absent
from 100
c. c.

Brooks & rivers

More or less polluted

7.7

53.8

34.6

3.8

Loch Loggan

Receives some farm
land drainage

1.2

33.0 49.4

16.4

Loch Ericht

Free from human &
animal pollution

1.0

19.0

80.0

Gumming in his report on the sanitary conditions of the Potomac
River watershed found a striking correlation between the proximity of
pollution and the incidence of colon bacilli.

69

TABLE XXXIII. ON THE RELATION OF THE NUMBER OF COLON BACILLI
AND PROXIMITY OF POLLUTION IN SURFACE WATER (POTOMAC RIVER).

(After Gumming 1916)

Sampling
Point

Distance from
source of pollution

Bacilli of colon group per c. c.
monthly average

Just below city

24 to 617

At Marvel Hall

14 miles

24 to 137

Maryland Point

42 miles

.04 to 6.9

Oyster region

64 to 103 miles

.01 to .04

The high incidence of the colon group in the vicinity of sewer out-
lets, its elimination as the distance from the source of pollution increases,
and its rarity at points remote from contamination is clearly shown.

In connection with the purification of water, some correlation has
been observed between the incidence of colon bacilli in the filtered water
and the typhoid fever rate in the community. Thus in Lawrence, Massa-
chusetts, Clark and Gage report the following incident, which occurred in
1898. In relaying some under drains of a filter the sand on the beds was
seriously disturbed resulting in a marked increase in tests for colon bacilli
in the filter effluent. This was followed by a rapid rise in the number
of cases of typhoid fever, and as the colon content fell to normal, the cases
of typhoid fever disappeared.

TABLE XXXIV. SHOWING RELATION OF COLON INCIDENCE TO TYPHIOD
MORBIDITY RATE, LAWRENCE, MASS., 1898-9.

(After Prescott & Winslow, 1915)

Percent of positive colon
tests in 1 c. c. samples

Number of cases of
typhoid fever

December, 1898

72

12

January, 1899

54

59

February, 1899

62

12

March, 1899

8

9
(all during the early part of
the month)

Caird, in a report on twenty years' experience with the filter plant at
Elmira, New York, also notes that there is some correlation between the
percent of positive colon tests and the typhoid fever rate.

Opinions as to the Value of the Colon Test. It appears, from
the foregoing discussion, that the colon group possesses many of the at-
tributes of an excellent indicator of pollution. Unfortunately, however,
it is not characteristic and distinctive of the human intestine. Some of its
members have been isolated from grain and other sources which are far
removed from human or even animal contamination. This has lead to
disagreement as to the value of the colon test in water analysis.

70

TABLE XXXV. INCIDENCE OF COLON GROUP AND TYPHOID DEATH RATE
IN FILTERED WATER, ELMIRA, N. Y. (After Caird 1919).

Year

% Samples showing colon group

Typhoid rate for 100,000

1903
1904
1905

20.00
13.33
1.37

64.5
53.3
25.2

1906
1907
1900

18.71
4.17
6.04

39.2
25.2
30.8

1909
1910
1911

5.47
0.00
0.00

33.6
26.4
13.2

1912
1913
1914

1.81
0.46
0.00

15.9
10.6
10.6

1915
1916
1917

0.17
0.35
1.43

9.6
9.6
11.9

phoi<
Incite nc

i Death f

and

Color

Gro

hlte
Eln

ed
ra 1

Ion Icemen

ite
:e

\

1903

I9O5

'07 '08

/9IO

1915 '/t '17

The English and American bacteriologists uphold it. The Germans
have objected strenuously to its inclusion in water work. The contention
of Kruse, Konrich and other German investigators that colon bacilli may
be found in any water irrespective of its source, provided a sufficiently
large sample is taken, and that of Chick, Houston, and other English author-
ities, that these organisms are not present in water unless it has received

71

sewage pollution, led Prescott and Winslow to conclude that it was the
number rather that the presence of colon bacilli that should be used
as a criterion for recent sewage pollution. Thus they state that the rind-
ing of a few colon bacilli in large samples of water or their occasional
isolation from small samples is not of any special significance but if these
organisms are present in a large proporotion of small samples (I c. c. or
less) then the evidence of recent sewage pollution is significant.

Gartner (1910) estimates the number of colon forms in cultivated
soil at 1,500,000 per square meter and naively remarks that it is no wonder
the rain should wash a few of them into neighboring wells. The recent
work of Johnson and Levine, Burton and Rettger, Chen and Rettger and
others have shown quite conclusively that the organisms in the soil are
markedly different from those habitually present in the human and animal
intestine. Furthermore the "B. coli" of Houston and other English in-
vestigators does not include many of the soil forms for they regard only
lactose positive-indol positive varieties as "typical" B. coli, whereas most
of the Bact. aerogenes and Bact. cloacae are indol negative. German
opinion is gradually becoming more favorable toward acceptance of the
colon test, provided that preliminary enrichment is carried out at 46° C.
(Eijkman test). This procedure, as has been pointed out before, serves
to eliminate the Bact. aerogenes and Bact. cloacae or characteristic soil
types. It appears that the contentions of the German school were really
against these soil forms as indices of pollution. It thus becomes possible
to reconcile their opinions with those of English and American bacteriol-
ogists if we recognize that not only is the numerical incidence of the colon
group significant as an index of pollution, but that the type of organism
must also be considered.

THE COLI AND AEROGENES SECTIONS AS INDICES
OF POLLUTION

The question arises as to whether the aerogenes and coli sections are to
be accorded the same sanitary significance. Considerable work is still needed
along these lines, but the following is presented as to (1) their distribu-
tion in nature, (2) correlation of their relative incidence with the sanitary
qualities of water, (3) relative viability in water, and (4) relative resis-
tance to treatment and chlorination.

Incidence of the Coli and Aerogenes Sections in Feces. Rogers,
Clark, and Evans called attention to the scarcity of the high ratio group
(aerogenes section) in cow dung. Only one colony of 150 fished was of
this type. Levine observed that among 117 cultures isolated from cow,
horse, sheep, pig, and man, not a single organism proved to be Bact. aero-
genes. A study of the literature employing the \7oges-Proskauer test as a
means of differentiation of the two colon sections disclosed that these ob-
servations as to the scarcity of the aerogenes types in human and animal
feces has long been recognized although not adequately appreciated.

Thus MacConkey (1905) remarks on the scarcity of Bact. (lactis)
aerogenes in human feces. In the examination of 241 strains obtained from

72

23 hunjan stools, only 4 gave the Voges-Proskauer reaction and of these
3 were obtained from a single sample. At the same time, a study of 51
strains from 5 samples of horses and 48 from 6 samples of cow manure
did not yield a single aerogenes culture. In 1909 he made further ob-
servations with similar results.

Ferriera, Horta and Paredes reported 8 out of 117 lactose fermenting
strains from human feces, and 2 of 81 strains obtained from 46 different
species of animals to belong to the aerogenes section as indicated by a
positive Proskauer test.

The work of Clemesha is particularly significant because of the large
number of cultures examined. Of 1207 human strains, about 6 percent
were found to be of the aerogenes type but a large number of these came
from a single sample. Among 1029 cultures from cow dung, Bad. (lactis)
aerogenes was encountered in about 10 percent. He records that in cow
dung, the Bact. aerogenes was found in small numbers, and Bact. cloacae
was sometimes very common, but that in human stools these types were
very rare and that a sudden increase in their prevalence was never ob-
served. Clemesha's observations were confirmed to a considerable extent
by R. G. Archibald of the Wellcome Tropical Research Laboratory in an
investigation of the water of Khartoum.

TABLE XXXVI. INCIDENCE OF AEROGENES TYPES AMONG COLON BACILLI

IN HUMAN DEJECTA.

Investigators

No. of strains
studied

% Aerogenes Section
(V. P._U:M. R )

MacConkey (1905)

241

1.7

Ferriera, Horta and Paredes (1908)

178

5.6

MacConkey (1909)

117

6.8

Archibald (1911)*

100

0.0

Clemesha (1912)

1207

4.6

Levine (1916)

25

0.0

Hulton (1916)

10

0.0

Rogers. Clark and Lubs (1916)

113

5.8

Rogers, Clark and Lubs (1918)

177

t23.0

Darling (1919)

20

0.0

Wood (1919)

33

0.0

Stokes (1919)

141

§16.3

Chen and Rettger (1920)$

173

0.0

Total

2534

5.9

*Includes strains obtained from latrines, directly from feces, and fecal suspensions
exposed to action of sunlight.

fThese comprised 46 strains isolated for the most part by special methods and 31 of
these strains were obtained from a single specimen. The true incidence of aerogenes types
is therefore considerably less than shown in the table.

§A11 Bact. cloacae.

^Includes animal strains.

73

Recent observations by Rogers and his associates, Darling, Chen and
Rettger, Stokes, and others are entirely in accord with the opinions recorded
that the aerogenes section is rare in feces of man and animals.

The relative presence of the coli and the 'aerogenes sections in human
and various animal feces is shown in Tables XXXVI. and XXXVII.

TABLE XXXVII. INCIDENCE OF AEROGENES TYPES AMONG COLON BACILLI

IN ANIMAL DEJECTA.

Investigators

Animal
source

No. of strains
studied

% Aerogenes
Section
(V.P.4_;M.R_)

MacConkey (1905)

Horse & cow

99

1.0

Ferriera, Horta & Paredes (1908)

Misc.

81

2.5

MacConkey (1909)

Horse

67

4.5

MacConkey (1909)

Misc.

87

0.0

Archibald (1911)

Cow & goat*

20

40.0

Clemesha (1912)

Cow

1029

10.7

Rogers, Clark & Evans (1914)

Cow

150

0.7

Levine (1916)

Misc.

92

0.0

Darling (1919)

Misc.

93

0.0

Wood (1919)

Misc.

99

7.0

Stokes (1919)

15

13.0

Total

1832

7.4

* Single samples from each species.

Incidence of the Coli and Aerogenes Sections on Grains and
in Soil. Many investigators have reported the presence of the colon
group on grains and plants. Prescott found lactose -fermenting bacteria
in flour, bran, corn meal, oats, and barley. Similar observations were
made by Papasotiriu in Germany. Cline and Houston found, what they
regarded typical colon baccili, in 3 out of 24 samples of wheat, oats, rice,
etc. Bettencourt and Borges (1908) succeeded in isolating such organ-
isms from vegetables and cereals. Konrich (1910) frequently observed
the colon group on cultivated plants and that even about six percent of
plants obtained from waste places showed colon bacilli. Fifty-five percent
of 300 samples of grain showed members of the colon group.

These investigators, however, did not differentiate between the Bact.
coli and Bact. aerogenes. Rogers, Clark, and Evans were the first to point
out that the colon-like forms, occurring on grains, were distinctly dif-
ferent from those observed in feces. They isolated 166 cultures from
33 samples of dried grains and 2 samples of green oats. Of these only
8 (4.8%) were the low ratio (coli section); (91.0%) were of the

74

high ratio type (aerogenes section), and 7 (4.2%) produced only hydrogen.
They noted further that none of the types found on grains were identical
with those obtained by them from bovine feces, the low ratio types from
grains being pigmented.

Stokes isolated 77 strains from corn meal, grape nuts, post toasties,
corn flakes, puffed wheat, quaker oats, rye, corn, wheat, barley, and oats,
of which 57 (74.0%) were of the aerogenes cloacae type. It is appar-
ent therefore that on grains and plant products, the high ratio V. P.
positive, methyl-red negative group predominates.

Houston, in 1897 and 1898. examined a number of samples of orchard,
garden pasture, and virgin soil, including polluted and unpolluted areas,
concludes that the true colon bacilli are very rarely or never found in virgin
soils, while they are present in large numbers in other areas, especially
those which have been contaminated from animal sources.

Konrich, in a study of 547 specimens of soils in Germany, detected
the colon group in 65 percent of 0.1 to 0.5 gram samples. He notes that

TABL,E XXXVIII. SOURCE, TREATMENT, ETC., OF 42 SAMPLES OF SOIL

EXAMINED IN 1915.

•- —

_£ D

^ c

Is

11

*s £

• . .

Sample

Source

Soil treatment or other remarks

O ^

Crop

>?
O c/>

Q ~

^'1

P 03

g bD

Z O

I

II

III

IV

VI
VII

VIII
IX

X

XI

XII

XIII

XIV

Corn field...
Corn field...
Corn field...
Corn field...
Corn field...
Corn field...
Expt. plot
Expt. plot
Expt. plot
Expt. plot

Manured, corn grown for three years

Manured, corn grown for three years

Manured, corn grown for three years

Manured, corn grown for three years

Manured, corn grown for three years

Manured, corn grown for three years

tons manure annually

114
113
112
111

Expt. plot 110

1913 |Corn

0

1913

Corn

14

1913 ICorn

4

1913

Corn

3

1913

Corn

2

1913

Corn

9

1914 (None

2 tons manure annually 1914

1 ton manure annually 1914

tons of clover chopped and ploughed
under annually |

2 tons of clover chopped and ploughed

1914

None
None

None

under annually

Expt. plot 109 1 ton of clover as in IV
Expt. plot 108 !2 tons of oat straw chopped and ploughed

under annually

Expt. plot 107 No treatment, check plot |

Expt. plot 106 2 tons of timothy chopped and ploughed

under annually

Expt. plot 105 jl ton timothy as in IX |

Expt. plot 104 18 tons of clover once in four years j

Expt. plot 103 8 tons of manure once in four years

Expt. plot 102 128 tons peat annually ,

Expt. plot 101 [Timothy grown each year and stubble
' plowed under

1914 |None
1914 JNone

|None
1914 |None

I

1914 |None
1914 |None
1913 |None

1913 INone

1914 (None

1914 (Timothy

18
0

0
1

0
1

26
24
14

18

,'r

75
TABLE XXXVIII. (continued)

Sample

Source

,
Soil treatment or other remarks

Date of last
treatment

Crop

Number of coli-like
organisms obtained

XV
XVI

XVII
XVIII
XIX
XX

XXI
XXII

XXIII
XXIV

XXV
XXVI
XXVII

XXVIII
XXIX
XXX

XXXI

XXXII
XXXIII
XXXIV
XXXV
XXXVI

Orchard
Orchard

Orchard

Clover sod

1913
1913
1913

1913
1913
1913

1913

1913
1913
1913
1913
1913
1913

| Clover

?

Blue grass
Clover

[Corn
[Corn

(Corn
|Corn
[Corn

Corn
Corn
Corn

Corn

Corn
Timothy
Clover
Corn
Corn
I Kafir corn

I 12

1 0
0
8
1

1
8

5

13

18
15
14

16
16
13

19

12
10
13
17
15
I 2

Clover crop planted in summer and
ploughed under in spring

Clean cultivation

Orchard
Sioux loam
Wabash silty
clay
Marshal silt
loam

Blue grass sod

Clover field recently in oats

Surface soil

Surface soil

Corrington
loam

Surface soil; corn field

Expt. plot 213
Expt. plot 214

Expt. plot 215
Expt. plot 216
Expt. plot 217

Expt. plot 218
Expt. plot 219
Expt. plot 220

Expt. plot 221
See XIV.

No treatment, check plot

Legume treatment, cow peas in July. 2
years out of four in rotation with corn..
8 tons manure

8 tons manure with cow peas

8 tons manure with 800 pounds bone
meal and cow peas

800 pounds bone meal and cow peas
8 tons manure and 800 pounds bone meal
800 pounds bone meal and 200 pounds K
in KC1 and cow peas

8 tons manure, 800 pounds bone meal and
200 pounds K in KC1

See XIV ...

Clover field
Corn field
Corn field
Corn field

Unknown

Same as A

Same as A

Unknown

the farther away the source was from cultivation, the smaller was the
proportion of positive results, but that they were never altogether absent.
Unfortunately, neither Houston nor Konrich differentiated between the aero-
genes and coli sections.

Johnson and Levine first pointed out that the coli-like microorganisms
of the soil were strikingly similar to those obtained by Rogers and his as-
sociates from grains and that they were quite distinct from the true Bact.
coli of feces.

The observations of Johnson and Levine are based upon 42 samples
of soil obtained from experimental plots, (data on treatment of which is
available for many years) from orchards, corn fields, and miscellaneous
sources in different parts of the state of Iowa.

The method of isolation of cultures was by planting on litmus lactose
agar directly or after preliminary enrichment in lactose broth, incubation

76

being at the body temperature. In the following table is given the source
and treatment of the soil, crops raised, and the incidence of coli-like bacteria
in the samples studied.*

It will be noticed that a number of plots which had not received
manure for many years (some about 15 years) yielded organisms of this
group. There was a marked correlation also between cropping and the
incidence of the colon group.

TABLE XXXIX. PRESENCE OF BACILLI OF COLON GROUP IN EXPERI-
MENTAL SOIL PLOTS AT AMES, IOWA

Coli-like organisms

Fallow plots

Cropped plots

Number

Percent

Number

Percent

Absent

7
2
4

53.8
15.4
30.8

0
0
11

0.0
0.0
100.0

Rare

Abundant

Of the 24 plots examined, 13 had been kept fallow and 11 had had
various crops, for the most part corn, raised upon them. Of the fallow
plots 7 (53.8%) showed no colon forms, in 2 (15.4%) such organisms
were rare, and relatively abundant only in 4 (30.8%). On the other
hand, bacilli of the colon group were readily isolated from each of the 11
plots upon which crops were grown.

One hundred and seventy-seven strains were obtained in pure cul-
ture. Of these 142 (80.4% ) are of the aerogenes section (the predominating
type being the Bact. cloacae) and 35 (19.6%) are of the coli section.

The preponderence of the aerogenes types in soil has recently been
confirmed in two very detailed studies by Burton and Rettger (1917) and
Chen and Rettger (1920).

Burton and Rettger isolated 193 coli-like bacteria from about 1000
samples of soil, twigs, leaves, and flowers.

Of these only 36 (18.7%) resembled Bact. coli whereas 157 (81.3%)
were of the aerogenes subgroup, 148 of the latter being Bact. cloacae.

Chen and Rettger in a study of a large number of samples of soil
of known sanitary quality observed also a great preponderence of the aero-
genes section but they found the Bact. aerogenes rather than the Bact.
cloacae, as recorded by Johnson and Levine and by Burton and Rettger,
to be the predominating species. Of 467 strains studied, 430 (92.1%)
were Bact. aerogenes, 17 (3.6%) Bact. cloacae, and only 20 (4.3%) were
of the coli section (V. P. negative-methyl-red-positive type).

The evidence is quite distinct and clear that the colon-like organisms
present in soil are of the aerogenes type and that they are more abundant in
cropped than in fallow areas. The relative incidence of the aerogenes and
coli sections in soil and grains is summarized in Table XL.

77

TABLE XL. INCIDENCE OF AEROGENES TYPES AMONG COLON BACILLI
ISOLATED FROM SOIL AND GRAINS.

Investigator

Source

No. of strains
studied

% Aerogenes
section
(V.P.4- M.R )

Levine & Johnson (1917)

Soil

177

80.4

Burton & Rettger (1917)

Soil

193

81.3

MacConkey (1909)

Soil

16

37.5

Chen & Rettger (1920)

Soil

467

95.7

Total

853

88.1

MacConkey (1909)

Grains

30

56.7

Rogers, Clark & Evans (1915) *

Grains

166 1 91.0

Wood (1919)

Grains and
Cereals

IS

73.5

Stokes (1919)

Grains

77

74.0

Total

288

81.7

*Differentiation based on gas ratio.

Incidence of the Coli and Aerogenes Sections in Water and
Milk. There is, at present, comparatively little information available as
to the relative abundance, of the coli and aerogenes subgroups in various
types of water. Rogers, in a study of 137 strains isolated from a stream,
found 66 percent were of the high ratio and aerogenes-cloacae and 33 percent
of the low ratio coli type. Greenfield and Skorup found the ratio just
the reverse. Among 405 colon strains, isolated from surface and ground
waters in Kansas, 70 percent were of the low ratio V. P. negative type and
30 percent of the high ratio V. P. positive type. Houston, Winslow and
Cohen, and Wood obtained similar findings to that of Greenfield. That
is, they also found a preponderance of the coli subgroup in water; while
Stokes, in a study of 528 cultures from water of Maryland, and Mac-
Conkey, among 49 strains from ponds, rain waters, and roof washings, ob-
tained a preponderance of the so-called non-fecal or high ratio aerogenes
and cloacae strains. Clemesha, in India, and Boles, in the Canal Zone,
point out that the Bact. coli is in excess immediately after pollution but
that later the Bact. aerogenes are the prevailing forms.

Although there is no agreement as to the relative incidence of these two
groups in water, it is apparent that the aerogenes section is proportionately
much more common than in feces and much less frequent than in soil. The
author thinks it is a reasonable assumption that the strains in water repre-
sent the resultant of sewage pollution, soil contamination, and proximity
of pollution.

The colon group isolated from milk is about evenly divided between
aerogenes and coli types. Wood, MacConkey, Orr, Rogers and Stokes

78

recorded 17.5, 29, 39, 47, and 69.4 percent respectively of the aerogenes
group in milk.

TABLE XLI. INCIDENCE OF AEROGENES TYPES AMONG COLON BACILLI
ISOLATED FROM WATER AND MILK

Investigators

Source

No. of strains
studied

% Aerogenes
section
V.P._1_;M.R.— )

Houston (1911)

Water

532

8.1

MacConkey (1909)

Water

49

59.2

Rogers (1916)

Water

137

66.0

Greenfield & Skorup (1917)

Water

405

30.0

Winslow & Cohen (1918)

Water

255

21.2

Wood (1919)

Water

231

33.3

Stokes (1919)

Water

528 63.7

Total

2137

35.2

Orr* (1908)

Milk

850

39.0

MacConkey (1909)

Milk

26

57.8

Rogers, Clark & Davis (1914)

Milk

124

47.5

Hulton (1916)

Milk

18

72.3

Wood (1919)

Milk

93

17.5

Stokes (1919)

Milk

271

59.8

Total

1382

43.1

*GUucose fermenters, lactose reaction not recorded.

The observations as to the incidence and distribution of the colon
group in nature may be summed up as follows:

The evidence indicates that lactose fermenting bacteria capable of
growing aerobically are widely distributed in nature and that they con-
stitute two distinct groups; a low ratio V. P. negative subgroup, which was
previously described as the coli section, is characteristic of the intestinal
tract of man and animals and relatively infrequent in localities not recently
polluted with human or animal intestinal material; and a high ratio V.-P.
positive group, termed the aerogenes section, which is very rare in the in-
testinal discharges of man and animals, but constitutes the predominating type
in soil and on grains. In water and milk the proportion of coli and aero-
genes strains naturally varies, the predominating type depending to a con-
siderable extent on the source of contamination.

Correlation of the Sanitary Survey and the Incidence of Coli
and Aerogenes Types in Water. If, as has been intimated in the fore-
going discussion, the coli and aerogenes sections are so characteristically

79

TABLE XLII. SUMMARY OF TYPES OF COLON BACILLI ISOLATED FROM

VARIOUS SOURCES

Source

No. of strains
observed

% Aerogenes
section

% Coli section

Human feces

2534

5.9

94.1

Animal feces

1832

7.4

92.6

Water

2137

35.2

64.8

Milk

1382

43.1

56.9

Grains

288

81.7

18.3

Soil

853

88.1

11.9

of different origin, should we not expect a correlation of the sanitary sur-
vey with the type of colon bacillus encountered in a water? Reliable in-
formation on this matter would certainly be a great aid in the interpreta-
tion of water analyses. Winslow and Cohen studied the distribution of these
organisms in waters of known sanitary quality with this idea in mind and
concluded that there was no relation whatsoever between the incidence of
aerogenes bacilli and the sanitary quality of the source. Thus among 94
strains isolated from polluted waters, 80 from unpolluted, and 81 from
stored waters in the vicinity of New Haven, Connecticut, organisms giving
the V.-P. reaction were encountered in 24, 24, and 15 percent respectively.
They conclude that the "significance of the ratio of Bact. coli and Bact.
aerogenes groups in a sanitary examination of water seems somewhat
dubious." Their results are quite the reverse of those reported by other
investigators.

Greenfield and Skorup conclude that there is a correlation between
the increased incidence of the Bact. coli strains and the sanitary survey
during dry weather.

TABLE XLIII. RELATIVE NUMBER OF AEROGENES AND COLI TYPES IN
WOLF CREEK. (Rogers 1918)

Miles

Pollution

Cultures isolated

Ratio aerogenes to coli

0

7

All aerogenes

1.1

Private Sewers

15

1.1

2.1

City sewer

18

0.8

3.1

11

0.1

5.1

20

9.0

7.1

10

All aerogenes

9.3

9

3.5

11.3

5

4.0

80

Rogers studied a section of a stream for a distance of about eleven
miles determining the ratio of Bad. aerogenes to Bact. coli above and be-
low sewer outlets.

It will be noted that above the city all cultures isolated were of the
aerogenes type and that after passing through the town, where the creek
became polluted, the Bact. coli strains predominated, being 10 times as
common, but farther down the stream, as the distance from the source of
pollution increases, the aerogenes forms again become markedly more
numerous. The indications are, therefore, strong that the ratio of aerogenes
to coli varies with the proximity of sewage pollution becoming greater as
the pollution is more remote.

Stokes remarks on the scarcity of the low ratio coli group in well
waters. Only one low ratio organism was found in 14 samples of water
from artesian wells containing members of the colon-aerogenes group.
Thirteen of these contained high ratio organisms including Bact. aerogenes
and Bact. cloacae.

In a study of 59 samples of raw waters, Stokes found 38 which showed
only one type of the colon group (24 contained Bact. aerogenes, 13 Bact.
coli. and 1 Bact. cloacae) determined by fishing 10 colonies from an Endo
plate. This striking fact that a large proportion of the samples harbored
a single species of the colon group in pure culture suggests the possibility
of a correlation between the source of pollution or contamination and the
type of colon organism present. He concludes that "In the bacteriological
examination of drinking water it would not be safe to assume that the
water was free from intestinal pollution if but two or three colonies were
studied from each specimen. However, if 20 to 25 colonies be picked
from each sample and all are found to be of the high ratio type, it might
be safe to regard the water as free from fecal contamination provided a
chemical examination and sanitary survey gave no evidence of pollution."

The writer would suggest that in place of picking such a large number
of colonies, which is impractical for routine, a differential medium like
the eosin-methylene-blue agar should be employed for confirming the pre-
sumptive test. The probable presence of the "fecal" Bact. coli type could
be easily recognized.

Perhaps the most detailed study on the correlation of sanitary quality
of ground water and the type of colon bacillus present was carried out by
D. R. Wood in England. A large number of samples from various sources,
200 of which contained lactose fermenting organisms, were investigated.
Sixty-six samples contained the V.-P. positive-methyl-red negative or Bact.
aerogenes types and 41 of these showed no evidence of recent excretal con-
tamination. He records also that in 22 deep well supplies of good repute
(these were sunk through lime-stone 250 to 400 feet into underlying sand,
used as public supplies for years, and appeared excellent bacterially, lactose
fermenters being rarely present and seldom in less than 100 c. c.) suddenly
showed a large proportion of the colon group in the spring of 1918. Lac-
tose fermenting organisms were obtained in two cases in 1 c. c., in several
instances in 10 c. c., but in every sample in 100 c. c. On further examina-

81

tion all but one of these organisms proved to be of the aerogenes section
(V.-P. positive, methyl red negative). As these wells were several miles
apart the simultaneous appearance of these organisms is difficult to explain.
Wood mentions that the time of year rather suggests some connection with
the sowing of grain.

It is felt that the relation of the sanitary survey to the type of colon
bacilli present in the water, particularly well supplies, is in need of con-
siderable investigation. The lack of correlation in the vicinity of the
New Haven, reported by Winslow and Cohen, is difficult to reconcile with
the observations of Stokes in Maryland and Wood in England. Possibly
local conditions, if taken into proper consideration, may explain the seem-
ing discrepancies.

On the Relative Viability of Bact. coli and Bact. aerogenes in
Water. Emulsions of feces were stored in water in bottles or suspended
in parchment sacks in running streams by Rogers. He reported that the
Bact. coli died off much more rapidly than the Bact. aerogenes. In water
kept in a bottle at 20° C., preliminary examination showed only Bact. coli
but after 3 days storage the ratio of Bact. aerogenes to Bact. coli was 1
to 6. This ratio gradually increased until after 278 days storage the Bact.
aerogenes was 39 times as prevalent as Bact. coli. This was of course a
prolonged period of storage but it indicates, nevertheless, that the aero-
genes group is more resistent and will persist for a much longer period
in water than will the coli group.

In running water the change in the ratio of Bact. aerogenes and Bact.
coli was much more rapid. Thus beginning with a fecal emulsion in parch-
ment sacks suspended in running water, an initial ratio of 1 Bact. aerogenes
to 2.3 Bact. coli was found after 7 days to contain 10 times as many Bact.
aerogenes as Bact. coli. The Bact. aerogenes were proportionately more
than 20 times as prevalent in a fecal suspension stored in running water for
a week. The death rate of Bact. coli (K=0.48) was about twice that of
Eact. aerogenes (K=0.93).

TABLE XLIV. CHANGES IN COLON BACTERIA IN RUNNING WATER.

(After Rogers 1918)

Age, days

Colon group, per c. c.

Ratio of Bact. aerogenes

to Bact. coli

0

190,000

1 to 2.3

1

130,000

1 to 4.0

2

19,000

1 to 3.3

3

9,000 1 to 1.2

4

20

1 to 1.1

7

30

1 to 0.11

Winslow and Cohen report similar, though not such striking results,
when storing mixtures of these organisms in bottles. They observed a

82

preliminary decrease in the proportion of Fact, aerogenes, after 24 hours,
followed by an increase. An initial incidence of 46 percent Bact. aerogenes
rose to 71 percent after 60 days' storage.

Savage and Wood, in studies on the vitality and viability of Strepto-
cocci in water also report a number of interesting observations on the rela-
tive incidence of the capsulated and non-capsulated members of the colon
group as affected by storage. The following experiment is illustrative of
their results.

To a liter of water which was not sterilized but which was known
to be free from the colon group, were added 0.03 c. c. of a 24 hour peptone
water culture of non-capsulated and capsulated colon organisms were add-
ed. The non-capsulated form (Bact. coli) with an initial count of 4,000
per c. c. fell to 25 per c. c., after 31 days, and 3 per c. c. after 33 days'
storage and was entirely absent in 50 days. The capsulated form (Bact.
aerogenes), on the other hand, was reduced from an initial count of 7,440
to 440 in 31 days, to 122 in 50 days, and living organisms were still pres-
ent in this water at the end of 14 weeks. In another instance, an initial
count of 8,000 non-capsulated organisms (Bact. coli) dropped to less than
1 per c. c., whereas the capsulated strain, with an initial count of 4,920,
was still present to the extent of 540 per c. c. after 28 days' storage.

TABLE XLV. ON THE VIABILITY OF COLON GROUP IN UNSTERILE TAP
WATER. (After Savage and Wood 1917).

Period of storage

Surviving bacteria per c. c.

Non capsulated (Bact. coli)

Capsulated (Bact. aerogenes)

At start

4,400

7,440

31 days

25

440

43 days

3

176

50 days

0

122

57 days

0

81

70 days

30

84 days

25

98 days

6

They also observed that when sewage was stored the Bact. coli strains
gradually disappear the ultimate colon survivors being all capsulated strains
(Bact. aerogenes).

Clemesha in India reports that Bact. aerogenes is relatively rare in
waters recently polluted but becomes extremely common 5 to 15 days after
contamination. He found the Bact. aerogenes more prevalent in rivers and
lakes shortly after rains but was later supplanted by another member of
the aerogenes section, Bact. cloacae, which was found to be the predom-
inating type during dry seasons.

83

Similarly, Bowles, working with waters in the Panama Canal Zone,
where the conditions simulate those of the tropics, notes that the aerogenes
types are extremely prevalent long after the Bact. coli forms have died off.
The supplies in the canal zone consist of impounded reservoirs and are
treated with sulphate, filtered, and in some instances chlorinated after fil-
tration. The reservoirs are policed and adequately protected against pol-
luion. In the Chagres River during the rainy weather Bact. coli was pres-
ent in 1 c. c. but during the dry season Bact. coli was found only in 50 c. c.
samples while Bact. aerogenes, he says, was present in large numbers. Dur-
ing the dry season, the reservoirs, acting as sedimentation basins, effect con-
siderable (wonderful) purification. He states that many examinations may
be made this time of year without getting any test for Bact. coli, but that
Bact. aerogenes is present.

There is no question, therefore, as to the relative viability of Bact.
coli and Bact. aerogenes (and Bact. cloacae) in water. The latter is much
more resistant, will persist for considerably greater periods, so that when
present alone (that is, not accompanied by Bact. coli) in natural waters,
it may merely indicate pollution or contamination so remote that the Bact.
coli, and consequently the dangerous intestinal disease producing forms,
have died off. In fact, Clemesha maintains that the presence of Bact. aero-
genes is a favorable indication that self-purification is rapidly taking place.

Relative Resistance of Coli and Aerogenes Types to Chlorina-
tion and Filtration. Greenfield and Skorup state that there is no differ-
ence in the resistance of these two forms to treatment.

Ellms in experimental studies with the Milwaukee water supply found
the aerogenes type less resistant particularly to chlorination, but that on sub-
sequent storage these forms became the predominating colon type.

TABLE XLVI. RELATIVE INCIDENCE OF COLI AND AEROGENES SECTIONS

IN RAW AND TREATED WATER SUPPLY OF MILWAUKEE.

(After Ellms 1920)

Lake

Settled

Filtered

Chlorinated

Filter A

Filter B

Filter1 A

Filter B

Basin

No. of strains

710

655

288

294

28

43

13

&V.-P.+

44
56

40

40

33

29

26

70

%V.-P.—

60

60

65

71

74

30

Levine recorded the following experiences with the water supply of
Dijon, France. The raw water examined daily for six months yielded 96
positive results for members of the colon group with Bact. aerogenes pres-
ent in 16.6 percent of these. The tap samples during the same period (the
water was treated with .07 to 0.1 part per million of chlorine, then stored
in distribution reservoirs) showed only three positively confirmed tests,
but Bact. aerogenes was present in each instance or 100 percent of the posi-
tively confirmed gas tubes.

84

Recently Hinman reported the relative incidence of these two groups
in the water supply of Iowa City which is both filtered and chlorinated. In
the raw water, out of 220 positively confirmed tests, aerogenes forms were
present in 35 or 16 percent. In the treated water only 11 samples were
confirmed for the colon group but of these 4 (36%) were Bact. aerogenes.

H. E. Jordan in a report on the Indianapolis water supply notes a
seasonal variation in the relative resistance of these forms to chlorination
which he found to be particularly effective against the coli section during
the summer.

TABLE XLVII. THE VIABILITY OF CAPSULATED AND NON-CAPSULATED

LACTOSE-FERMENTING ORGANISMS IN BOILED HARD WATER.

(After Wood 1919)

Non capsulated

Capsulatec

1 Organisms

Duration of
Experiment

"Typical" Bact. coli
V.P.— ; M.R.+

Strain T.
V.P.+ jM.R.-jIndoJ-
Bact. aerogenes

Strain N. D.
V.P.+ :M.R.-;Indol+
Bact. aerogenes

Start

1,250 per c. c.

580 per c. c.

Not enumerated, approx-
imately the same as
Strain T.

2 weeks

1 to 10 per c. c.

1 to 10 per c. c.

4 weeks

Present in 100 c. c.

Present in 0.1 c. c.

5 weeks

Absent from 100 c. c.

Present in 0.1 c. c.

6 weeks

Absent from 100 c. c.

Present in 0.1 c. c.

7 weeks

Present in 0.1 c. c.

12 weeks

Present in 0.1 c. c.

Present in 1 c. c.

19 weeks

Present in 0.1 c. c.

25 weeks

Present in 0.1 c. c.

33 weeks

2,800 per c. c.

39 weeks

2,000 per c. c.

1 year

Present in 1 c. c.

Present in 1 c. c.

1 year, 4 months

Present in 1 c. c.

1 year, 9 months

Present in 1 c. c.

2 years

Present in 1 c. c.

3 years, 6 months

Still present

The evidence as to the effect of filtration and chlorination, particularly
when followed by storage, on the relative prevalence of the coli and aero-
genes types is meager and somewhat conflicting. Further observations on
this phase are needed and would be of value. Clemesha states that the
aerogenes group, for a while at least, multiplies in water. It would be in-

85

teresting and significant to know whether its increased prevalence in stored
chlorinated water reported by Hinman, Ellms, H. E. Jordan, and the writer
is due to such a secondary multiplication or merely to survival over the
less viable Bact. coli.

That under certain conditions the Bact. aerogenes may multiply and
survive for an incredibly long time in water is shown by the following
experiment of Wood. He inoculated a hard water, which had been pre-
viously boiled and which contained no measurable trace of saline or or-
ganic ammonia, with Bact. coli and Bact. aerogenes and determined their
incidence and relative viability. The results are indicated in Table XLVII.

The Bact. coli disappeared quite rapidly; an initial count of 1,250
per c. c. decreasing to less than 1 in 100 c. c. after five weeks' storage. With
the aerogenes strains, however, there was an increase from 580 to 2,800
per c. c. after 33 weeks' storage and the organism was still present in one
experiment after 3% years.

Ellms in his studies on the Milwaukee supply detected colon forms
in the chlorinated stored water (basin samples) on only five of the ex-
perimental runs but in three of these the colon index (see table XLVIII) was
considerably higher than in the chlorinated filter effluents.

TABLE XLVIII. COLON INDEX IN CHLORINATED FILTERED AND STORED

WATER.

(Calculated from report of Ellms 1920)

Operation period

Colon bacilli per liter

Ratio of basin to chlorinated
filter effluents

Filter A*

Filter B*

Basin

9-8 —10-21, 1919

4

4

2

0.5

10-21—10-25, 1919

0

6

33

11.0

11-6 —11-12, 1919

74

61

50

.7

11-12—11-24, 1919

14

45

100

3.3

12-6 —12-9 , 1919

21

17

367

19.3

% cultures V.P.-f

*29

26

70

* Chlorinated filter effluents.

Taking into consideration that the filter effluents showed only 26-29
percent V. P. positive strains as compared with 70 percent after storage in
the basin, the writer feels that the rise of the colon index in the latter was
most likely due to multiplication of the aerogenes and cloacae forms.

That the colon index may increase on storage of a filtered chlorinated
water may be shown quite strikingly from the report of the Indianapolis
Water Company.

The incidence of colon bacilli in tap samples was at times 11 to 13
times as great as in the water works plant effluent.

86

TABLE XLIX. COLON GROUP PER 100 C. C. IN INDIANAPOLIS WATER SUPPLY

(Calculated from report of H. E. Jordon 1920)

Average for 3 years 1917-18-19.

Plant effluent

Tap

Ratio of tap to plant
effluent

* % Aerogenes in
plant effluent

Jan.

2.1

3.1

1.5

30

Feb.

0.87

1.94

2.2

40

Mar.

0.72

0.83

1.2

66

Apr.

0.50

0.73

1.5

69

May

0.63

0.36

1.0

t87

June

0.37

1.33

3.6

58

July

0.21

2.87

13.5

75

Aug.

0.13

1.47

11.0

82

Sept*

0.33

4.40

13.2

50

Oct.

0.47

1.67

3.6

43

Nov.

0.47

0.93

2.0

45

Dec.

2.17

1.50

0.7

19

*Average for five years includes two yews before chlorination.

•f-In two years before chlorination ratio of basin to plant effluent was 13.7.

In general the proportion of aerogenes (M. R. — -V. P. +) strains
in the plant effluent was greatest in the warmer months. The relative in-
cidence of aerogenes and coli types in the tap samples is not stated but it
will be noticed from Table XLIX. that secondary multiplication was most
marked in July, August and September. The temperature of the water
at these times is quite likely to be very near the optimum for growth of
Bact. arogenes and closely allied bacilli.

The foregoing observations are not sufficiently extensive to warrant
final conclusions. They nevertheless do indicate that under some con-
ditions there may be a secondary rise in the colon index of a filtered or
chlorinated water on storage, and that this is probably due to the growth
of the aerogenes section, for the increase is associated with those months
when the aerogenes types were relatively more numerous and when the
temperature of the water was near the optimum for growth of these forms.

Are all Varieties and Species of the Aerogenes and Coli Sec-
tions of the Same Sanitary Significance? Although the aerogenes
group is rare in human feces, it is nevertheless present in small numbers,
and similarly Bact. coli, although comparatively infrequent, is occasionally
encountered in presumably non-polluted sources. It certainly would be a
great aid if some test could be devised which would distinguish the Bact.
aerogenes from human feces and those from soils and grains, or which

87

would distinguish the Bact. coli of the human intestine from those in the
soil or of animal intestinal origin.

Rogers suggested that fermentation of adonite was a convenient differ-
ential index for the separation of the human from the grain strains of Bact.
aerogenes. Of 46 human strains, all fermented adonite, whereas of 111
strains isolated from grains, only 12.6 percent were adonite fermenters.

This differentiation was incorporated in the Standard Methods of 1917,
which divided the Bact. aerogenes forms into two subgroups, (1) the adonite
fermenting type supposedly of human origin and (2) the adonite non-fer-
menting type regarded of non-fecal origin.

Other investigators, however, do not agree as to the value of adonite
for this purpose. Monfort assigns to adonite about the same significance
as to dulcite, saying that it is of rather dubious importance and of little
significance as an index of pollution. Winslow and Cohen, in a study of
water of known sanitary quality, found no evidence to support the conten-
tion that adonite fermentation by the Bact. aerogenes was indicative of fecal
origin. In fact, their results were exactly the converse; 59 percent of the
aerogenes strains from polluted water fermented adonite, whereas 90 per-
cent from non-polluted sources were adonite fermenters which would in-
dicate that the adonite fermenters (fecal varieties) were more prevalent
in non-polluted water.

Chen and Rettger report 152 (34%) of their 447 aerogenes strains
from soil to be adonite fermenters. It would appear, therefore, that adonite
fermentation can not be considered, for the present at least, a reliable
criterion for the identificaiton of Bact. aerogenes of fecal origin.

In 1905, MacConkey first subdivided the colon group into separate
species, suggesting that information may thereby be obtained as to the
relative value of specific members of the group as indicators of pollution.
Clemesha advocates strongly such consideration and points out that the
dulcite non-fermenters1 ( MacConkey 's groups 1 and 4) are resistant in
water, whereas the dulcite fermenting forms (groups 2 and 3 of MacConkey)
are sensitive, dying off rapidly. He contends that for tropical water a
distinction must be made between the resistant dulcite non-fermerters and
the non-resistant dulcite fermenting organisms. Whether these sugges-
tions are as applicable to the temperate climate of our country is doubtful
except possibly the hottest season of the year or in our tropical possessions.
In this connection it should be pointed out that the observations of Houston
do not agree at all with those of Clemesha as to the relative resistance of the
dulcite fermenters and non-fermenters. The so-called sensitive dulcite
fractors were less prevalent while the supposedly resistant forms were found
to be more numerous in raw than in stored or filtered water.

Levine suggests that there is some correlation between the species of
the coli section and their source, and that the determination of species may
have some bearing upon interpretation of analyses.

From a study of 333 colon strains isolated from various sources, as
indicated in Table L., he observes with reference to the coli section that
Bact. communior was extremely abundant in the horse (79%) and sheep

(72.8%) but much less frequent in the cow (30%) and pig (29%);
whereas in sewage and man it was relatively rare (7.7 and 8% respectively).

TABLE L. DISTRIBUTION OF ORGANISMS FROM DIFFERENT SOURCES
AMONG THE VARIOUS SPECIES AND VARIETIES.

B. c/oa-
COf

B. aero-
qenes

B.com-
munior

B.oeapol-
/fanus

0. C03-

coroba

6.c<

/;

B. Qcidi-lactici

Ton/

com(nff)t5

'mmebiJis

6rvtntM

immobili

c_//

N°

68

54

se

O

O

2

0

7

0

177

•JO/I

%

49.7

30.5

14.7

1.1

4.0

Hnrto

Ho

O

0

15

0

0

4-

O

0

O

/<?

%

79.0

21.0

Sheep

M>

0

0

16

0

5

1

0

0

0

£?

%

72.8

22.7

4.5

Cow

Ho

O

O

6

4

O

9

O

/

0

&

%

30.0

20.0

45.0

5.0

p,9

/Yo

O

O

9

O

/

II

1

9

O

31

%

29.0

32

35.6

32

29.0

M>

/

B

3

3

2

/

13

2

7

39

JSWCXJi

%

2.6

20,5

7.7

7.7

5.1

2.6

30.8

5.1

17.9

flan

Ho

0

O

2

0

/

5

5

1

II

25

7.

8.0

4.0

20.0

20.0

4.6

44.O

Total

89

62

77

7

9

33

18

20

18

333

The Bact. neopolitanum was present only in bovine feces and sewage,
comprising 20 percent of the bovine and 7.7 percent of the sewage.

Bact. coscoroba occurred as follows: sheep 22.7 percent, pig 3.2 per-
cent, sewage 5.1 percent and human 4.0 percent.

The relation of sucrose fermentation and the source is especially em-
phasized. The sucrose fermenting strains are relatively uncommon in
human feces, whereas they constitute the predominating type in animal
feces and in the soil. This low incidence in human feces is confirmatory
of the observations of numerous other investigators.

In this connection, it may be well to recall that when Durham sug-
gested the name B. coli communior for the sucrose fermenting variety be-
cause of its greater prevalence, his observations were based on the intestinal
contents of animals for which this fact holds true, but, as has been pointed
out above, the sucrose positive strains are relatively scarce in man.

Bact. coli, like Bact. communior was isolated from all of the sources
tested, but a rather distinct correlation with the source is observed with
the varieties Bact. coli-communis and Bact. coli-immobilis. The former
comprise" 1.1 percent of soil; 21 percent of horse; 4.5 percent of sheep;
45 percent of cow; 35.6 percent of pig; 2.6 percent of sewage; and 20
percent of human strains. Bact. coli-immobilis was not obtained from

89

the soil, horse, sheep or cow. but it made up 3.2 percent of the pig, 30.8
percent of the sewage, and 20 percent of the human strains.

TABLE LI. FERMENTATION OF SUCROSE BY BACT.-COLI-L1KE BACTERIA

FROM HUMAN FECES.

Investigators

Number of
organisms
studied

Number of
sucrose
fermenters

Percentage of
sucrose
fermenters

Houston, 1902-3

100

30

30

MacConkey, 1905 and 1909....

419

142

33.9

Ferreira, Hoita, Paredes, 1908

117

44

37.6

Winslow and Walker 1907

25

8

32

Howe, 1912

540

324

60

Clemesha, 1912

1200

348

29

Browne, 1915

175

20

11.3

Levine, 1916

25

3

12

Total

2601

919

35.3

Bact. acidi-lactici was not obtained from the horse nor sheep, and only
rarely from the cow (5.0%) or soil (4.0%). The motile variety Bact.
acidi-lactici var. Grunthali was particularly abundant among the pig cul-
tures (29%) and rare in sewage (5.1%) and man (4%). The non-motile
Bact. acidi-lactici var. immobili was restricted to man and sewage entirely,
comprising 44 percent of the human and 17.9 percent of the sewage strains.

If the sucrose negative forms are more indicative of human pollution
it would be anticipated that they would be more prevalent in the more in-
tensely polluted waters. Observations by the writer on 78 samples col-
lected in France were as follows:

Among 34 samples in which the coli section was present in 1 c. c.
or smaller quantities, sucrose negative strains were detected in 23 or (68%),
whereas of 44 samples containing the coli section only in 10 c. c. or larger
quantities but 13 (29%) showed sucrose negative coli strains. That is,
the more polluted supplies apparently did contain a greater proportion
of sucrose non-fermenters. More extensive work on the correlation of
species of colon bacilli with the source, character of pollution, and history
of waters is certainly desirable.

Resume. From considerations of the requirements for an index of
pollution, the colon group appears to be a convenient and desirable one.
It is not however, an ideal indicator for the species which it comprises are
not all of equal sanitary significance.

The evidence seems to be clear and definite that the colon group com-
prises two subgroups or sections which are characteristically of different
habitat; one, typified by Bact. coli, is present in large numbers in feces
and sewage, whereas the other, exemplified by Bact. aerogenes, is rare in

90

such objectionable matter but predominates on the presumably harmless
soil and grains. The aerogenes group, as indicated* by laboratory experi-
ments and observations in the field, is much more viable in water, where
it will persist for long periods, and seems capable of growing, to some
extent, in stored treated supplies.

It is not possible from our present limited knowledge of these two
groups to put forth any definite rules for interpreting the significance of
their presence in water, but it is felt that the presence of Bact. aerogenes
alone should not be regarded as objectionable as is the presence of the
Bact. coli in equal numbers. If the sanitary survey is favorable and there
is no evidence of the true Ract. coli types in a water supply under different
weather conditions, then a considerably greater number of Bact. aerogenes
may be tolerated.

The presence of Bact. aerogenes alone (i. e. not associated with Bact.
coli) in a supply may indicate merely remote pollution or soil contamina-
tion which is not as objectionable and certainly not as dangerous as
sewage pollution.

Differentiation of the coli and aerogenes types in routine water an-
alysis is obviously desirable as it may assist in the detection of the probable
source and nature of the contamination.

91

VI. THE SPORE FORMING LACTOSE FERMENTERS AND
THEIR SIGNIFICANCE IN WATER ANALYSIS.

Spore forming lactose fermenters are not infrequently encountered
in water. They are for the most part anaerobes resembling the Cl. welchii
(B. aerogenes capsulatus of Welch) or the Cl. enteritidis sporogenes group
of Kline, but recently spore formers capable of growing on aerobic plates
have been reported and isolated by Meyers, Ewing and Ellms and Hinman
and Levine. They interfere seriously with the presumptive test for the
colon group.

Sanitary Significance. Very little is known as to the source, dis-
tribution, or pathogenicity of the aerobic sporing types. Ellms reports
their presence in feces while Ewing emphasizes that they were present in
water only after heavy rains, so that they may represent soil forms. No
reliable conclusion can be drawn at present as to their sanitary significance.

The anaerobes may be frequently encountered, often in large numbers,
in the intestinal tract of man. Kline and Houston report 30 to 2,200
CL (enteritidis) sporogenes per c. c. in sewage, whereas in waters of good
quality such forms are often absent even from large volumes. These obser-
vations have sometimes led to the contention that the presence of Cl. (enter-
itidis) sporogenes or closely allied bacteria in water is indicative of fecal
pollution. The employment of these organisms as indices of dangerous
pollution appears to the author unwarranted, undesirable and impractical
for,

1. They are not characteristic of the human intestine.

2. There is a very little correlation between the incidence of Sporogenes
and Welchii types in water and the sanitary survey.

3. They are extremely resistant.

The anaerobic lactose fermenters are verv widely distributed in nature.
They are encountered in large numbers in manures from various animals, in
decomposing organic materials, and in the soil. They cannot be considered
distinctive or characteristic of the human intestinal tract.

In surface waters, the number of spores is remarkably constant, quite
independent of the degree of sewage pollution as indicated by sanitary in-
spection.

Thus Gumming reports "Unlike B. coli which varies many thousand
percent, from several hundred per c. c. to less than 1 in 10 c. c., according
to the intensity of pollution, these spores were found often in the best
river water in 10 c. c. and seldom showed an average much above 4 or
5 per c. c "

"Their number furnishes no clue to the degree of pollution and puri-
fication as does the number of B. coli "

"The generally uniform distribution of organisms of this group in
surface water, even in those not highly contaminated with sewage, and with
no considerable increase in polluted waters, indicates that this group is
not, as has sometimes been supposed, an organism characteristic of the
intestine."

92

c p- 73

* i A >> %1 »

*2 3 "3 "3 "-.2

Not pathogen

1

I

« c

^

S K®

O o c

*8.

•*•

i!

I;Hri

ill

*|S|R

g O S 33
/. — Jr -

"2^
•a*

09

r
jz =

« Ilia

i -ol

s^S£^_

i

O Hi

111

•SsSI I4ti
a- =' 'slf

w

«

es oval and
erminal ; not
ily formed.

Spo
h

sn
rea

PI

S

s

Ce
ter

lifi

s s

2 S

O 02

+ 1

+ 1

+ 1

+

.2

03 G3CQOQ

O </5

•aS'i

l?l

"S£S
^"«5

U

'C

H

"3 e
*&

^

93

Furthermore, as these anaerobes are extremely resistant they would
persist in water for considerable periods beyond any possible danger of
transmission of intestinal disease (as typhoid, dysentery, cholera) the or-
ganisms of which (being non spore bearing) die off quite rapidly in water.
Again they resist disinfection and other methods of water purification,
which experience has shown efficient and sufficient for elimination of
disease producing microorganisms. The presence of these spore formers
in a treated water is entirely insignificant as an index of danger from ty-
phoid or similar diseases.

It is conceivable, however, that some members of the Welchii group
may themselves be responsible for intestinal disturbances. Herter in this
country, and Klein in England have considered Cl. welchii as a cause of
diarrhea in children. Such affections have been traced in a few instances
to milk.

As an index of specific infection, the determination of the Welchii
group in water analysis may possibly be of value, but if employed for this
purpose, it would obviously be necessary to first determine the particular
species or variety which is an intestinal pathogen. The terms Cl. welchii
and Cl. (enteritidis) sporogenes, as employed by water works opeartors
and analysts, designate not a specific organism but •the whole group of
anaerobic lactose fermenters. Some of these are known to be responsible
for serious wound infections but many are harmless and no specific strain
has been definitely proven to be pathogenic when restricted to the intes-
tinal tract.

The Anaerobic Sporing Lactose Fermenting Species and Vari-
eties. These are for the most part rather long rods, possessing central or
subterminal spores usually slightly larger than the diameter of the cell.
During the war considerable work on anaerobes was in progress and an
extensive literature has developed which was carefully summarized in the
Special Report No. 39 of the Medical Research Committee of England.

Table LIIL, taken for the most part from the foregoing report sum-
marizes the characteristics of the lactose fermenting anaerobes.

TABLE LIIL FERMENTATION REACTION OF LACTOSE FERMENTING SPORING

ANAEROBES.

Species

•g

g

3

0

|

0

a

o

o

£

cd

>;

o

13
Q

""oS
CA)

2

C/2

3

c

I-H

Cl. welchii*
Cl. oedematis
Cl. chauvoei
Cl. aerofetidum
Cl. butyricum
Cl. multi-fermentans
Cl. tertium
CL sphenoides

+ !
+
+
+
+
+
+
+

+

__!_

+

+

j_
±

+
+
•f

+
±

+
+
+
+

+

+
+
+
+
+
+
+
+

+

+

+
+
+
±

—

| _
i -

1 -
1 +

L±

H-

+
. -

-
-

- I
+

+

+
+
+

+

+

+
-H

-H

4-

*Four types have been differentiated on innlin and glycerol fermentation.
Type I. Inulin_|_, glycerol_|_. Type II. Inulin — , glycerol_|_. Type III.
glycerol — . Type IV. Inulin — , glycerol — .

Inulin_|_,

94

KEY TO THE MORE COMMON SPORE PRODUCING LACTOSE FERMENTING

ANAEROBES.

I. Gelatin liquefied (generally pathogenic).

A. Coagulated serum, liquefied (non-pathogenic)

1. Cl. aerofetidum

B. Coagulated serum not liquefied (pathogenic)

1. Non motile 2. Cl. welchii

2. Motile

a. Sucrose fermented salicin not attacked

3. Cl. chauvoei

b. Sucrose not attacked, salicin fermented

4. Cl. oedematis

II. Gelatin not liquefied (non pathogenic)

A. Non motile or very faintly motile 5. CL tertium

B. Motile

1. Salicin not fermented 6. Cl. butyricum

2. Salicin fermented

a. Glycerol and inulin fermented 7. CL multifermentans

b. Glycerol and inulin not fermented.

8. CL sphenoides

Creel made a very interesting and, from the standpoint of the pre-
sumptive test, an important observation in his study of drinking waters on
railroad trains. He found two types of anaerobes which he designates
"Group A and B" respectively. "Group A" comprises very long Gram-
negative bacili (6-9 microns long and about 0.5 microns in width) whose
spores are very much larger than the diameter of the cell. It is particularly
significant to note that this group grows very rapidly producing consider-
able gas in lactose broth, under anaerobic conditions, but that no gas is
evolved in the lactose bile medium. Creel's anaerobes of "Group B" were
found to be non-motile, capsulated, resembling in general CL welchii. This
group will not produce gas in lactose broth unless the medium has been
freshly boiled or steamed in the Arnold until all air is expelled but in or-
dinary lactose bile, on the other hand, gas is formed very readily.

These observations may explain the controversy as to the relative value
of lactose broth and bile as a presumptive test for the colon group. In
dealing with waters -containing anaerobes of "Group A" lactose bile will
prove more reliable whereas if "Group B" is the predominating anaerobe
then lactose broth will be found to give a higher proportion of confirmed
presumptive tests.

Isolation of Anaerobes. Several methods have been suggested and
employed for the isolation of anaerobic spore forming lactose fermenters
from water. That originally described by Kline is as follows:

Milk is inoculated with the water under examination and heated for
10 minutes at 80° C. to destroy vegetative cells. The tube is then cooled,
made anaerobic, and incubated at the body temperature. In about 24
to 36 hours, a characteristic, so-called "sporogenes" reaction will be ob-

95

served. This is described by Kline as follows:

"The cream is torn or altogether dissociated by the development of
gas so that the surface of the medium is covered with stringy, pinkish-white
masses of coagulated casein enclosing a number of gas bubbles. The main
portion of the tube formerly occupied by the milk now contains a color-
less, thin, watery whey, with a few casein lumps adhering here and there
to the sides of the tube. When the tube is opened, the whey has a smell of
butyric acid and is acid in reaction. Under the microscope the whey is
found to contain numerous rods, some motile, others motionless."

The method employed by Creel, which was said to be very efficient
for isolation in pure culture is given herewith.

Petri dishes are selected having covers considerably larger than the
inner plates. Agar is poured into the inner dish, allowed to harden and
then inoculated by smearing over the surface with material from a broth
or lactose bile tube suspected of containing an anaerobe. The inoculated
plate is then inverted into the large cover. Three grams of pyrogallic
acid are placed in the cover, one c. c. saturated potassium * hydroxide is
inserted with a pipette, and the two dishes are immediately sealed with
melted paraffin. Incubation is at the body temperature.

In the Standard Methods of Water Analysis A. P. H. A. 1912, a de-
tailed procedure for the detection of these anaerobes is recommended as
given below:

"B. sporogenes* is indicated by a vile odor which is produced in the
liver broth fermentation tubes used in the regular test for general gas form-
ing bacteria. The specific tests are made as follows:

1. Inoculate various dilutions (usually 0.1, 1.0 and 10.0 c. c.) of
water, or of sewage in higher dilutions, into fermentation tubes containing
liver broth and incubate for 24 hours at 37° C. If B. sporogenes is pres-
ent in the dilutions used, there will be vigorous gas formation, accompanied
by an offensive odor, and numerous large spores will be present.

2. Transfer the entire contents of each tube showing gas plus char-
acteristic odor into separate sterile Erlenmeyer flasks or large test tubes
and heat to 80° C. for 10 minutes.

3. One (1) c. c. (not more) of broth containing sediment is with-
drawn from the bottom of each of the flasks or tubes which have been
heated, and is planted separately into a second set of sterile liver broth
fermentation tubes and incubated for 24 hours at 37° C. after which time
gas formation and characteristic odor will again be observed. Microscopic
examination will reveal the presence of numerous large sluggishly motile
bacilli containing spores. Usually B. sporogenes is now present in pure
culture.

4. A stab culture made from this 24 hour liver broth culture into
dextrose liver gelatine or nutrient gelatin will demonstrate the presence
of B. sporogenes by the following characteristic growth. After 48 hours
incubation at 20° C., a distinct anaerobic growth will be observed begin-

*Term employed to designate the Welchii group and not a specific organism.

96

ning about two centimeters below the surface. Liquefaction will be well
advanced and gas bubbles will accumulate at the top of the liqufied area.

5. In order to obtain colonies of B. sporogenes on agar plates it is
necessary to transplant a few drops of broth and sediment from the second
set of fermentation tubes, in step 3, into a third set of tubes and incubate
for three to five hours at 37° C. After that period a distinct anaerobic
growth will be observed in the closed arm, and a few bubbles of gas will be
seen at the top. The B. sporogenes is now in the vegetative state and this
is the only condition in which it will grow on the plates.

The contents of the closed arm are transferred to the open bulb by
tilting forward, and plated in dilutions of 1.0 to .00001 c. c. on dextrose
liver agar, and incubated for 12 to 18 hours in hydrogen at 37° C. Typical
colonies will then be visible consisting of one or more gas bubbles sur-
rounded by a delicate white fringe. The plate cultures also have a dis-
agreeable cheesy odor.

6. From one of these typical colonies a deep stab culture is made
into dextrose* liver agar and incubated for 24 hours. A distinct anaerobic
growth will be observed along the line of puncture and sometimes the
agar is split into two or three layers by the gas evolved.

7. A sub-culture may also be made into litmus milk and incubated
for 48 hours, anaerobically, after which time there will be a complete
separation of curd and whey and a strong odor of butyric acid. Some-
times the curd adheres to the sides of the tubes and has a peculiar shredded
appearance."

It should be emphasized that none of the foregoing methods will
yield all the sporing lactose fractors. Detailed procedures for isolation
of specific forms together with the special media necessary are described
in the English Report on Anaerobic Bacteria and Infections, to which ref-
erence has already been made.

The Aerobic Sporing Lactose Fermenters. Meyers, in 1918,
isolated a spore forming lactose fermenter capable of growing on the sur-
face of Endo agar, from the water supplies of Newport and Covington,
Kentucky and from tannery wastes.

The organism is Gram negative, grows readily on Endo agar, pro-
ducing a red colony in 24 hours, which shows a distinct metallic luster
after 48 hours, and in Clark and Lubs medium it is alkaline to Methyl
Red and positive for the Voges Proskauer Reaction. The more important
characteristics as detailed by Meyers are:

Agar slant. Growth quite distinctive. At 37° C., in 24 hours, thin
transparent veil-like growth over entire surface except the very top. Growth
lobate along upper edge. Microscopically — in 24 hours mainly vegetative
forms, in 48 hours spore-bearing forms and later only free spores.

Endo's plates at 37° C. In 24 hours, colonies pink with red center,
irregular contour, one to two m. m. diameter, little or no sheen. Colonies
48 hours, deep red, much sheen in colonies and surrounding medium. Lat-
ter point distinctive.

97

Gelatin stab at 20° C. In 48 hours, beginning liquefaction; in 72
hours liquefaction infundibuliform, slight precipitate.

Carbohydrates. In standard extract broth to which has been added
one percent of the following carbohydrates, acid and gas are formed: (1)
glucose, (2) laevulose, (3) raffinose, (4) maltose, (5) sucrose, (6) lactose,
(7) inulin, (8) starch, (9) glycerol, (10) mannitol. No acid or gas and
little growth in dulcite broth, which remains clear and limpid. In other
broths gas usually appear in 24 hours. Media uniformly clouded, slight
stringy precipitate, no pellicle. Media 48 hours, slightly viscous.

Clark and Lubs. Typical reaction of 'Grain' type coli, in 48 hours
at 37° C., i. e. reaction alkaline to methyl red, Voges-Proskauer test posi-
tive.

Indol production in 1 percent peptone, four days at 37° C. No
indol detected when tested for by the nitrite and by Ehrlich's p ar a- dimethyl -
amido-benzaldehyde method.

Glucose-neutral-red broth. Same reaction as Bact. coli i. e. yellow
fluorescence with gas formation.

Litmus milk at 37° C. In 24 hours acid, in 48 hours partially
reduced, coagulated with extrusion of whey; beginning digestion of curd.

Lactose bile at 37° C. In 96 hours, no gas or growth.

Chromogenesis. None.

Ewing, in 1919, reported a similar bacillus in the water supply of
Baltimore and noted that its presence was associated with heavy rainfall.

TABLE LIV. ELIMINATON OF SPURIOUS PRESUMPTIVE TESTS.
(After Hall and Ellefson 1918)

Series

Gentian
violet
concentration

Positive
presumptive
tests

Colon
group
present

% Presumptive
tests confirmed

A

None
1-100,000

21

20

12
15

57.1
75.0

B

None
1-20,000

44
33

26

27

59.1
81.8

C

None
1-20,000

85

74

58
61

68.2
82.4

D

Samples heated
600. C 30 min.

None
1-20,000

81
5

3

1

3.7
20.0

Inhibition of Growth of Spore Formers. Hall has suggested the
use of gentian violet in the lactose broth presumptive test tube to eliminate
spurious presumptive tests. In a series of examinations with water and
pure cultures, he found that 1-20,000 to 1-100,000 gentian violet in lactose
broth exerted but little inhibitory influence on the growth of Bact. coli;
whereas the anaerobes were almost completely checked.

Table LIV. shows the marked inhibitory action of gentian violet on
the anaerobes. Not only are spurious presumptive tests eliminated but
the total of successful isolations of colon bacilli is increased. This

98

would indicate that the presence and growth of the Welchii group is detri-
mental to the successful isolation of Bact. coll. This work of Hall and
Ellefson has recently been confirmed by Wagner and Monfort.

Ellms (1919) reports sporing lactose decomposing aerobes from the
water supply of Milwaukee, and the feces of children. These strains differ
from those of Meyers and Ewing in that they are Gram positive and are
acid to methyl red. The importance of these aerobic spore formers to
the bacteriologist and engineer is apparent. Being much more resistant
than the non-sporing colon group, they would naturally survive the ordinary
methods of water treatment, and as they are capable of growing aerobic-
ally, they may be mistaken for Bad. coli, Bact. aerogenes, etc., in the or-
dinary routine examination. Their presence in a water may conceivably
account for the apparently poor results sometimes obtained in purification
processes.

Muer and Harris, of the Mount Prospect Laboratory, observed that in
lactose peptone bile a concentration of 1-700 to 1-1000 brilliant green would
not appreciably affect the volume of gas produced by Bact. coli in seven
days, whereas Cl. ivelchii would not produce gas until the brilliant green
was diluted to 1-50,000. This observation is quite remarkable for, as is
well known, brilliant green is frequently used to inhibit Bact. coli. The
writer has observed a dilution of even 1-2,000,000 has a marked inhibitory
effect on the rate of growth of Bact. coli in peptone water. Their results,
however, are very distinct and significant. Probably the bile reacts in
some way with brilliant green. It is well known, for example, that in eosin
brilliant green agar a much higher concentration of brilliant green can be
employed without affecting Bact. typhi than in the Andrade brilliant green
medium of Krumwiede. The following table shows very clearly the in-
hibitory action of brilliant green on the anaerobic spore formers. It ap-
pears also that the growth of the anaerobes interferes with the isolation
of Bact. coli. In 17 samples Bact. coli was isolated when brilliant green
lactose bile was employed but was missed when plain lactose peptone bile
was used.

TABLE LV. COMPARISON OF 115 SAMPLES OF WATER PLANTED IN BOTH

PLAIN LACTOSE PEPTONE BILE AND BRILLIANT-GREEN LACTOSE BILE.

(After Muer and Harris 1920)

Medium

Number of dilutions from
which Bact. coli was
isolated

Number of dilutions from
which Cl. welchii was
isolated

Plain lactose-peptone bile

34

18

Brilliant-green lactose bile

51

0

The chief interest of the water bacteriologist in these spore formers
is that the anaerobes interfere and confuse the presumptive test rendering
confirmation necessary, while the aerobic forms complicate the confirmatory
test as well, making it essential to resort to more detailed identification tests
where there is reason to suspect this type present.

99

APPENDIX A. ROUTINE METHODS OF WATER ANALYSIS
AND THE COLON INDEX.

Although the American Public Health Association has been issuing
standard methods of water analysis for some fifteen years, there is still
a marked lack of uniformity in the methods employed in different labora-
tories. Thus Norton (1918), in a tabulation of 23 laboratories, found
that 13 employed lactose broth, 8 lactose bile, 1 lactose and dextrose broth,
and 1 lactose agar and dextrose broth for primary inoculation or prelim-
inary enrichment. The most commonly employed routine methods are
given here.

I. The Treasury Department Standard for the Examination
of Water on Interstate Common Carries. The following method for
the examination of water on Interstate common carriers has been formu-
lated by a committee of prominent sanitarians. The permissible limits
of bacteriological impurity are stated as follows:

1. The total number of bacteria developing on standard agar plates,
incubated 24 hours at 37° C., shall not exceed 100 per cubic centimeter;
provided, that the estimate shall be made from not less than two plates,
showing such numbers and distribution of colonies as to indicate that the
estimate is reliable and accurate.

2. Not more than one out of five 10 c. c. portions of any sample ex-
amined shall show the presence of organisms of the Bacillus coli group when
tested as follows:

(a) Five 10 c. c. portions of each sample tested shall be planted,
each in a fermentation tube containing not less than 30 c. c. of lactose
peptone broth. These shall be incubated 48 hours at 37° C. and observed
to note gas formation.

(b) From each tube showing gas, occupying more than five per-
cent of the closed arm of the fermentation tube, plates shall be made
after 48 hours' incubation, upon lactose litmus agar or Endo's medium.

(c) When plate colonies resembling B. coli develop upon either of
these plate media within 24 hours, a well-isolated characteristic colony
shall be fished and transplanted into a lactose-broth fermentation tube,
which shall be incubated at 37° C. for 48 hours.

For the purpose of enforcing any regulations which may be based
upon these recommendations the following may be considered sufficient
evidence of the presence of organisms of the Bacillus coli group.

Formation of gas in fermentation tube containing original sample of
water (a).

Development of acid-forming colonies on lactose-litmus-agar plates
or bright red colonies on Endo's medium plates, when plates are prepared
as directed above under (b).

The formation of gas, occupying 10 percent or more of closed arm of
fermentation tube, in lactose peptone broth fermentation tube inoculated
with colony fished from 24 hour lactose litmus agar or Endo's medium
plate.

100

These steps are selected with reference to demonstrating the presence
in the samples examined of aerobic lactose-fermenting organisms.

3. It is recommended, as a routine procedure, that in addition to
five 10 c. c. portions, one 1 c. c. portion, and one 0.1 c. c. portion of
each sample examined be planted in a lactose peptone broth fermentation
tube, in order to demonstrate more fully the extent of pollution in grossly
polluted samples.

4. It is recommended that in the above-designated tests the culture
media and methods used shall be in accordance with the specifications
of the committee on Standard Methods of Water Analysis of the American
Public Health Association, as set forth in "Standard Methods of Water
Analysis" (A. P. H. A., 1912).

II. English Procedure (After Savage). Add 0.1 and 1.0 c. c.
of water respectively to tubes of lactose bile salt broth in double tubes.
Add 10 c. c. to a similar tube, but containing lactose bile salt broth of
double strength. To the remainder in the bottle, after all the different
amounts of water have been withdrawn for the different parts of the
examination, add the contents (about 10 c. c.) of a tube of four times
strength neutral red broth. Replace the glass stopper. Four times
strength bile salt broth may be used, and, if the examination is for B. coli
alone, is preferable, but by using neutral red broth the mixture is also avail-
able for the examination for streptococci.

If a 2-ounce sample is collected, the amount remaining in the bottle
will be about 30 c. c. If a large sample of water is collected, then 50 c. c.
should be added by sterile pipette to a tube of four times strength neutral
red broth large enough to hold the added water.

The tubes are labeled, incubated at 37° C., and examined after 24
and after 48 hours.

If the 0.1, 1.0, and 10 c. c. tubes show no gas after 48 hours, it can be
assumed that B. coli is absent in these amounts. Then, in every case, the
larger amount (i. e. the 30 c. c. in the bottle) should be examined for this
organism. The alteration of the red color to yellow, with the presence
of fluorescence, is an indication of the probable presence of B. coli.

If gas is present in the tubes containing smaller amounts, use the one
showing gas in the tube with the least quantity of added water for in-
oculating plates of solid media. In this way it can be definitely ascertained
whether B. coli is present or absent in 50 c. c. or less, and if present,
approximately in what numbers.

To isolate the B. coli group organism, a trace of the positive tube
selected is distributed over the surface of a plate containing neutral red
lactose bile salt agar (L. B. A.), fuchsin agar or some other medium se-
lected. L. B. A. is recommended as most suitable. Several colonies
should be subcultivated and worked out.

Subcultivation upon or in the following five media is recommended
for routine work, i. e.:

(a) Gelatine slope (for morphology, motility, cultural appearance,
and liquefaction).

101 " A ; -r^^S ",' - A

(b) Litmus-milk at 37° C.

(c) Lactose-peptone litmus solution (in a double tube).

(d) Peptone water (for indol production).

(e) Saccharose peptone litmus solution (in a double tube).

III. American Public Health Association. Standard Method.
The 1920 report of the Committee on Standard Methods of Water Analysis
of the American Public Health Association suggested the following:

It is recommended that the B. coli (colon) * group be considered as
including all non-spore-forming bacilli which ferment lactose with gas for-
mation and grow aerobically on standard solid media.

The formation of 10 percent or more gas in a standard lactose broth
fermentation tube within 24 hours at 37° C. is presumptive evidence of
the presence of members of the B. coli group, since the majority of the
bacteria which give such a reaction belong to this group.

The appearance of aerobic lactose-splitting colonies on lactose-litmus-
agar or Endo's medium plates made from a lactose-broth fermentation tube
in which gas has formed confirms to a considerable extent the presumption
that gas-formation in the fermentation tube was due to the presence of
members of the B. coli group.

To complete the demonstration of the presence of B. coli as above
defined, it is necessary to show that one or more of these aerobic plate
colonies consists of non-spore-forming bacilli which, when inoculated into
a lactose-broth fermentation tube, form gas.

It is recommended that the standard tests for the B. coli group be
either (a) the Presumptive, (b) the Partially Confirmed, or (c) the Com-
pleted test as hereafter defined, each test being applicable under the
circumstances specified.

A. Presumptive Test. 1. Inoculate a series of fermentation
tubes with appropriate graduated quantities of the water to be tested. In
every fermentation tube there must always be at least three times as much
medium as the amount of water to be tested. When necessary to examine
larger amounts than 10 c. c. as many tubes as necessary shall be inoculated
with 10 c. c. each.

2. Incubate these tubes at 37° C. for 48 hours. Examine each tube
at 24 and 48 hours, and record gas-formation. The records should be
such as to distinguish between:

(a) Absence of gas-formation.

(b) Formation of gas occupying less than 10 percent of the closed
arm.

(c) Formation of gas occupying more than 10 percent (10%) of
the closed arm.

More detailed records of the amount of gas formed, though desirable for
purposes of study, are not necessary for carrying out the standard tests pre-
scribed.

*Parenthesis author's.

102

3. The formation within 24 hours of gas occupying more than 10
per cent. (10%) of the closed arm of fermentation tube constitutes a
positive presumptive test.

4. If no gas is formed in 24 hours, or if the gas formed is less than
10 percent (10%), the incubation shall be continued to 48 hours. The
presence of gas in any amount in such a tube at 48 hours constitutes a
doubtful test, which in all cases requires confirmation.

5. The absence of gas formation after 48 hours' incubation consti-
tutes a negative test. (An arbitrary limit of 48 hours' observation doubt-
less excludes from consideration occasional members of the B. coli group
which form gas very slowly, but for the purposes of a standard test the
exclusion of these occasional slow gas forming organisms is considered
immaterial ) .

B. Partially Confirmed Test. 1. Make one or more Endo's
medium or lactose-litmus-agar plates from the tube which, after 48 hours'
incubation, shows gas formation from the smallest amount of water tested.
(For example, if the water has been tested in amounts of 10 c. c., 1 c. c.,
and 0.1 c. c. gas is formed in 10 c. c., and 1 c. c., not in 0.1 c. c. the test need
be confirmed only in the 1 c. c. amount).

2. Incubate the plates at 37° C., 18 to 24 hours.

3. If typical colon-like red colonies have developed upon the plate
within this period, the confirmed test may be considered positive.

4. If, however, no typical colonies have developed within 24 hours,
the test cannot yet be considered definitely negative, since it not infre-
quently happens that members of the B. coli group fail to form typical
colonies on Endo's medium or lactose-litmus-agar plates, or that the colon-
ies develop slowly. In such case, it is always necessary to complete the
test as directed under "C" 2 and 3.

C. Completed Test. 1. From the Endo's medium or lactose-litmus-
agar plate made as prescribed under "B", fish at least two typical colon-
ies, transferring each to an agar slant and a lactose broth fermentation
tube.

2. If no typical colonies appear upon the plate within 24 hours, the
plate should be reincubated another 24 hours, after which at least two
of the colonies considered to be most likely B. coli, whether typical or
not, shall be transferred to agar slants and lactose broth fermentation
tubes.

3. The lactose broth fermentation tubes thus inoculated shall be in-
cubated until gas formation is noted; the incubation not to exceed 48
hours. The agar slants shall be incubated at 37° C. for 48 hours, when
a microscopic examination shall be made of at least one culture, selecting
one which corresponds to one of the lactose broth fermentation tubes which
has shown gas-formation.

The formation of gas in lactose broth and the demonstration of non-
spore-forming bacilli in the agar culture shall be considered a satisfac-
tory completed test, demonstrating the presence of a member of the B. coli
group.

103

The absence of gas-formation in lactose broth or failure to demon-
strate non-spore-forming bacilli in a gas-forming culture constitutes a
negative test.

APPLICATION OF PRESUMPTIVE, PARTIALLY CONFIRMED,
AND COMPLETED TESTS.

A. The Presumptive Test. When definitely positive, that is showing
more than 10 percent (10%) of gas in 24 hours, is sufficient:

(a) As applied to all except the smallest gas-forming portion of each
sample in all examinations.

(b) As applied to the smallest gas-forming portion in the examina-
tion of sewage or of water showing relatively high pollution,
such that its fitness for use as drinking water does not come
into consideration. This applies to the routine examinations of
raw water in connection with control of the operation of puri-
fication plants.

2. When definitely negative, that is showing no gas in 48 hours, is

final and therefore sufficient in all cases.

3. When doubtful, that is showing gas less than 10 percent (10%)
(or none) in 24 hours, with gas either more or less than 10 per-
cent in 48 hours, must always be confirmed.

B. The Partially Confirmed Test. 1. When definitely positive, that
is, showing typical plate colonies within 24 hours, is sufficient:

(a) When applied to confirm a doubtful presumptive test in cases
where the latter, if definitely positive, would have been suffi-
cient.

(b) In the routine examination of water-supplies where a sufficient
number of prior examinations have established a satisfactory
index of the accuracy and significance of this test in terms of
the completed test.

2. When doubtful, that is, showing colonies of doubtful or negative
appearance in 24 hours, must always be completed.

C. The Completed Test. The completed test is required as applied to
the smallest gas-forming portion of each sample in all cases other
than those noted as exceptions under the "presumptive" and the "par-
tially confirmed" tests.

The completed test is required in all cases where the result of the
confirmed test has been doubtful.

IV. Modification of A. P. H. A. Method. The following pro-
cedure, which is a modification of the A. P. H. A. method has proved
very satisfactory and convenient in the hands of the author.

104

PROCEDURE FOR THE EXAMINATION OF WATER. (BACTERIOLOGICAL)

Steps in procedure

Further tests
required

1. Plate two 1 c. c. and one 0.1 c. c. or other ap-
propriate portions of the sample on plain agar,
and incubate for 24 hours at 37° C.

2. Inoculate 10 c. c., 1 c. c. or other portions
of the sample into lactose broth (or lactose
peptone water). Incubate at 37° C.

3. Optional if a very rapid result is neces-
sary. 8 to 10 hours after incubation perform
a Preliminary Confirmatory test as follows:
Divide an Eosine-methylene blue, (or Endo, or
Litmus-lactose) agar plate into sectors. Streak
out a drop or loop of a fermentation tube con-
taining 10 c. c. of the sample on one of the
sectors, and in a similar manner streak out
other fermentation tubes on the remaining
sectors. Incubate at 37° C.

1. Count agar plates made the previous day, re-
cord and discard petri dishes.

2. Record presence or absence of gas after
hours incubation as follows: 10% or more, —
Positive; less than 10%, — Doubtful; no gas
—Negative.

3. Gas formation (any amount) accompanied by
a positive preliminary confirmatory test, con-
stitutes what is known as a Partially Con-
firmed test for the colon group.

4. If 10% or more gas is formed, and the prelim

inary confirmatory test is negative or was not 7

made, the result is regarded as a positive Pre' ^e or raw
sumptive test for the colon group.

5. If less than 10% gas is formed, and the pre-
liminary confirmatory test is negative or was
not made, the result is regarded as a doubtful
presumptive test, and should be confirmed.

None

'D'

water

1: If no gas was formed after 24 hours' incuba-
tion, but is present after 48 hours, the test is
regarded as doubtful and inconclusive. This
must be confirmed.

'D'

105
PROCEDURE FOR THE EXAMINATION OF WATER. (BACTERIOLOGICAL)

D

Steps in procedure

Further tests
required

To confirm the presence of the colon group in
a tube showing gas, streak out a loop of the
medium onto cosine methylene blue agar (or
Endo or litmus-lactose agar.) Incubate over-
night at 37° C.

(a) Presence of characteristic colonies on the
agar constitutes a Partially Confirmee
Test for the colon group. Recorc
whether the colony resembles the coli or
aerogenes sections.

(b) If no growth develops on the agar plate
it is considered that the gas produced in
the fermentation tube was due to anaero
bic organisms and not to coli-like forms
Record probably anaerobe.

(c) If characteristic colonies of Bact. coli
or its close allies are not present, the
plates must be examined further before
reporting absence of the colon group.

1. To determine whether Bact. coli or its close
allies is present on a negative or questionable
cosine methylene blue (Endo, or litmus-lac-
tose) agar confirmatory plate.

(a) Pick one or two colonies which most re-
semble Eact. coli or Bact. aerogenes and
plant into lactose broth (or lactose pep-
tone water), and incubate for 24 hours at
37° C. If 10% or more gas is formed,
record colon group present.

(b) A Gram stain should be made of the col-
ony before inoculation into lactose broth
to insure that a Gram-negative coli-like
form in pure culture is being fished.

Clark and Lubs medium or glucose broth may be
inoculated, if desired to test for the M. R. or V. P.
reaction, respectively.

None

None

'E'

THE COLON INDEX

Estimation of the Incidence of the Colon Group. It is apparent
that all agree in the use of the preliminary enrichment method for the de-
termination of the presence of the colon group but a moment's thought

106

will show that it is extremely difficult to estimate the number of colon
forms present from the result obtained. Perhaps the point can best be
illustrated by considering a specific example.

Suppose a sample of water was examined and gave the following re-
sult for colon types: 1 c. c.+ ; O.I c. c.+ ; and 0.01 c. c.— . What was the
incidence of the colon group per unit volume? Should 10 colon bacilli
present per c. c. be recorded? It might be said, as is commonly stated,
that there were more than 10 but less than 100 colon forms, and yet it is
conceivable that there might be less than 10. Assume that there were five
Bact. coli per c. c. In that case, in taking out a 0.1 c. c. sample the analyst
would be just as likely to catch a Bact. coli as to miss one. The mere
detection of the organism in a sample is not necessarily a safe criterion
for regarding the organism constantly present in that quantity of water.
On the other hand, the absence of colon forms in 0.01 c. c. is no justi-
fication for stating that such organisms would be absent if another 0.01
c. c. sample were taken, for if there were 50 colon bacilli present per c. c.,
the analyst would be just as likely to catch an organism as to miss one
in a single 0.01 c. c. sample. It is apparent, therefore, that from the an-
alysis presented it is extremely difficult to express by a single figure, the
number of colon types present per unit volume. If, however, instead of
having taken one portion of each dilution ten had been employed, a very
much closer approximation to the actual number of organisms could be
made. Similarly if the dilutions indicated were taken on ten different
days a reasonably close estimate could then be made of the average num-
ber of colon bacilli present during that period.

In water works operation, and for comparison of the efficiencies of
different plants, we are not concerned with a single analysis but with a
series of analyses extending over a long period, perhaps a month or a
year. A number of methods have been suggested for calculating the in-
cidence of the colon group or what is known as the "Colon Index." .

The most commonly employed method, and the one recommended by
the A. P. H. A. is that of Phelps. More complicated but probably more
accurate methods are described by McCrady, Wolman and Weaver, and
Stein.

The Phelps method is based on the assumption that the most probable
number of organisms present in any specimen is the reciprocal of the
highest positive dilution; thus in the above example (1 c. c.-f ; 0.1 c. c.+ ;
0.01 c. c.— ) 10 colon forms per cubic centimeter would be considered
the most probable number. To obtain the colon index for a month or a
year it is merely necessary to add the reciprocals of the highest positive
dilutions for the individual (daily or otherwise) tests and divide by the
total number of tests; this will give the average, but not necessarily the
most probable, number for the period under consideration. An example
follows:

107

Day

Results of Daily Test

Probable incidence
of colon group

1 c. c.

0.1 c. c.

0.01 c. c.

0.001 c. c.

i

3

1
1

+
+
+
+
+
+
+

+ 1 +++++

+

—

10
100
10
10
10
1
10

Total (for estimating averages) 151

Average of 7 tests 21.6 per c. c.

In 1915 McCrady proposed a method for the calculation of the most
probable number of colon bacilli from a series of fermentation results
but his methods have not been employed because of the cumbersome cal-
culations involved. Wolman and Weaver, in 1917, simplified the McCrady
formula and presented some graphs by means of which the tedious com-
putations are almost entirely removed.

If A=total number of tubes inoculated with 10 c. c.

B=total number of tubes inoculated with 1 c. c.

C=total number of tubes inoculated with 0.1 c. c.

a=number of 10 c. c. tubes positive

b= number of 1 c. c. tubes positive

c= number of 0.1 c. c. tubes positive

then x — the most probable number of colon bacilli per 100 c. c. — may
be obtained by trial substitutions of values for x in the following equation.

lOOa lOb Ic

100A - 10B - 1C=-

— .99X 1— ,999X

From the charts accompanying Wolman and Weaver's paper, the
values for x corresponding to any proportion of positive tests may be
read off directly when dealing with a single dilution. When concerned
with several dilutions the above formula must be employed, but the val-
ues of (1— .9X), (1— .99X), and (1— .999X) for any assumed value of
x are also given on these charts so that the actual mathematics involved
is reduced to merely simple arithmetic.

Stein in 1918, from a consideration of the laws of probability, evolved
a curve from which the colon index, together with its reliability could be
read directly if (1) the number of observations (2) the proportion of
positive tests, and (3) the size of the test samples, were known.

For further details the reader is referred to the original papers by
Stein, M. F. Journal of Bacteriology, IV, 1919, p. 243, and Wolman and
Weaver, Journal of Infectious Diseases, XXL, 1917, p. 287. In the Pub-
lic Health Journal (Canadian) IX., 1918, p. 201, McCrady presents a set
of tables for the interpretation of fermentation-tube results, which almost
completely relieve the analyst of mathematical calculations.

108
APPENDIX B.— CULTURE MEDIA

Numerous mediums have been utilized in the bacteriological examina-
tion of water. The preparation of the more important is described here.

I. ADJUSTMENT OF REACTION OF CULTURE MEDIA

(A. P. H. A. 1920)

1. Phenol Red Method for adjustment to a hydrogen-
ion concentration of PH+ = 6.8-8.4. Withdraw 5 c. c. of the
medium, dilute with 5 c. c. of distilled water, and add 5 drops of a solu-
tion of phenol red (phenol sulphone phthalein). This solution is made
by dissolving 0.04 grams of phenol red in 30 c. c. of alcohol and diluting
to 100 c. c. with distilled water.

Titrate with a 1:10 dilution of standard solution of NaOH (which
need not be of known normality) until the phenol red shows a slight but
distinct pink color. Calculate the amount of the standard NaOH solution
which must be added to the medium to reach this reaction. After the
addition check the reaction by adding 5 drops of phenol red to 5 c. c. of
the medium and 5 c. c. of water.

2. Titration with phenolphthalein. (For the convenience of
those who wish to retain the use of this method for the present it is given
here, but it is recommended that as soon as possible the more accurate meth-
od of determining the hydrogen-ion concentration be substituted.)

In a white porcelain dish put 5 c. c. of the medium to be tested, add
45 c. c. of distilled water. Boil briskly for one minute. Add 1 c. c. of
phenolphthalein solution (5 grams of commercial salt to one liter of 50
percent alcohol). Titrate immediately with a n/20 solution of sodium
hydrate. A faint but distinct pink color marks the true end point. This
color may be precisely described as a combination of 25 percent of red
(wave length approximately 658) with 75 percent of white as shown by
the disks of the standard color top made by the Milton Bradley Educational
Co., Springfield, Mass.

All reactions shall be expressed with reference to the phenolphthalein
neutral point and shall be stated in percentages of normal acid or alkali
solutions required to neutralize them. Alkaline media shall be recorded with
a minus ( — ) sign before the percentage of normal acid needed for their
neutralization and acid media with a plus ( + ) sign before the percentage
of normal alkali solution needed for their neutralization.

The standard reaction for culture media for water analysis shall be
+ 1.0 percent, as determined by tests of the sterilized medium. As ordin-
arily prepared, broth and agar will be found to have a reaction between
+0.5 and +1.0. For such media no adjustment shall be made. The re-
action of media containing sugar shall be neutral to phenolphthalein.
Whenever reactions other than the standard are used, it shall be so stated.

II. STANDARD STOCK MEDIA

A. Nutrient Agar (A. P. H. A., 1920). 1. Add 3 grams of beef
extract, 5 grams of peptone and 12 grams of agar, dried for one-half hour

109

at 105° C. before weighing, to 1,000 c. c. of distilled water. Boil over a
water bath until all agar is dissolved, and then make up the loss by
evaporation.

2. Cool to 45° C. in a cold water bath, then warm to 65° C. in the
same bath, without stirring.

3. Make up lost weight and adjust the reaction to a faint pink with
phenol red, or if the phenolphthalein titration is used, and the reaction
is not already between +0.5 and +1, adjust to +1.

4. Filter through cloth and cotton until clear.

5. Distribute in test-tubes, 10 c. c. to each tube, or in larger contain-
ers, as desired.

6. Sterilize in the autoclave at 15 pounds (120° C.) for 15 minutes
after the pressure reaches 15 pounds.

B. Nutrient Gelatin (A. P. H. A., 1920). 1. Add 3 grams of
beef extract and 5 grams of peptone to 1,000 c. c. of distilled water and add
100 grams of gelatin dried for one-half hour at 105° C. before weighing.

2. Heat slowly on a steam bath to 65° C until all gelatin is dissolved.*

3. Make up lost weight and adjust the reaction to a faint pink with
phenol red, or if the phenolphthalein titration is used, and the reaction
is not already between +0.5 and +1, adjust to +1.

4. Filter through cloth and cotton until clear.

5. Distribute in test-tubes, 10 c. c. to each tube, or in large con-
tainers as desired.

6. Sterilize in the autoclave at 15 pounds (120° C.) for 15 minutes
after the pressure reaches 15 pounds.

C. Nutrient Broth. (A. P. H. A., 1920). 1. Add 3 grams of
beef extract and 5 grams of peptone to 1,000 c. c. of distilled water.

2. Heat slowly on steam bath to at least 65° C.

3. Make up lost weight and adjust the reaction to a faint pink with
phenol red, or if the phenolphthalein titration is used, and the reaction is
not already between +0.5 and +1, adjust to +1.

4. Cool to 25° C. and filter through paper until clear. •

5. Distribute in test-tubes, 10 c. c. to each tube.

6. Sterilize in the autoclave at 15 pounds (120° C.) for 15 minutes
after the pressure reaches 15 pounds.

D. Media for Indol Test. (A. P. H. A., 1920). To 1,000 c. c.
of distilled water add 0.3 gram tryptophane, 5 grams dipotassium hydro-
gen phosphate (K2HP04), and 1 gram peptone. Heat until ingredients
are thoroughly dissolved, tube (6 to 8 c. c.), and sterilize in autoclave for
15 minutes after the pressure reaches 15 pounds. Some American peptones
are standardized to contain a uniform amount of tryptophane. If such
peptone is used the tryptophane in the above formula may be omitted and
the peptone increased to 5 grams.

E. Litmus Milk. (After Prescott and Winslow). The milk to be
used as a culture medium shall be as fresh as possible, "Certified Milk"
being ordinarily the best obtainable in city laboratories. It shall be placed

*The solution of the gelatin will be facilitated by allowing it . to soak in the cold
one-half hour before heating.

110

in a refrigerator over night to allow the cream to rise and the suspended
matter to settle. The skimmed milk shall be siphoned off into a flask for
use. It will be found more convenient, however, to allow the milk to stand
in a separatory funnel. Should the milk be too acid the reaction shall be
corrected to +1 by the addition of normal sodium hydrate. It is then ready
to be tubed and sterilized. Litmus milk shall be prepared as above, with
the addition of sterile 1 percent azolitmin. As it is impossible to make
each lot of litmus milk with the same shade of color, it is recommended
that a control tube be always exposed with the inoculated tubes for the
purposes of comparison.

III. MEDIA FOR PRELIMINARY ENRICHMENT OR THE PRE-
SUMPTIVE TEST.

A. Lactose Broth. (Standard Methods A. P. H. A., 1920). Sugar
broths shall be prepared in the same general manner as nutrient broth with
the addition of 0.5 percent of the required carbohydrate just before sterili-
zation. The removal of muscle sugar is unnecessary as the beef extract
and peptone are free from any fermentable carbohydrates. The reaction
of sugar broths shall be a faint pink with phenol red or, if on titration
with phenolphthalein the reaction is not already between neutral and +1,
adjust to neutral. Sterilization shall be in the autoclave at 15 pounds
(120° C.) for 15 minutes after the pressure reaches 15 pounds, provided
the total time of exposure to heat is not more than one-half hour; other-
wise a 10 percent solution of the required carbohydrate shall be made in
distilled water and sterilized at 100° C. for iy2 hours, and this solution
shall be added to sterile nutrient broth in amounts sufficient to make a
0.5 percent solution of the carbohydrate and the mixture shall then be
tubed and sterilized at 100° C. for 30 minutes, or it is permissible to add
by means of a sterile pipette directly to a tube of sterile neutral broth
enough of the carbohydrate to make the required 0.5 percent. The tubes so
made shall be incubated at 37° C. for 24 hours as a test for sterility.

B. Lactose (Peptone) Bile. The lactose bile medium consists
of sterilized undiluted fresh ox gall (or a 10 percent solution of dry fresh
ox gall) to which has been added 1 percent of peptone and 1 percent of
lactose. The addition of peptone is important.

C. Lactose Bile Salt Broth. (After Savage).

Sodium taurocholate 5 grammes

Lactose 5 grammes

Peptone 20 grammes

Water 1000 c. c.

These constituents are heated together until the solids are dissolved.
The mixture is filtered, and sufficient neutral litmus solution is added to
give a distinct color. The medium is then distributed into Durham's fer-
mentation tubes and sterilized by steaming for twenty minutes on three
successive days.

The sodium taurocholate prevents the growth of many saprophytic
bacteria.

Ill

The presence of fermenting organisms, including B. coli is shown when
the medium turns red (due to acid production) and gas is formed in the
inner tube.

D. Glucose Broth. Same as lactose broth, substituting glucose for
the disaccharid.

E. Liver Broth. (After Prescott and Winslow). 1. This
medium is made from a hot infusion of beef liver instead of fresh meat,
and is, in other respects, with the exception that phosphate is added the
same as dextrose broth, but it is a richer food medium for bacteria. It
gives gas formation with all species which ferment dextrose and develops
attenuated bacteria, whether gas-forming or not, to a better degree than
does beef broth. It is also especially suited to the rejuvenation of species
in pure culture.

Formula

Beef Liver 500 gm.

Peptone 10 gm.

Dextrose 10 gm.

Di-Potassium Phosphate (K9HP04) 1 gm.

Water . ' 1000 gm.

2. Chop 500 gm. of beef liver into small pieces and add 1000 c. c.
of distilled water. Weigh the infusion and container.

3. Boil slowly for 2 hours in a double boiler, starting cold and stir-
ring occasionally.

4. Make up any loss in weight by evaporation and pass through a
wire strainer.

5. To the nitrate add 10 gm. of peptone, 10 gm. of dextrose and 1
gm. of potassium phosphate.

6. After warming this mixture in a double boiler and stirring it for
a few minutes to dissolve the ingredients, titrate with N/20 sodium
hydrate, using phenolphthalein as an indicator, and neutralize with normal
sodium hydrate.

7. Boil vigorously for 30 minutes in a double boiler, and 5 minutes
over a free flame with constant stirring to prevent the caramelization of the
dextrose.

8. Make up the loss in weight by evaporation and filter through
cotton flannel and filter paper.

9. Tube and sterilize in an autoclave for 15 minutes at 120° C. (15
pounds).

The following media have been suggested for the elimination of spur-
ious presumptive test:

F. Gentian Violet Lactose Broth. (Hall and Ellefson). The
medium consists of 1 percent lactose broth containing 1-20,000 gentian vio-
let.

112

G. Brilliant Green Lactose Bile (Muer and Harris). The com-
position of the medium used is as follows:

Distilled water 1,000 grams

Ox gall (dried) 50 grams

Peptone 10 grams

Lactose 10 grams

Brilliant-green 0.1 grams

Directions for Preparation

1. Heat 1 liter of distilled water in double boiler until water in outer
vessel boils.

2. Add 50 grams of dried ox gall and 10 grams of peptone, stirring
until all ingredients are dissolved.

3. Continue boiling for one hour.

4. Remove from flame and add 10 grams of powdered lactose.

5. Filter through cotton flannel until clear.

6. To each liter of the filtrate add 10 c. c. of a 1 percent solution of
brilliant-green.

7. Tube and sterilize in autoclave for 15 minutes at 15 pounds pres-
sure.

IV. MEDIA FOR DIRECT ISOLATION OR CONFIRMATION OF
PRESUMPTIVE TEST.

A. Litmus Lactose Agar. (Wurtz Agar). (Standard Methods
A. P. H. A., 1920).

Litmus-lactose-agar shall be prepared in the same manner as nutrient
agar with the addition of 1 percent of lactose just before sterilization. The
reaction shall be a faint pink with phenol red, or, if on titration with
phenolphthalein the reaction is not already between neutral and +1, adjust
to neutral. One c. c. of sterilized litmus or azolitmin solution shall be
added to each 10 c. c. of the medium just before it is poured into the
petri dish, or the mixture may be made in the dish itself.

B. Fuchsin Sulphite (Endo) Agar. Endo agar consists of
nutrient lactose agar containing basic fuchsin decolorized with sodium
sulphite as an indicator. Many modifications have been described. Those
more commonly employed and method of use are listed below:

(a). Endo's Medium (Standard Methods A. P. H. A., 1920)

1. Add 5 grams of beef extract, 10 grams of peptone and 30 grams of
agar dried for one-half hour at 105° C. before weighing, to 1,100 c. c. of dis-
tilled water. Boil on a water bath until all the agar is dissolved and then
make up the loss by evaporation.

2. Cool the mixture to 45° C. in a cold water bath, then warm to
65° C. in the same bath without stirring.

3. Make up lost weight, titrate, and if the reaction is not already
between neutral and +1, adjust to neutral.

4. Filter through cloth and cotton until clear.

5. Distribute 100 c. c. or larger known quantities in flasks large
enough to hold the other ingredients which are to be added later.

113

6. Sterilize in the autoclave at 15 pounds (120° C.) for 15 minutes
after the pressure reaches 15 pounds.

7. Prepare a 10 percent solution of basic fuchsin in 95 percent
alcohol, allow to stand 20 hours, decant and filter the supernatant fluid.
This is a stock solution.

8. When ready to make plates melt 100 c. c. of agar in streaming
steam or on a waterbath. Dissolve 1 gram of lactose in 15 c. c. of dis-
tilled water, using heat if necessary. Dissolve 0.25 gram anhydrous sodium
sulphite in 10 c. c. water. To the sulphite solution add 0.5 c. c. of the
fuchsin stock solution. Add the fuchsin-sulphite solution to the lactose
solution and then add the resulting solution to the melted agar. The lac-
tose used must be chemically pure and the sulphite solution must be made
up fresh.

9. Pour plates and allow to harden thoroughly in the incubator be-
fore use.

(b) Endo's Medium (Hygienic Laboratory Modification). 1. The
Hygienic Laboratory-Endo medium consists of a 3 percent agar which is
titrated and corrected to +0.5 to phenolphthalein, to which is added 3.7
cubic centimeters of a 10 percent solution of anhydrous sodium carbonate
per liter. For convenience it is flasked, sterilized, and stored in 200 cubic
centimeter quantities. When ready to use, the following ingredients are
added to 200 cubic centimeters of agar as follows:

2. Dissolve 2 grams C. P. lactose in 25 to 30 cubic centimeters of
distilled water, with the aid of gentle heat.

3. Dissolve 0.5 gram of anhydrous sodium sulphite in 10 to 15 cubic
centimeters of distilled water.

4. To the sulphite solution add 1 cubic centimeter of saturated so-
lution of basic fuchsin in 95 percent alcohol.

5. Add the fuchsin-sulphite solution to the lactose solution, and
then add the whole to the agar. Pour plates at once and, after hardening,
dry for 15 minutes in the incubator.

(c) Endo's Medium (Med. Dept. U. S. Army Modification).

1. Into a container put 1 liter of tap water, marking the level of the
fluid. Add 30 grams of thread agar, 10 grams of peptone, 5 grams of NaCl, 5
grams of beef extract. Cook until dissolved — it is best to autoclave thirty
minutes at 15 pounds; filter through sterile gauze or cotton. If necessary
clear with egg. For this purpose, for each liter beat up the white of one
egg with 10 c. c. of warm water until the egg is well mixed. Add this to
agar cooled to 55° C., mix thoroughly, heat for 30 minutes or autoclave and
filter through cotton.

2. This stock agar is kept on hand in quarter-liter flasks or bottles.
Agar is standardized just before use and reaction adjusted to 0.2 percent
acid to phenolphthalein. Before use, fuchsin and sodium sulphite are
added. A filtered, saturated solution of basic fuchsin in 95 percent alcohol
is kept on hand. A 10 percent solution of dry sodium sulphite crystals
in sterile water is freshly made.

, 114

3. Teague has shown that a 10 percent solution of crystalline sodium
sulphite can be heated for twenty minutes at 15 pounds pressure with prac-
tically no change, and that the 10 percent sodium sulphite solution covered
with a layer of liquid petroleum about one cm. thick and sterilized in the
autoclave can be kept at room temperature for three weeks and probably
much longer with but very slight change.

4. One and eight-tenths c. c. of fuchsin solution is added per liter
to the agar. After this has been done the sodium sulphite solution is
added gradually until the hot agar is almost decolorized — usually about
25 c. c. to the liter. A pale rose color should be present in the hot agar,
which fades to a very faint pink on cooling; 10 grams of lactose is dis-
solved in a little water, filtered and added to each liter.

Various fuchsin solutions may differ and the absolute quantities given
above may not be exactly the proper balance in separate lots. These are
approximate, however, and the proper balance can easily be attained by a
little preliminary testing in which sodium sulphite solution is added to
small quantities of fuchsin solution in a test-tube.

The finished product is poured into large sterile Petri dishes. The
cover is left off until the agar is hard. Smears are made on these plates.

It is helpful to lay a piece of filter paper into the lid of the petri plate
in order to absorb liquid evaporating from the agar in the incubator. If
there is not enough filter paper for this, the plate should be placed up-
side down in the incubator.

(d) Endo's Medium (Kendal's Modification). (1) Preparation
of Agar. — (a) Prepare plain, sugar-free nutrient agar, using 15 grams of
agar per liter.

(b) Adjust the reaction to a point just alkaline to litmus.

(c) Flask the agar, 100 c. c. to a flask, and sterilize in the autoclave.

(2) Preparation of Indicator — (a) Prepare a 10 percent solu-
tion of basic fuchsin in 96 percent alcohol. This solution is fairly stable
if kept away from light.

(b) Prepare a 10 percent aqueous solution of chemically pure anhy-
drous sodium sulphite (1 gram in 10 c. c. water). This solution does not
keep.

(c) Add 1 c. c. of "2, a" to 10 c. c. of "2, b" and heat in the
Arnold sterilizer for 20 minutes. The color of the fuchsin is nearly dis-
charged if the solutions are of proper strength. This solution must be
prepared each day — it does not keep.

(3.) Preparation and Use of Endo medium — (a) Add 1 gram
of C. P. lactose (free from dextrose) to 100 c. c. of agar and place in the
autoclave until melted and the lactose is thoroughly dissolved.

(b) Add a sufficient volume of "2, c" (about 1 c. c.) to impart a
faint pink color to the medium.

(c) Pour into sterile Petri dishes and allow to harden in a dark place
with the covers partly removed. When cool the medium should be color-
less when viewed from above and a very faint pink when viewed from the

115

edge. The medium must be kept in a dark place because the color is re-
stored by the action of daylight.

(e) Endo's Medium (Robinson and Rettger's Modification).

Water 1,000 c. c.

Agar (powdered) 25 grams

Peptone (American brand) 10 grams

Meat extract (Liebig's) 5 grams

Sodium carbonate (10% sol.) 10 c. c.

Lactose, c. p 10 grams

Fuchsin (sat, alcoh. sol.) 5 c. c.

Anhydrous sodium bisulphite (10% sol.) 10 c. c.

Dissolve the agar, meat extract, and peptone. Make neutral to litmus
paper, steam in the autoclave at 12 to 15 pounds extra pressure for 35 to
40 minutes, filter through absorbent cotton and cheesecloth, add the sodium
carbonate solution, and heat for about 10 minutes in a boiling water bath.
Introduce the lactose and the fuchsin into the hot liquid. The medium
will now be brilliant red. Finally add the bisulphite solution. The hot
medium is light red in color, is filled into large test-tubes, 20 cubic cen-
timeters in each tube, and sterilized for five to seven minutes at 10 pounds
extra pressure. When the medium has cooled completely it should be
of a light pink or flesh color in the tubes, but transparent and practically
colorless in the large Petri dishes. The tubed agar may be kept for sev-
eral weeks in the refrigerator.

(f) A Simplified Endo's Medium (Levine) The medium is pre-
pared as follows:

Distilled Water 100 c. c.

Peptone (Difoo) 10 grams

Dipotassium phosphate (K2HP04) 2 to 5 grams

Agar 15 to 30 grams

The ingredients are boiled until dissolved and any loss due to evapora-
tion is made up with distilled water.

No adjustment of reaction is made and filtration is not necessary if
the medium is to be used for streak plate cultures.

Measured quantities are placed in flasks or bottles and sterilized for
15 minutes at 15 pounds.

For use, the agar prepared as above is melted and the following ma-
terials added to each 100 c. c. of medium.

20 percent lactose solution 1 gram or 5.0 c. c.

10 percent (saturated) alcoholic solution of basic fuchsin 0.5 c. c.

Freshly prepared 10 percent sodium sulphite solution 2.5 c. c.

Plates are poured, allowed to harden in the incubator, and inoculated
in the ordinary way.

(g) Fuchsin (Endo) Agar (Savage's Modification). 1. Peptone,
10 grams; Liebig's extract of beef, 10 grams; sodium chlorid, 5 grams, are
boiled up in an enamelled dish with 1 liter of distilled water. The mix-
ture is then poured into a flask, 30 grams of powdered agar added, and
the whole heated in the autoclave at 115° C. for one hour. The flask is

116

removed, and, after cooling to about 60° C., the white of one egg mixed
with a little distilled water is added. The contents are coagulated by
heating in current steam in the usual way, filtered, and the filtrate made
up to 1 liter. The mixture is made neutral, litmus paper being used as
the indicator. Then 19 c. c. of normal sodium carbonate solution and
10 grams of chemically pure lactose are added. The flask is replaced for
30 minutes in the steam sterilizer. Almost invariably there is a consider-
able precipitate, and the mixture has to be again filtered.

2. Seven c. c. of the fuchsin solution (see below) are added, fol-
lowed by 25 c. c. of a quite freshly prepared 10 percent sodium sulphite
solution. The mixture becomes much less red, but is not immediately
decolorized. It is then tubed, conveniently into small flasks, each con-
taining 50 to 60 c. c. of media, and sterilized in current steam for two
days, 30 minutes each day.

3. The fuchsin solution is made as follows: Three grams of pow-
dered crystalline fuchsin are placed in a dry flask, and 60 c. c. of abso-
lute alcohol are added. The contents are thoroughly mixed, and the flask,
tightly stoppered, allowed to stand for exactly 24 hours at 20° to 22° C.
The alcoholic extract is then decanted and preserved in a clean glass-stop-
pered bottle. Made in this way a uniform fuchsin extract is obtained
which keeps well, and the same quantity of fuchsin is added each time a
fresh batch of medium is prepared; a matter of much importance.

The medium must be stored in the dark, since light gradually turns
it red. When solidified it is almost free from color.

(h.) Conradi-Drigalski Agar (After Prescott and Winslow) These
authors have modified lactose litmus agar by adding to it nutrose and
crystal violet and by using three percent of agar instead of one percent.
The crystal violet strongly inhibits the growth of many other bacteria, es-
pecially cocci, which would also color the medium red; the 3 percent
agar makes the diffusion of the acid which is formed more difficult.

Three pounds of chopped beef are allowed to stand 24 hours with
two liters of water. The meat infusion is boiled one hour and filtered.
Twenty gm. of Witte's peptone, 20 gm. of nutrose, and 10 gm. of NaCl
are then added, and the mixture boiled another hour. After filtration
and the addition of 60 gm. of agar the mixture is boiled for three hours,
made alkaline and filtered. In the meantime 300 c. c. of litmus solution
(Kahlbaum) are boiled for 15 minutes with 30 gm. of lactose. Both
solutions are then mixed and the mixture, which is now red, made faintly
alkaline with 10 percent soda solution. To this feebly alkaline mixture
4 c. c. of hot sterile 10 percent soda solution are added and 20 c. c. of a
sterile solution (1 to 1000) of crystal violet (Hochst B.).

(i) Bile Salt (Rebipel) Agar (After Savage). Sodium tauro-
cholate 5 grams, Witte's peptone 20 grams, and distilled water 1 liter,
are boiled up together, 20 grams of agar are- added and dissolved in the
solution in the autoclave in the ordinary way. The medium is cleared with
white of egg and filtered. After filtration, 10 grams of lactose and 5

117

c. c. of recently prepared 1 percent neutral red solution are added. The
medium is then tubed and sterilized for 15 minutes on three successive
days.

(j) Aesculin Agar. (After Eyre) (B. coli and allied organisms
give black colonies surrounded by black halo.)

Measure out 400 c. c. distilled water into a tared 2-liter flask.
Weigh out

Agar 15 grams

Peptone 10 grams

Sodium taurocholate 5 grams

and make into a thick paste with 150 c. c. distilled water.
Add this paste to the distilled water in the flask.

Dissolve the ingredients by bubbling live steam through the mixture.
Weigh out

Aesculin 1.0 gram

Ferric citrate 0.5 gram

and dissolve in a second flask containing 100 c. c. distilled water.

Mix the contents of the two flasks — adjust the weight to the calculated
medium figure (in this case 1031.5 grams) by the addition of distilled
water at 100° C.

Clarify with egg and filter.

Tube and sterilize as for nutrient agar.

(k) A Simplified Eosine Methylene Blue Agar. (Levine)

Distilled water 1000 c. c.

Peptone (Difco) 10 grams

Dipotassium phosphate (K2HP04) 2 grams

Agar : 15 grams

Boil ingredients until dissolved and make up any loss due to evapora-
tion with distilled water.

Place measured-quantities (100 or 200 c. c.) in flasks or bottles, and
sterilize in the autoclave at 15 pounds pressure for 15 to 20 minutes.

Just prior to use add, to each 100 c. c. of the melted agar prepared
as above, the following:

Lactose, sterile 20% solution .5 c. c. or 1 gram dry substance

Eosine yellowish 2% aqueous sol. 2 c. c.

Methylene-blue 0.5% aqueous sol. 2 c. c.

Pour medium into perti dishes, allow them to harden, and inoculate
by streaking on the surface.

There is no adjustment of the reaction and filtration of the medium
is not necessary.

V. SPECIAL MEDIA FOR DIFFERENTIATION OF THE COLI
AND AEROGENES SECTIONS.

A. Clark and Lubs Medium. (Standard Methods A. P. H. A.,
1920). 1. To 800 c. c. of distilled water add 5 grams of Proteose-Pep-
tone, Difco., or Witte's Peptone (other peptones should not be substituted),
5 grams c. p. dextrose, and 5 grams dipotassium hydrogen phosphate

118

(K2HP04). A dilute solution of the K2HP04 should give a distinct
pink with phenolphthalein.

2. Heat with occasional stirring over steam for 20 minutes.

3. Filter through folded filter paper, cool to 20° C. and dilute to
1,000 c. c. with distilled water.

4. Distribute 10 c. c. portions in sterilized test tubes.

5. Sterilize by the intermittent method for 20 minutes on three suc-
cessive days.

B. Synthetic Medium. (After Clark and Lubs). (Standard
Methods A. P. H. A., 1920).

1. Na2HP04 (anhydrous) 7 grams

or

Na2HP04. 2H,0 8.8 grams

KHphthalate 2 grams

Asparatic acid 1 gram

Dextrose 4 grams

Warm distilled water 800 c. c.

2. When solution is complete, cool and make up to 1 liter at room
temperature.

3. Heat in an autoclave for 15 minutes after the pressure has reached
15 pounds, provided the total time of exposure to heat is not more than
one-half hour.

4. The hydrogen-ion concentration of the medium is fixed by the
composition. It should be very close to PH 7.0, slightly red with phenol
red. All materials should be re-crystallized or if used from stock fur-
nished by manufacturers, should be carefully examined. The di-sodium
hydrogen phosphate may be used either as the anhydrous salt obtained by
dessication in vacuo at 100° C. or else as the salt containing two mole-
cules of water of crystallization. This is obtained by exposing the re-
crystallized Na,HP0412H,0 for two weeks. Use 0.88 percent of Na,HP04-
2H20.

C. Uric Acid Medium (Koser)

Distilled amrnonia-free water 1,000 c. c.

NaCl 5.0 gm.

MgS04 0.2 gm.

CaCL, 0.1 gm.

K2HP04 1.0 gm.

Glycerol 30.1 gm.

Uric acid 0.5 gm.

This combination gives a colorless and clear medium. It is filled into
ordinary test tubes and sterilized in the autoclave at 13 to 15 pounds extra
pressure for 15 minutes. A slight turbidity may be apparent after auto-
claving, due, presumably, to a finely divided precipitate of calcium sul-
phate. On cooling, the solution becomes clear.

On the addition of 1.5 percent of washed shred agar to the solution
mentioned in the foregoing an agar medium was obtained on which the
same distinction between the two types may be brought out.

119
REFERENCES

1. Archibald, R. G., 1917. Lactose fermenting bacilli in surface water, feces, etc.

Wellcome Tropical Research Laboratories, 4th Report, 319.

2. Ayers, S. H., Rupp., 1918. Simultaneous acid and alkaline bacterial fermenta-

tion from dextrose and the salts of the organic acids, respectively, J. Infect.
Dis., XXIIL, 188.

3. Barber, M. A., 1908. The rate of multiplication of B. coli at different tempera-

tures. J. Infect. Dis. V., 379-400.

4. Bartow, E., 1916. Examination of drinking water on railway trains. Univ. 111.

Bull. No. 13. Water Survey Series 12 (1916), 173-182. Journal American
Water Works Association, 274-82.

5. Bergy, D. H. and Deehan, S. J., 1908. The colon-aerogenes group of bacteria.

J. Med. Research, XIX., 175.

6. * Bender, McCrady and Lafreniere, 1916. Occurance and significance of certain

types of B. coli in waters of the province of Quebec. Bull. San. Quebec, 16,
93-97.

7. Bowles, J. T. B., 1916. A study of the prevalence of B. coli in some tropical

surface waters of Panama and some waters of Mexico. Am. J. Pub. Health,
VI., 1173-78.

8. * Brown, A., 1906. Le roye neutra et le diagnostic rapide de la souillure des

eaux de boisson par le colibacille. Bull, de ITnst. Pasteur, IV., 56.

9. Browne, W. W., 1916. A comparison of acid production of B. coli group isolated

from various sources. Am. J. Pub. Health VI., 39-48.

10. Browne, Wm. W., 1914. The production of acid by the B. coli group. J. Infect.

Dis. XV., 580-608.

11. Browne, Wm. W., 1913. A comparative study of the Smith fermentation tube and

the inverted vial in the determination of sugar fermentation. Am. J. Pub.
Health, III., 701.

12. Burling, H. A. and Levine, M., 1918. Concentration of glucose and lactose and

viability of coli-like bacteria. Am. J. Pub. Health, VIII., 306-307.

13. Burton, L. V. and Rettger, L. F., 1917. Correlations in the colon- aerogenes

group. J. Infect. Dis. XXL, 162-95.

14. Caird, James M., 1919. The Operation of an American or rapid water filtration

plant for twenty years at Elmira, New York. J. Am. Water Works Assn. VI.,
409-21.

15. Chamot, E. M. and Sherwood, C. M.. 1917. Culture media employed in the

bacteriological examination of water. IV. Neutral red-lactose-peptone media.
J. Am. Chem. Soc, XXXIX., 1755-66.

16. Chen, Chen Chong and Rettger, Loe F., 1920. A correlation study of the colon-

aerogenes group of bacteria, with special reference to the organisms occurring
in the soil. J. Bact. V., 253-299.

17. * Chick, H. Distribution of bacterium coli commune. Thompson-Yates Laboratory

Reports, No. 3, 1.

18. * Clark, H. W. and Gage, S. DeM., 1900. Significance of the appearance of

B. coli communis in filtered water. J. of Boston Society of Med. Research,
IV., 178.

19. Clark, Wm. Mansfield and Lubs, Herbert A., 1915. The differentiation of bacteria

of the colon-aerogenes family by the use of indicators. J. Infect. Dis. XVIL,
160-173.

20. Clark, Wm. M., 1915. The final hydrogen-ion concentrations of cultures of

Bacillus coli. J. Biol. Chem. XXIL, 87.

.21. Clark, W. M. and Lubs, H. A., 1916. The colorimetric determination of the
hydrogen-ion concentration of bacteriological culture media. J. Wash. Acad.
Sci. VI., 483.

22. Clark, W. M. and 1 H H. A., 1917. Improved chemical methods for differ-

entiating bacteria 01 ' n coli-aerogenes family. J. Biol. Chem. XXX., 209-34.

23. Clark, W. M., 1918. Ditu Nation of bacteria of colon-aerogenes family. J.

Am. Water Works Assn. V. ^-35.

120

24. Clark, W. M. and Cohen, Barnett, 1918. Influences of the PH of media upon

the reproduction of some common bacteria. Abs. Bact., II., 10.

25. Clark, W. M., Lubs, H. A., Rogers, L. A., 1918. Characteristics of bacteria of

the colon type occurring in human feces. J. Bact. Ill, 231-52.

26. Clark, W. M. and Lubs, H. A., 1919. The colorimetric determination of the

hydrogen-ion concentration and its application in bacteriology. J. Bact. II.,
1 and 109.

27. Clemesha, W. W., 1912. Bacteriology of surface waters in the tropics. The

Calcutta and London.

28. Clemesha, W. W., 1912. A Criticism of Dr. A. C. Houston's report on the bi-

ological characters of B. coli isolated from (1) raw, (2) stored river water,
and (3) stored and filtered water. J. Hyg. XII., 463.

29. Cohen, B. and Clark, W. M., 1919. The growth of certain bacteria in media

of diffrent hydrogen-ion concentrations. J. Bact. IV., 409.

30. Copeland, W. R. and Hoover, C. P., 1911. The interpretation of tests for B. coli

communis. J. Infect. Dis. IX., 343.

31. Creel, R. H., 1914. Examination of drinking water on railway trains. Bull. 100,

Hyg. Lab. Wash.

32. Gumming, H. S., 1916. Investigations of the pollution and sanitary conditions

of the Potomac river water shed. Bull. 104, Hyg. Lab. Wash.

33. De Bord, Geo. G., 1917. The fuchsin aldehyde reaction on the Endo medium.

J. Bact. II., 309-314.

34. Dolt, M. L., 1908. Simple synthetic media for the growth of B. coli and for

its isolation from water. J. Infect. Dis. V., 616.

35. Durham, H. E., 1901. Some theoretical considerations upon the nature of ag-

glutinis, together with further observations upon Bacillus typhi abdominalis,
Bacillus enteritidis typhi abdominalis, Bacillus enteritidis, Bacillus coli com-
munis, Bacillus lactis aerogenes, and some other Bacilli of allied character.
J. Exper. Med. V., 353.

36. Dyer, H. G. and Keith, Jr. S. C., 1893. Notes on Normal intestinal Bacilli of

the horse and of certain other domesticated animals. Tech. Quarterly, VI.,
256-257.

37. Darling, Chester A., 1919. Characteristics of the colon type of bacteria found

in feces. Am. J. Pub. Health, Concord, IX., 884.

38. Eijkmann, 1904. Die Garungs probe bei 46<> als Hilfsmittel bei Trinkwasser

untersuchung. Central. Bakteriol. Orig. XXXVII., 74.

39. Ellms, Joseph W., 1920. Report of experiments on the purification of the water

supply of Milwaukee, Wisconsin.

40. Endo, S., 1903-4. Ueber ein Verfahren zum Nachweis der Typhus-bacillen.

Centralb. f. Bakt. I. Abt. XXXV., 109.

41. Escherich, J., 1885. Die Darmbakterien des Neugeborenen und Sauglings.

Fortschr. der Med. III., 515, 547.

42. Ewing, C. Leroy, 1919. Presence of a spore bearing aerobic gas forming bacillus

in Baltimore city drinking water. Am. J. Pub. Health, XX., 257-8.

43. Ferreira, A., Horta, A., and Paredes, C., 1908. Recherches sur le B. coli de

1'intestine des manniferes et des oiseaux. Archives do Real Institute Bacter-
iologico Camara Pestana, Tome II, Facs. II, 203.

44. Ferreira, A., Horta, A., and Paredes, C., 1908. Recherches sur le B. coli com-

munis de 1'intestin de I'homme. Archives do Real Institute Bacteriologico
Camara Pestana, Tome II, Fasc. II, 153.

45. Fitzgerald, J. G., 1914. A biometrical study of the mucosus capsulatus group.

J. Infect. Dis. XV., 208-78.

46. *Flint, J. M., 1896. Notes on the distribution of Bacillus coli communis. J.

Am. Med. Assn. XXVL, 410-411.

47. Ford, W. W., 1903. The classification and distribution of the intestinal bacteria

in man. Studies from the Royal Victoria Hos^ al I, also studies from the
Rockefeller Institute for Medical Research IT

48. Frost, W. H., 1916. Confirmation tests for . coli in routine water. Am. J.

Pub. Health, VI., 585-8.

121

49. Gage, S. D., 1901. Notes on testing for B. coli in water. J. Appl. Micros. IV.,

1403.

50. Gage, S. D. and Phelps, E. B., 1903. Notes on B. coli and allied forms with

special reference to the neutral red reaction. Transaction of the Am. Pub.
Health Assn., 1902 Meeting XXVIII., 402.

51. Gartner, A., 1910. Das Bacterium coli als Indikator fur fakale verurreinigung

eines Wassers. Zeitschrift fur Hygiene, LXVIL, 55.

52. Graf, A. V. and Nolte, A. G., 1916. Some practical Methods of bacteriological

water analysis and results obtained therewith. Am. J. Pub. Health, VI., 1109-19.

53. Greenfield, Myrtle, 1916. The presence of soil and fecal strains of organisms of

the colon-aerogenes group in the water of Kansas. J. of Inf. Dis. XIX., 647-
651.

54. Greenfield, Myrtle and Skourup, W. N., 1917. Seasonal distribution of soil and

fecal strains of coli-aerogenes group in surface waters. J. Ind. & Eng. Chem.
IX., 989-90: 675-678.

55. Hale, F. E. and Melia, T. W., 1910. Studies on inhibition, attenuation and re-

juvenation of B. coli. Am. J. Pub. Hyg. XX., 622.

56. Hale, F. E. and Melia, T. W., 1911. (The) Value of the esculin test. J. Am.

Pub. Health Assn. I., 833.

57. Hall, Ivan C. and Ellefson, L. J., 1918. The elimination of spurious presumptive

test for B. coli in water by use of gentian violet. J. Bact. III., 329-54.

58. Hall, I. C. and Ellefson, L. J., 1919. Further studies on gentian violet as a means

of eliminating spurious presumptive tests for B. coli in water. J. Am. Water
Works Assn. VI., 67-77.

59. Harden, A., 1901. Action of B. communis on dextrose. Trans, of the Chem-

ical Society (1901) 604.

60. Harden, A., 1905. Action on glucose of lactose fermenting bacteria of feces.

J. Hyg. V., 488.

61. Harden, A. and Walpole, S. G., 1905-1906. Chemical action of B. lactis aero-

genes (Escherich) on glucose. Production of 2-3, butylene-gycol and acetyl-
methyl-carbinol. Proc. Roy. Soc. Series B, LXXVIL, 399.

62. Harden, A., 1905-1906. On the Voges-Proskauer reaction for certain bacteria.

Proc. Roy. Soc. LXXVIL, 424.

63. Harden, A. and Norris, D., 1911. The diacetyl reaction for proteins. J. Phy-

siol. XLVII, 332.

64. Harden, A. and Norris, D., 1911-1912. Bacterial production of acetyl-methyl-

carbinol and 2-3 butylene-glycol from various substances. Proc. Roy. Soc.,
Series B. LXXXIV., 492; LXXXV., 73.

65. Hasseltine, H. E., 1917. The bacteriological examination of water. Pub. Health

Rep. 32, 1879.

66. Hauser, S. J., 1917. Methods used at Cincinnati filtration plant for isolating

and identifying Bacillus coli. Engineer. & Contr. XLVIIL, 296-297. Am.
J. Pub. Health VII., 1050-1051.

67. Hinds, M. E., 1916. The factors which influence longevity of B. coli and B. ty-

phosus in waters. Univ. of 111. Bull. Water Survey Series No. 13, 225-33.

68. Hinman, Jack J. Jr., 1920. Standards of quality of water. J. Am. Water Works

Assn. VII., 821.

69. Hiss, P. H., 1902. New and simple media for the differentiation of the colonies

of typhoid, colon and allied bacilli. J. Med. Research, VIII., 148.

70. Holmes, W. L, 1916. The use of decolorized acid fuchsin as an acid indicator

in carbohydrate fermentation tests with some remarks on acid production
by bacteria. J. Infect. Dis. XV., 227-232.

71. Horton, E. F., 1903. The Colon bacillus in ground waters. J. Hyg. III., 155.

72. Houston, A. C., 1901. Report on the chemical and bacteriological examination

of Chichester well-water. Supplement to the 13th Annual Report of the Local
Gov. Board containing the Report of the Medical Officer for 1900-01, 511.

73. Houston, A. C., 1903. Report on the bacteriological examination of the normal

stools of healthy persons. Supplement to the 32nd Annual Report of the Local
Government Board containing the Report of the Med. Officer for 1902-03, 511.

122

74. Houston, A. C., 1903. Report on the chemical and bacteriological examination

of Tunbridge wells deep well water. Supplement to the 32nd Annual Report
of the Local Government Board containing the Report of the Medical Officer
for 1902-03, 581.

75. Houston, A. C., 1904. The bacteriological examination of the river Thames and

the Thames Estuary. 4th Report of the Roy. Comm. on Sewage Disposal,

Ml.. J.«

76. Houston, A. C., 1904. Note on the absence of Bacillus coli (or coli-like microbes)

from one cubic centimeter to one thousand cubic centimeters of certain samples
of unpolluted or only slightly polluted water. J. Path, and Bact. X., 456.

77. Houston, A. C., 1905. Report on the bacteriological examination of (1) the

normal stools of healthy persons; (2) the intestinal contents of sea-fowls and
fish; and (3) certain of our public water supplies. Supplement to the 33rd
Annual Report of the Local Government Board containing the Report of
Medical Officer for 1904-05, 326.

78. Houston, A. C., 1908. The vitality of the typhoid bacillus in artificially infected

samples of raw Thames and Lee river waters with special reference to the
question of storage. First report on Research, Metropolitan Water Board,
London.

79. Houston, A. C., 1911. Seventh Report on Research Work, Metropolitan Water

Board, London.

80. Houston, A. C., 1913. Studies in water supply. London, MacMillan & Co.

81. Houston, A. C., 1899. Bacteriological examination of drinking water with par-

ticular reference to the relations of Streptococci and Staphylococci with water
of this class. Supplement to the 28th Annual Report of the Local Government
Board containing the Report of the Medical Officer for 1898-99, 439.

82. Houston, A. C., 1898. Chemical and bacteriological examination of soils with

reference to the amount and nature of organic matter and the number and char-
acter of the bacteria contained in them. Supplement to the 28th Annual
Report of the Local Government Board containing the Report of the Medical
Officer, 1898-99, 439.

83. Howe, E. C., 1912. A biometric investigation of certain non-spore-forming in-

testinal bacilli. Science N. S. XXXV., 225.

84. Hulton, Florence, 1916. The correlation of certain reactions of colon bacilli

and like organisms with source.. J. Infect. Dis. XIX., 606-613.

85. Irons, E. E., 1901. Some observations on methods for the detection of B. coli

communi in water. Trans. Am. Pub. Health Assn., 1900 Meeting XXVI., 310.

86. Jackson, D. D., 1906. A new solution for the presumptive test for Bacillus coli.

Biological studies by the pupils of William Thompson Sedgwick, Boston, 1906.

87. Jackson, D. D., 1907. The use of lactose-bile medium in water analysis. J.

Infect. Dis. Supplement No. 3, 30. Also Trans. Am. Pub. Health Assn., 1906
Meeting XXXII., Part II., 30.

88. Jackson, D. D. and Muer, T. C., 1911. Liver broth. J. Am. Pub. Health Assn.

I., 927.

89. Jackson, D. D., 1911. Classification of the B. coli group. J. Infect. Dis. VIII.,
241.

90. Johnson, B. R. and Levine, Max, 1917. Characteristics of coli-like microorgan-

isms from the soil. J. Bact. II., 379-401.

91. Jordan, E. 0., 1901. The relative abundance of Bacillus coli communis as an

index of the self-purification of streams. J. Hyg. I., 295.

92. Jordan, E. O., 1913. The inhibitive action of bile upon B. coli. J. Infect. Dis.

XII., 326-34.

93. Jordan, E. 0., Russell and Zeit, 1904. The longevity of typhoid bacillus in

water. J. Infect. Dis. I., 641.

94 Jordan, H. E., 1920. Fifteen years' filtration practice in Indianapolis. From
proc. of the 13th annual convention of the Indiana Sanitary and Water Sup-
ply Assn. Indianapolis Water Company.

95. Judson, E., 1919. Differentiation of B. coli. J. Am. Water Works Assn. VI.,
243-47.

123

96. Kahn, Reuben L., 1918. The differentiation of lactose-fermenting anaerobes

from B. coli. J. Bact. III., 561-64.

97. Kahn, Reuben L., 1918. Note on the nature of the reaction of B. coli on Endo's

medium. J. Bact. IK, 547-54.

98. Kahn, Reuben L., 1918. A simplified confirmatory test for B. coli. J. Bact.

III., 555-560.

99. Kahn, Reuben L., 1919. The examination of B. coli in water. Am. J. Pub.

Health IX., 333-36.

100. Keyes, F. G., 1909. The gas production of Bacillus coli. J. Med. Research XXI.,

69.

101. Keyes, F. G. and Gilespie, L. T., 1913. A contribution to our knowledge of

the gas metabolism of bacteria. J. Biol. Chem. XIII., 291.

102. Kendall, A. I. and Wakder, A. W., 1910. The isolation of Bacillus dysenteriae

from stools. J. Med. Research XXIII., 48.

103. Kellermann, K. F. and Scales, F. M., 1917. A preliminary report on the life

cycle of Bacillus coli. Abs. of Bact. L, 27.

104. Klein, E., 1898. Report on the morphology and biology of Bacillus enteritidis

sporogenes, on the association of this microbe with infantile diarrhoea and with
cholera nostras; and on its relation with milk, with sewage and with manure.
Supplement to the 27th Annual Report of the Local Government Board con-
taining the Report of the Medical Officer for 1897-98, 210.

105. Klein, E., 1901-2. Differentiation of the several anaerobic microbes commonly

present in the intestinal contents of man and other animals. Local Govern-
ment Board Medical Officer's Report, p. 404.

106. Kligler, I. J., 1914. Studies on the classification of the colon group. J.

Inf. Dis. XV., 187.

107. Konrich, 1910. Sm. Bewertung des Bacterium coli im Wasser. Kleinisches

Jahrbuch. XXIIL, 1.

108. Koser, S. A., 1918. The employment of uric acid synthetic medium for the

differentiation of B. coli and B. aerogenes. J. Infect. Dis. XXIIL, 377-379.

109. Laybourn, R. L., 1920. The fermentation of polysaccharids by Bacillus aero-

genes. J. Infect. Dis. XXVI., No. 5, 418-423.

110. Lemoigne, M., 1912. Fermentation du sucre par le Bacillus subtilis. Production

du 2.3-butylene-glycol. Compt. Rend. CLV., 792.

111. Letton, H. P., 1917. The Treasury Department Standard for drinking water;

its value and enforcement. J. Am. Water Works Assn. IV., 478-88.

112. Levine, Max, 1916. The correlation of the Voges-Proskauer and Methyl-Red

reaction in the colon-aerogenes group of bacteria. J. Infect. Dis. XVIII.,
358-367.

113. Levine, Max, 1916. Acid production and other characters of B. coli-like bacteria

from feces and sewage. J. Infect. Dis. XIX., 773-805.

114. Levine, Max, 1916. On the significance of the Voges-Proskauer reaction. J.

Bact. L, 153-164.

115. Levine, Max, 1917. Some differential reactions for coli-like bacteria. Am. J.

Pub. Health L, 784-785.

116. Levine, Max, Weldin, J. C. and Johnson, Beryl R., 1917. The Voges-Proskauer

and correlated reactions of coli-like bacteria. J. Infect. Dis. XXL, 39-46.

117. Levine, M., 1918. Presumptive test for Bacillus coli. Engin. Contr. XLIX., 34.

118. Levine, Max, 1918. A simplified fuchsin-sulphite (Endo) Agar. Abs. Bact.

1918, No. 38.

119. Levine, Max, 1918. Presumptive test for B. coli. J. Am. Water Works Assn.

V., 168-71. Engin. Contr. XLIX., 34.

120. Levine,. M., 1918. Differentiation of B. coli and B. aerogenes on a simplified

eosin-methylene blue agar. J. Infect. Dis. XXIIL, 43-47.

121. Levine, M. and Burling, H. A., 1918. Concentration of glucose and lactose and

viability of colon-like bacteria. Am. J. Pub. Health, VIIL, 306-307.

122. Levine, Max, 1918. A statistical classification of the colon-cloacae group. J.

Bact. III. 253-276.

124

123. Levine, Max, 1920. Experiences with the presumptive and confirmatory test for

B. coli in water analysis for the American Expeditionary Forces. Trans Am
Water Works Assn. VII., 188-92.

124. Longley, F. F. and Baton, W. V. C., 1907. Notes on the determination of B. coli

in water. J. Infect. Dis. IV., 397.

125. Marshall, W. W., 1907. The Para-dimethyl-amido-benzaldehyde test for indole.

J. Hyg. VII., 581.

126. Massachusetts State Board of Health, 1901. 32nd Annual Report, 1900.

127. Massachusetts State Board of Health, 1900. 31st Annual Report, 1899.

128. Massachusetts State Board of Health, 1899. 30th Annual Report, 1898.

129. Medical Research Committee, 1919. Report on the anaerobic infections of

wounds and the bacterioligical and serological problems arising therefrom.
Special reports Series No. 39, London.

130. Metcalf, H., 1905. Organisms on the surface on grains with special reference

to Bacillus coli. Science XXII., 434-441.

131. Meyer, E. M., 1917. The use of a 3% lactose-litmus agar plate for the demon-

stration of B. coli in water examination. J. Bact. II., 237-40.

132. Meyer, E. M., 1918. An aerobic spore-forming bacillus giving gas in lactose

broth isolated in routine water examination. J. Bact. III., 9-14.

133. *Moore, V. A. and Wright, F. R. A., 1900. A comparison of B. coli communis

from different species of animals. J. Boston Med. Soc. IV., 175.

134 Moore, V. A., and Wright, F. R. A., 1902. Preliminary observations on B. coli
communis from certain species of animals. Science n. 8. XV., 372.

135. Morse, R. B., and Wolman, A., 1918. The practicability of adopting standards

of quality for water supplies. J. Am. Water Works Assn. V., 198-228.

136. Mudge, Courtland S., 1917. The effect of sterilization on sugars in culture media.

J. Bact. II., 403-415.

137. Murray, J. J., 1916. A comparative study of colon bacilli isolated from horse,

cow and man. J. Infect. Dis. XIX., 161-174.

138. Muer, Theodore C. and Harris, Robt. L, 1920. Value of brilliant green in

eliminating errors due to anaerobes in the presumptive tests for B. coli. Am.
J. of Public Health, X., 874-6.

139. MacConkey, A., 1900. Experiments on the differentiation and isolation from

mixtures of the Bacillus coli communis and Bacillus typhosus by the use of
sugars and the salts of bile. The Thompson- Yates Laboratories Report HI,
151.

140. MacConkey, A. and Hill, C. A., 1901. Simple test for fecal contamination. The

Thompson-Yates Laboratories Report IV., 151.

141. MacConkey, A., 1905. Lactose-fermenting bacteria in faeces. J. Hyg. IX., 33.

142. MacConkey, A., 1909. Further observations on the differentiation of lactose-

fermenting bacilli with special reference to those of intestinal origin. J. Hyg.
IX., 86.

143. McCrady, M. H., 1918. The numerical interpretation of fermentation-tube re-

sults. J. Infect. Dis. XVII., 183-212.

144. Norton, J. F., 1918. A comparison of methods for the examination of water

at filtration plants. J. Infect. Dis. XXIII., 344-350.

145. Obst, M. M., 1916. Bile compared with lactose broth for determining the pres-

ence of B. coli in water. J. Bact. I., 73-79.

146. Orr, 1908, Report on an investigation as to the contamination of milk. Cited

by Wood, D.

147. Penfold, W. J., 1912. On- the specificity of bacterial mutations, with a resume of

the results of an examination of bacteria found in faeces and urine which
undergo mutation when grown on lactose media. J. Hyg. XII., 195.

148. *Pennington, M. E. and Kusel, G. C., 1900. An experimental study of the gas

producing power of Bacillus coli communis under different conditions of en-
vironment. Jr. Am. Chem. Soc. XXII., 556.

149. Perry, Margaret C. and Monfort, W. F., 1921 Some atypical colon aerogenes

forms isolated from natural waters. J. Bact. VI., 53-69.

150. Phelps, E. B., 1907. A method for calculating the numbers of B. coli from the

results of dilution tests. Am. J. Pub. Hyg. XVIII., 141.

151. Prescott, S. C, 1902. On the apparent identity of the cultural reactions of

B. coli communis and certain lactic bacteria. Science, N. S., XV., 363.

152. Prescott, S. C., 1902. A Note on methods of isolating colon bacilli. Science

N. S. XVI., 671.

153. Prescott, S. C., 1903. On certain precautions required in making and inter-

preting the so-called "Colon Test" for potable waters. Medicine XL, 20.

154. Prescott, S. C. ^and Baker, S. K., 1904. The cultural relations of bacillus coli

and Houston's sewage Streptococci and a method for the detection of these
organisms in polluted waters. J. Infect. Dis. I, 193. Also Trans. Am. Pub.
Health Assn., 1903, Meeting XXIX., 369.

155. Prescott, S. C., 1906. The occurrence of organisms of sanitary significance on

grains. Biol. Studies of Pupils of W. T. Sedgwich, 208-222, Boston.

156. Prescott, S. C. and Winslow, C. E. A., 1915. Elements of water bacteriology,

New York.

157. Purdy. W. C. and Butterfield, C. T., 1918. The effect of plankton animals upon

bacterial death rates. Am. J. of Pub. Health, VII., 499-505.

158. Race, J., 1914. Significance of B. coli in water examination. Am. J. Pub.

Health, XXVI., 758-759.

159. Race, J., 1914. Bacillus coli in water supply. Mun. Engin. XLVII., 12-14.

160. *Rettger, L. T., 1903. An experimental study of the chemical products of B.

coli and B. lactis aerogenes. Am. J. Physiol. VIIL, 284.

161. Rettger, Neva, 1919. The value of the presumptive test for B. coli based on the

routine use of lactose bile and lactose broth. J. Bact. IV., 609-14.

162. Robinson, H. C. and Rettger, L. F., 1916. Studies on the use of brilliant green

and a modified Endo's medium in the isolation of Bacillus typhosus from
feces. J. Med. Res. New Series, XXIX., 363.

163. Rogers, L. A., Clark, Wm. M. and Evans. A. D., 1914. The characteristics of

bacteria of the colon type found in bovine feces. J. Infect. Dis. XV., 100.

164. Rogers, L. A., Clark, Wm. M. and Davis, B. J., 1914. The colon group of bac-

teria. J. Infect. Dis. XIV., 411.

165. Rogers, L. A., Clark, Wm. M. and Evans, A. C., 1915. The characteristics of

bacteria of the colon type occurring on grains. J. Infect. Dis. XVII., 137-59.

166. Rogers, L. A., 1917. The viability of colon-aerogenes bacteria in water. Abst.

Bact. 59.

167. Rogers, L. A., Clark, Wm. M. and Lubs, H. A., 1918. The characteristics of

bacteria of the colon type occurring in human feces. J. Bact. III., 231-252.

168. Rogers, L. A., 1918. The occurrence of different types of the colon-aerogenes

group in water. J. Bact. III., 313-28.

169. Russell, H. L. and Fuller, C. A., 1906. The longevity of Bacillus typhosus in

natural waters and in sewage. J. of Infect. Dis. Supplement No. 2, 40. Also
Trans, of the Am. Pub. Health Assn. 1905 Meeting 31, Part II.

170. Salter, Raymond C., 1919. Observations on the growth of B. coli. J. Infect. Dis.

XXIV., 260-84.

171. Savage, W. G., 1905. Bacteriological examination of tidal mud as an index of

pollution of the river. J. Hyg. V., 146.

172. Savage, W. G., 1916. Bacteriological examination of food and water. University

Press, Cambridge, England, 2nd Ed. N. Y., G. P. Putman's Sons, 200 pp.

173. Savage, W. G. and Wood, D. R., 1917. The vitality and viability of Streptococci

in water. J. Hyg. XVL, 227-239.

174. Sawin, L. R., 1907. Experience with lactose-bile medium for the detection of

B. coli in water. J. Infect. Dis. Supplement No. 3, 33. Also Trans, of the
Am. Pub. Health Assn., 1906 Meeting XXXII., Part II, 33.

175. * Sawin, L. R., 1910. Pancreation-bile-salt medium for the detection of B. coli

in water. Am. J. of Public Hyg, XX., 672.

176. Smith, E. G. B., 1905. Notes on the occurrence on grain of organisms resem-

bling Bacillus coli communis. Science, XXL, 710-711.

126

177. Smith, T., 1893. The fermentation-tube with special reference to anaerobiosis

and gas production among bacteria. The Wilder Quarter-Century Book, 187.

178. Smith, T., 1893. A new method for determining quantitatively the pollution of

water by faecal bacteria. Thirteenth Annual Report of the State Board of
Health of New York for 1892, 712.

179. Smith, T., 1895. Ueber den Nachweis des Bacillus coli communis im Wasser

Centralb. f. Bakt. XVIIL, 494.

180. Smirnow, M. R., 1916. Biological variations of bacteria. Induced variations

in the cultural characters of B. coli. J. Bact. I. 385.

181. Stein, M. F., 1919. Critical study of the bacterial count in water and sewage.

Am. J. Pub. Health, VIII., 820-9.

182. Stein, M. F., 1919. The interpretation of B. coli test results on a numerical

and comparative basis as applied to the investigation of water and sewage. J.
Bact. IV., 243-265.

183. Stokes, W. R., 1904. A simple test for the routine detection of the colon bacillus

in drinking water. J. Infect. Dis. I., 341.

184. Stokes, W. R., 1919. The presence of fecal and non-fecal types of the colon

bacillus in various types of water. Am. J. Pub. Health, IX., 571-74.

185. Stovall, W. D., and Nichols, M. S., 1918. The methyl red and Voges-Proskauer

reactions with special reference to routine water analysis. J. Infect. Dis.
XXIII., 229-239.

186. Thompson, J., 1911. The chemical action of B. cloacae (Jordan) on glucose and

mannitol. Proc. Roy. Soc. (B) LXXXIV., 500.

187. Treasury Dept. Bacteriological standards for drinking water. Pub. Health

Reports XXIX, 2959-2966. J. Am. Water Works Assn. IV., 927-946.

188. Twort, F. W., 1907. The fermentation of glucosides by bacteria of the typhoid-

coli group and the acquisition of new fermenting powers by bacillus dysen-
teriae and other microorganisms. Proc. Roy. Soc. Series B, LXXIX., 329.

189. Voges and Proskauer, 1898. Beitrag zur Ernahrungs physiologic und zur differ-

ential diagnose der Bacterien der hamorrhagishen septicamie. Ztschr. f.
Hyg. XXVIII., 20.

190. Wagner, E. A., and Monfort, W. F., 1921. Lactose broth for isolating Bact. coli

from water.

191. Walpole, S. G., 1910-11. The action of B. lactis aerogenes on glucose and man-

nitol. The effect of free oxygen on their production. The action of B. lactis
aerogenes on fructose. Proc. Roy. Soc. (B), LXXXIIL, 272.

192. West, F. D., 1909. Notes on the Voges and Proskauer reactions for Bacillus

coli communis. Am. J. Pub. Hyg. XIX., 227.

193. *Whipple, G. C., 1903. On the practical value of presumptive tests for Bacillus

coli in water. Technol. Quart. XVI., 18.

194. Winslow, C. E. A. and Hunnewell, M. P., 1902. A study of the distribution of

the colon bacillus and of the sewage Strcptocci of Houston in polluted and un-
polluted waters. J. Med. Research, VIII., 502.

195. Winslow, C-E. A., 1903. The occurrence of the colon bacillus on the hands.

J. Med. Research, X., 463.

196. Winslow and Walker, 1907. Note on the fermentative reactions of the B. coli

group. Science N. S. XXVI., 797.

197. Winslow, A., 1916. Tests for B. coli as an indicator of water pollution. J. Am.

Water Works Assn. III., 927-39.

198. Winslow, C-E. A. and Cohen, Barnett, 1918. The relative viability of B. coli

and B. aerogenes types in water. J. Infect. Dis. XXIII., 82-89.

199. Winslow, C-E. A. and Cohen, Barnett, 1918. The distribution of B. coli and B.

aerogenes types in polluted and unpolluted water. J. Infect. Dis. XXIIL,
90-101.

200. Winslow, C-E. A., Kligler, I. J. and Rothberg, W., 1919. Studies on the classi-

fication of the colon typhoid group of bacteria with special reference to their
fermentative reactions. J. Bact. IV., No. 5, 429.

201. Wolman, Abel, 1917. The quality of water and confirmatory tests for B. coli. J.

Am. Water Works Assoc., IV., 200-5.

127

202. Wolman, A. and Weaver, H. L., 1917. A modification of the McCrady method

of the numerical interpretation of fermentation-tube results. J. Infect. Dis.
XXL, 287-91.

203. Wood, Denys R., 1919. Studies in the bacteriology of water. Inst. of Water

Engin., London.

204. Wood, Denys R., 1919. Recent advances in the differentiation of lactose fer-

menting bacilli with special reference to water and food products. J. Hyg.
XVHL, 46-58.

205. Yule, U., 1916. An Introduction to the Theory of Statistics. London.

*Original articles not available.

PRESS OF THE

NEWS PRINTING Co.
NEWTON, low A

GENERAL LIBRARY
UNIVERSITY OF CALIFORNIA—BERKELEY

RETURN TO DESK FROM WHICH BORROWED

This book is due on the last date stamped below, or on the

date to which renewed.
Renewed books are subject to immediate recall.

•i lasts-

Mar57DW
REC'D LD

31957

•

CLU BIO-MEO
NTERLIBRARY LOAN

OCT 221979

OCT 1

21-100m.l,'54(1887sl6)476

522564

UNIVERSITY OF CALIFORNIA LIBRARY