Full text of "Elements of water bacteriology, with special reference to sanitary water analysis"

See other formats

ELEMENTS

WATER BACTERIOLOGY

WITH SPECIAL REFERENCE TO

SANITARY WATER ANALYSIS

SAMUEL GATE PRESCOTT

Associate Professor of Industrial Microbiology in the Massachusetts

Institute of Technology

AND

CHARLES-EDWARD AMORY WINSLOW

Associate Professor of Biology, College of the City of New York, and Curator
of Public Health, American Museum of Natural History, New York

THIRD EDITION, REWRITTEN

FIRST THOUSAND

NEW YORK

JOHN WILEY & SONS, INC.

LONDON: CHAPMAN & HALL, LIMITED

1913

QTK 1 0-

CJDO .

BY
:S. C. PRESCOTT AND C.-E. A. WINSLOW

THE SCIENTIFIC PRESS

ROBERT DRUMMOND AND COMPANY

BROOKLYN. N. V.

DEDICATED

William

BY TWO OF HIS PUPILS,
AS A TOKEN OF GRATEFUL AFFECTION

PREFACE TO FIRST EDITION

THE general awakening of the community to the
importance of the arts of sanitation — accelerated by the
rapid growth of cities and the new problems of urban
life — demands new and accurate methods for the study
of the microbic world. Bacteriology has long since
ceased to be a subject of interest and importance to
the medical profession merely, but has become intimately
connected with the work of the chemist, the biologist,
and the engineer. To the sanitary engineer and the
public hygienist a knowledge of bacteriology is indis-
pensable.

In the swift development of this science during the
last ten years perhaps no branch of bacteriology has
made more notable progress than that which relates to
the sanitary examination of water. After a brief
period of extravagant anticipation, and an equally
unreasonable era of neglect and suspicion, the methods
of the practical water bacteriologist have gradually
made their way, until it is recognized that, on account
of their delicacy, their directness, and their certainty,
these methods now furnish the final criterion of the
sanitary condition of a potable water.

vi PREFACE TO FIRST EDITION

A knowledge of the new science early became so
indispensable for the sanitary expert that a special
course in the Bacteriology of Water and Sewage has for
some years been given to students of biology and sani-
tary engineering in the Biological Department of the
Massachusetts Institute of Technology. For workers
in this course the present volume has been especially
prepared, and it is fitting, we think, that such a manual
should proceed from an institution whose faculty,
graduates, and students have had a large share in shap-
ing the science and art of which it treats. We shall
be gratified, however, if its field of usefulness extends
to those following similar courses in other institutions,
or occupied professionally in sanitary work.

The treatment of the subject in the many treatises
on General Bacteriology and Medical Bacteriology is
neither special enough nor full enough for modern
needs. The classic work of Grace and Percy Frank-
land is now ten years old; and even Horrocks' valuable
" Bacteriological Examination of Water " requires to
be supplemented by an account of the developments
in quantitative analysis which have taken place on this
side of the Atlantic.

It is for us a matter of pride that Water Bacteriology
owes much of its value, both in exactness of method
and in common-sense interpretation, to American
sanitarians. The English have contributed researches
of the greatest importance on the significance of certain
intestinal bacteria; but with this exception the best
work on the bacteriology of water has, in our opinion,
been done in this country. Smith, Sedgwick, Fuller,

PREFACE TO FIRST EDITION vii

Whipple, Jordan, and their pupils and associates (not
to mention others) have been pioneers in the develop-
ment of this new field in sanitary science. To gather
the results of their work together in such form as to
give a correct idea of the best American practice is
the purpose of this little book; and this we have tried
to do with such completeness as shall render the volume
of value to the expert and at the same time with such
freedom from undue technicality as to make it reada-
ble for the layman. It should be distinctly understood
that students using it are supposed to have had before-
hand a thorough course in general bacteriology, and
to be equipped for advanced work in special lines.
BOSTON, March 10, 1904.

PREFACE TO THIRD EDITION

A SECOND edition of this work was called for in 1908
and it was rewritten in that year, with the inclusion
of much new material in the chapters dealing with
the isolation of the typhoid bacillus and of intestinal
bacteria, and with the addition of a new chapter on the
bacteriology of sewage and sewage effluents. In the
same year there appeared an excellent volume on
Water Bacteriology by Dr. W. G. Savage, which showed
the English methods of investigation and interpre-
tation to be closely in accord with those used in
America.

In the five years which have elapsed since our second
edition was published, there has again been important
progress along many lines in sanitary bacteriology;
and in particular the publication in 1912 of a new edition
of the Report of the Committee on Standard Methods
of the Laboratory Section of the American Public
Health Association has made necessary a change in
many details of current practice.

We have, therefore, prepared at this time a some-
what far-reaching revision of our book. Newer ideas

x PREFACE TO THIRD EDITION

on the effect of temperature upon the viability of
bacteria in water are included in Chapter I. The recent
recommendations of the Committee on Standard
Methods are discussed in Chapters II and IV; in
particular, Chapter IV, dealing with the 37° count,
has been expanded. We cannot bring ourselves to
agree with the recommendation of the committee that
the 37° count should replace the 20° count; but we are
entirely in accord with the resolution adopted by the
Laboratory Section of the American Public Health
Association at its Washington meeting that both
determinations should be made in ordinary routine
water examinations. Indeed, this is the position we
have maintained in both our earlier editions.

Chapter V, dealing with the isolation of specific
pathogenes from water, has been extensively rewritten
and extended. The use of the Jackson bile medium
for the preliminary enrichment of the typhoid bacillus
has become general since 1908 and a number of suc-
cessful isolations have been reported by its use; so that
this procedure promises to be of increasing importance
in the future.

In regard to the isolation and identification of
bacilli of the colon group we feel that the time has come
for a change from the usual American practice of the
past. The five standard tests for " typical B. coli "
established by the Committee on Standard Methods
in its 1905 report have come to seem more and more
illogical and unscientific to most practical water bac-
teriologists. The conviction has grown that they go
either too far or not far enough. For waters, in the

PREFACE TO THIED EDITION xi

United States, at least, it seems clear that all of the
lactose-fermenting group of bacilli are significant of
pollution from human or animal sources when present
in considerable numbers. The 1912 report of the
Standard Methods Committee apparently takes this
view in one place, while retaining the five tests in another
section. We have felt it best to place ourselves fairly
and fully in line with the view that the whole group
of lactose-fermenting bacilli is significant and that
the lactose bile fermentation test is a sufficient identi-
fication of the colon group for ordinary sanitary pur-
poses. This broad definition is the one upon which
we have based our general discussion of the colon
group in Chapters VI and VII. In Chapter VIII
we have discussed the subdivisions of the group as
worked out by MacConkey and others and their special
significance with respect to recent and remote pollu-
tion as suggested by the researches of Houston and
Clemesha.

The growing importance of the application of bac-
teriology to the sanitary study of shellfish has led
us to include a new chapter dealing with this subject,
based largely upon the recent report of the Committee
of the Laboratory Section of the American Public
Health Association.

Throughout the book we have resorted freely to
the use of tables of actual data for the illustration of
the various points discussed, believing that ample
familiarity with practical examples furnishes the only
sound basis for judgment in sanitary water exam-
ination.

xii PEEFACE TO THIKD EDITION

For the benefit of the student the chapters have
been sub-divided into sections with prominent headings
indicating the general topics under discussion.

Massachusetts Institute of Technology,
BOSTON, Mass.

College of the City of New York,
NEW YORK, N. Y.

June i, 1913.

TABLE OF CONTENTS

CHAPTER I

PAGE

THE BACTERIA IN NATURAL WATERS i

CHAPTER II

THE QUANTITATIVE BACTERIOLOGICAL EXAMINATION or WATER. . 29

CHAPTER III

THE INTERPRETATION OF THE QUANTITATIVE BACTERIOLOGICAL
EXAMINATION 51

CHAPTER IV

DETERMINATION OF THE NUMBER OF ORGANISMS DEVELOPING AT
THE BODY TEMPERATURE 61

CHAPTER V
THE ISOLATION OF SPECIFIC PATHOGENES FROM WATER 74

CHAPTER VI

THE COLON GROUP OF BACILLI AND METHODS FOR THEIR ISOLATION 99

xiii

xiv TABLE OF CONTENTS

CHAPTER VII

PAGE

SIGNIFICANCE or THE PRESENCE OF THE COLON GROUP IN WATER. 140

CHAPTER VIII

VARIETIES OF COLON BACILLI AND THEIR SPECIAL SIGNIFICANCE. 174

CHAPTER IX
OTHER INTESTINAL BACTERIA 201

CHAPTER X

THE SIGNIFICANCE AND APPLICABILITY OF THE BACTERIOLOGICAL
EXAMINATION 215

CHAPTER XI

BACTERIOLOGY OF SEWAGE AND SEWAGE EFFLUENTS 228

CHAPTER XII

BACTERIOLOGICAL EXAMINATION OF SHELLFISH 244

APPENDIX 265

ELEMENTS OF WATER BACTERIOLOGY

CHAPTER I
THE BACTERIA IN NATURAL WATERS

Bacteria and Their Nutritive Relations. Bacteria are
the most numerous and the most widely distributed
of living things. They are present not merely at the
surface of the earth or in the bodies of water which
partially cover it, as is the case with most other living
things, but in the soil itself, and in the air above, and
in the waters under the earth.

Probably no organisms are more sensitive to external
conditions, and none respond more quickly to slight
changes in their environment. Temperature, moisture,
and oxygen are of importance in controlling their distri-
bution; but the most significant factor is the amount
of food supply. Bacteria and decomposing organic
matter are always associated, and for this reason a
brief consideration of the general relation of bacteria
to their sources of food supply must precede the study
of their distribution in any special medium.

The bacteria possess greater constructive ability
than any animal organisms. They lack, however,

2 ELEMENTS OF WATER BACTERIOLOGY

the power of green plants to build up their own food
from compounds like carbon dioxide and nitrates which
have no stored potential energy. The food require-
ments of various bacterial types differ, however, widely
among themselves. Fischer (1900) has divided the
whole group into three great subdivisions according
to the nature of their metabolism. The Proto trophic
forms are characterized by minimal nutrient require-
ments, including organisms like the nitrifying bacteria
which require no organic compounds at all, but derive
their nourishment from carbon dioxide or carbonates,
nitrites and phosphates, or from inorganic ammonium
salts. A second group of Metatrophic bacteria includes
those forms which require organic matter, nitrogenous
and carbonaceous, but are not dependent on the fluids
of the living plant or animal. Finally, the Para trophic
bacteria are the true parasites, which exist only within
the living tissues of other organisms. These sub-
divisions, like all groups among the lower plants, are
not sharply defined, and the metatrophic bacteria in
particular exhibit every gradation, from types which
grow in water with a trace of free ammonia to organisms
like the colon bacillus which normally occur on the
surface of the plant or animal body, feeding upon the
fluids or on the extraneous material collected upon its
surface.

The vast majority of bacteria belong to the second,
or metatrophic group, living as saprophytes on dead
organic matter wherever it may occur in nature, and
particularly in that diffuse layer of decomposing plant
and animal material which we call the humus, or surface

THE BACTERIA IN NATURAL WATERS 3

layer of the soil. Wherever there is life, waste matter
is constantly being produced, and this finds its way
to the earth or to some body of water. The excretions
of animals, the dead tissues and broken-down cells of
both animals and plants, as well as the wastes of domestic
and industrial life, all eventually find their way to the
soil. In a majority of cases these substances are not
of such chemical composition that they can be utilized
at once by green plants as food, but it is first necessary
that they go through a decomposition or transforma-
tion in which their chemical nature becomes changed;
and it is as the agents of this transformation that
bacteria assume their greatest importance in the world
of life.

We may take the decomposition of a comparatively
simple excretory product, urea, as an example of the
part which the bacteria play in the preparation of
plant food. Through the activity of an enzyme pro-
duced by certain bacteria this compound unites with
two molecules of water and is converted into ammonium
carbonate,

+ 2H20 = (NH4)2C03.
NH2

This, however, is only part of the process. While
green plants can derive their necessary nitrogen in part,
at least, from ammonium compounds it is a well-
established fact that this element is often obtained
more readily from nitrates, and there are other bacteria
which as a further step oxidize the ammoniacal nitro-

4 ELEMENTS OF WATEE BACTERIOLOGY

gen to a more available form. This process of oxida-
tion is known as nitrification, and takes place in a suc-
cession of steps, the organic nitrogen being first con-
verted to the form of ammonium salts, and these in
turn to nitrites and nitrates, the oxygen used coming
from the air. Several groups of organisms are instru-
mental in bringing about this conversion. It is gen-
erally assumed that one group attacks the ammonium
compounds and changes them to nitrites; while another
group completes the oxidation to nitrates. In the
latter form nitrogen is readily taken up by green plants
to be built up into more complex albuminoid sub-
stances (organic nitrogen) through the constructive
power of chlorophyll.

This never-ending cycle is illustrated in the accom-
panying figure, devised by Sedgwick (Sedgwick, 1889) to
illustrate the transformations of organic nitrogen in
nature, the increasing size and closeness of the spiral
on the left-hand side indicating the progressive com-
plexity of organic matter as built up by the chlorophyll
bodies of green plants in the sunlight, and the other
half of the figure the reverse process, carried out largely
by the bacteria. In nature there are many short
circuits, as, for instance, when dead organic matter
is used as food for animals and built up into the living
state again without being nitrified and acted upon
by green plants; but the complete cycle of organic
nitrogen is as indicated on the diagram.

We have dwelt thus at length upon the general
relation between bacteria and organic decomposition
because in this relation will be found the master key

THE BACTEEIA IN NATURAL WATERS 5

to the distribution of bacteria in water as well as in
other natural habitats. It is true that certain peculiar
forms may at times multiply in fairly pure waters; but,
in general, large numbers of bacteria are found only in
connection with the organic matter upon which they
feed. Such organic matter is particularly abundant
in the surface layer of the soil. Here, therefore, the
bacteria are most numerous; and in other media their

THE SPHERE

OF
ORGANISMS

AND
THE HISTORY

OF
ORGANIC MATTER,

numbers vary according to the extent of contact with
the living earth.

Classification of Waters. Natural waters, then,
group themselves from a bacteriological standpoint
in four well-marked classes, according to their relation
to the rich layers of bacterial growth upon the surface of
the globe. There are first the atmospheric waters which
have never been subject to contact with the earth;
second, the surface-waters immediately exposed to such

6 ELEMENTS OF WATER BACTERIOLOGY

contamination in streams and pools; third, stored waters,
the lakes and large ponds in which storage has reduced
bacterial numbers to a state of comparative purity;
and fourth, the ground-waters from which previous
contamination has been even more completely removed
by filtration through the deeper layers of the soil.

Bacterial Content of Various Waters. Even rain and
snow, the sources of our potable waters, are by no
means free from germs, but contain them in numbers
varying according to the amount of dust present in
the air at the time of the precipitation. After a long-
continued storm the atmosphere is washed nearly free
of bacteria, so that a considerable series of sterile plates
may often be obtained by plating i-c.c. samples. These
results are in harmony with the observations of Tissandier
(reported by Duclaux, 1897), wno found that the dust
in the air amounted to 23 mg. per cubic meter in Paris
and 4 mg. in the open country. After a rainfall these
figures were reduced to 6 mg. and 0.25 mg., respectively.

With regard to what may be considered normal values
for rain it is difficult to give satisfactory figures. Those
obtained by Miquel (Miquel, 1886) during the period
1883-1886 showed on the average 4.3 bacteria per c.c.
in the country (Montsouris) and 19 per c.c. in Paris.
Snow shows rather higher numbers than rain. Janowski
(Janowski, 1888) found in freshly fallen snow from 34
to 463 bacteria per c.c. of snow-water.

As soon as the rain-drop touches the surface of the
earth its real bacterial contamination begins. Rivulets
from ploughed land or roadways may often contain
several hundred thousand bacteria to the cubic centi-

THE BACTERIA IN NATURAL WATERS 7

meter; and furthermore the amounts of organic and
mineral matters which serve as food materials, and thus
become a factor in later multiplication of organisms,
are greatly increased.

In the larger streams several conditions combine to
make these enormous bacterial numbers somewhat
lower. Ground-water containing little microbic life
enters as a diluting factor from below. The larger
particles of organic matter are removed from the flow-
ing water by sedimentation; many earth bacteria,
for which water is an unfavorable medium, gradually
perish; and in general a new condition of equilibrium
tends to be established. It is difficult, however, to
find a river in inhabited regions which does not con-
tain several hundreds or thousands of bacteria to the
cubic centimeter. Furthermore, heavy rains which
introduce wash from the surrounding watershed may
at any time upset whatever equilibrium exists, and
surface-waters are apt to show sudden fluctuations
in their bacterial content.

Seasonal Variation of Bacteria in Surface Waters.
Sharp variations in bacterial content are particularly
apt to occur in the spring and fall as a result of the
rain and melting snow at those seasons. The high
numbers shown for various rivers in the table on page
8 illustrate this point.

The rainfall is the main factor which causes these sea-
sonal variations; but its specific effect differs with dif-
ferent streams. The immediate result of a smart shower
is always to increase contamination by introducing fresh
wash from the surface of the ground. More prolonged

ELEMENTS OF WATER BACTERIOLOGY

SEASONAL VARIATIONS IN BACTERIAL CONTENT OF
RIVER WATERS. BACTERIA PER C.C. MONTHLY
AVERAGE

River.

Year.

Jan.

Feb.

Mar.

April.

May.

June.

Thames x

1905-6

2075

1,679

1,161

277

I 064

Lea1

1905-6

5,192

3,083

1,308

471

1,350

598

New 1

1905-6

I 455

I 304

291

I4.Q

-? cr 2

108

Mississippi 2 . . . .

1900-01

972

2,871

i,795

3,597

2,152

2,007

Potomac 3

1906-7

4,400

I,OOO

11,500

3,700

7 co

77 300

Merrimac 4

14,200

14.800

10,300

3,600

1,900

9,600

Susquehanna 5. .

1906

9,510

21,228

31,326

39,905

6,187

2,903

River.

Year.

July.

/ ug.

Sept.

Oct.

Nov.

Dec.

Thames 1

190=5—6

952

I 633

74-O

Lea1

1905-6

1,190

7,946

2.CKO

New1

1905-6

450

718

621

Mississippi 2 . . . .

1900-01

1,832

805

2,021

Potomac 3

1906-7

2,700

3,000

6,2OO

2,300

1, 800

6,9OO

Merrimac 4

1905

3,900

19,500

I3-500

39,800

8,700

Susquehanna $ .

1906

685

1,637

836

7,575

26,224

37,525

1 Houston, 19060, 19066. 4 Massachusetts, 1906.

2 New Orleans, 1903. 5 Harrisburg, 1907.

3 Figures obtained through courtesy of F. F. Longley.

moderate rain, however, exerts an opposite effect, and
after the main impurities which can be washed away
have been removed, may dilute the stream with water
purer than itself. What the net effect of rain may be
depends, therefore, on the character of the stream.
A river of fairly good quality shows its highest numbers
in rainy periods. With a highly polluted stream, on
the other hand, the constant influx of sewage over-
balances occasional contributions of surface contamina-
tion. Thus Gage (1906) shows in the following table
that the bacterial content of the Merrimac is highest
when the stream is lowest, that is, when its sewage
content is least subject to dilution.

THE BACTERIA IN NATURAL WATERS

VARIATIONS IN BACTERIAL CONTENT, MERRIMAC RIVER
GAGE (1906)

Flow of St-eam. Cubic

Bacteria

per c.c.

B. coli

per c.c.

Square Mile of Watershed

Canal.

ntake.

Canal.

Intake.

Less than i

7,500

10,800

.88

1—2

6 800

6 200

2-4

3,600

5,6oo

Over 4

3 AOO

2 IOO

The contrast between the two classes of rivers is well
brought out in a study of the Lahn and the Wieseck,
by Kisskalt (1906); and the table below, compiled from
his data, gives an excellent idea of the total numbers of
bacteria and their seasonal fluctuations in a stream of
fair quality (the Lahn) and a highly polluted one (the
Wieseck). In the former case the bacterial numbers
are highest when rain brings surface pollution; in the lat-
ter, when the sewage constantly present is least diluted.

MONTHLY VARIATIONS OF BACTERIA IN A NORMAL AND
POLLUTED STREAM

KISSKALT, 1906

Date.

Bacteria

per c.c.

Date.

Bacteria

per c.c.

1904.

Lahn.

Wieseck.

1904-5-

Lahn.

Wieseck.

3l8

104 ooo

December *

I 22O

21 2OO

July
August
October l . . .
October l . . .

132

840

1,235
420

156,800

98,400
28,400
58,000

January l . .
February 1 .
March1....
April !

3,668
5,38o

I,2IO
4,025

29,920
1 1 ,9OO
8,250
5,QIO

November

2 34O

7Q 2OO

May

C7o

14 800

November l .

I,74O

5 2 ,000

June . ,

686

^o 180

December l . .

780

28,600

1 Rain or high water due to previous thaw.

ELEMENTS OF WATER BACTERIOLOGY

Effect of Storage upon Bacteria in Water. In stand-
ing waters all the tendencies which make for the reduc-
tion of bacteria are intensified, and when a river
passes into a natural or artificial reservoir a notable
reduction in numbers occurs. The table below shows
the striking effect produced upon the water of the
Potomac River by its successive passage through the
three reservoirs of the Washington water supply in
the first nine months of 1907. We owe these figures
to the courtesy of Mr. F. F. Longley, the engineer then
in charge of the Washington filter plant.

REDUCTION OF BACTERIA IN WASHINGTON RESERVOIRS.
BACTERIA PER C.C., MONTHLY AVERAGE, 1907

Potomac
River.

Dalecarlia
Reservoir.

Georgetown
Reservoir.

Washington
City Reservoir.

January
February ....
March .

4,400
I,OOO
II 500

2,400

95°
8 300

2,2OO
I,OOO
7 2OO

950

750
•2 600

April

3.700

2 IOO

I 4OO

47 ^

M!ay

7 ^O

•7 tro

•2 2 C

June
July . .

2,300

2 7OO

950
600

600
3^O

IOO
1 60

August
September . . .

3,000
6. 200

275

425
I.QOO

230

The still more striking results obtained at London
are indicated in the table on page n.

When the water which enters a pond or a reservoir
has already undergone considerable storage and reached
a comparatively stable condition, the diminution due to
additional storage may be almost negligible. Thus
Philbrick (1905) found that the influent water of the
Chestnut Hill Reservoir of the Metropolitan Water

THE BACTEEIA IN NATURAL WATERS

Works of Boston contained on the average during the
eleven years, 1893-1903, 220 bacteria per c.c., and the
effluent 179. In many individual months, and in some
whole years, the effluent contained more than the
influent.

AVERAGE REDUCTION OF BACTERIA BY STORAGE AT
LONDON

(HOUSTON, 1909)

Water

Storage,
Days.

Bacteria per c.c.

Gelatin
20°.

Agar
37°.

Bile-salt
Agar 37°.

Raw Thames River

4405
175
208
362
8135
67

280
34
44
52
382
II

5
8

Do. stored at Staines

95
15
14

"5s"

Do stored at Chelsea

Do. stored at Lambeth
Raw Lee River

Do. stored

The seasonal variations in the bacterial content of a
large pond or lake follow a somewhat different course
from those observed in a stream. Philbrick, in , the
paper just cited, gives the figures tabulated below for
the Chestnut Hill Reservoir of the Metropolitan Water
Works (Boston). The averages are based on weekly
analyses covering the eleven years, 1893-1903.

MONTHLY VARIATIONS IN BACTERIAL CONTENT OF
CHESTNUT HILL RESERVOIR, 1893-1903

Month.

0. | K.

Bacteria

per c.c.

123

134

103

ELEMENTS OF WATER BACTERIOLOGY

The marked increase in April and September is the
notable feature of these analyses; and this is due to the
effect of the spring and fall overturns which, in the
months in question, stir up the decomposing organic
matter at the bottom and distribute it through the
reservoir. The slight, but steady, increase during the
warm months from May to August is also probably
significant.

On the whole it may be said that the bacterial content
of a lake or pond should not be more than one or two
hundred per c.c. and may often be under a hundred.
The student will find numerous analyses of natural
waters in Frankland's classic work (Frankland, 1894).
He notes, for example, that the Lake of Lucerne con-
tained 8 to 51 bacteria per c.c., Loch Katrine 74, and
the Loch of Lintralthen an average of 170. The water
of Lake Champlain examined by one of us (S. C. P.) in
1896 contained on an average 82 bacteria per c.c. at a
point more than two miles out from the city of Burling-
ton. Certain surface water-supplies near Boston studied
by Nibecker and one of us (Winslow and Nibecker, 1903),
gave the following results:

City.

Number of
Samp es.

Average Number
of Bacteria per c.c.

Wakefield

Lynn

Plymouth

•2 r

Cambridge

Salem

222

Medford . .

C.24.

Taunton

Peabody

I4.I

THE BACTEEIA IN NATURAL WATERS

In sea-water, too, bacterial numbers are small, as
noted by Russell at Naples (Russell, 1891) and Wood's
Hole (Russell, 1892), and in salt as in fresh water the
amount of bacterial life decreases in general as one
passes downward from the surface and outward from
the shore. Otto and Neumann (1904) obtained the
results summarized below at various points on the
high seas between Portugal and Brazil. Near the
European coast numbers were much higher.

BACTERIA IN THE ATLANTIC OCEAN. (OTTO AND
NEUMANN, 1904.) BACTERIA PER C.C.

Nearest Land.

Depth ir

Meters.

IOO

2OO

Canary Islands

1 20

Cape Verde Islands

St Paul Island

480

Pernambuco

168

Drew (1912) finds high numbers of bacteria in surface
sea- water off the Bahamas, ranging from 13,000 to
16,000, falling off below 200 fathoms (in the cold bottom
waters at 10° C. or below) to o to 17.

Factors Influencing the Diminution of Bacteria in
Surface-waters. The decrease in numbers which takes
place when a surface-water is stored in a pond or reservoir
indicates that the forces which tend to produce bacterial
self-purification are important ones. It is necessary
to consider in somewhat more detail just what these
forces are, in order to gauge their potency in any
particular instance.

14 ELEMENTS OF WATER BACTERIOLOGY

Chief of them appear to be sedimentation, the activ-
ity of other micro-organisms, light, temperature, and
food-supply, and perhaps more obscure conditions such
as osmotic pressure.

The subsidence of bacteria, either by virtue of their
own specific gravity, or as the result of their attachment
to particles of suspended matter, is unquestionably
partly, if not largely, responsible for changes in the
number of bacteria in the upper layers of water at rest
or in very sluggish streams. The results of numerous
investigations by different workers seem to indicate that
sedimentation of the bacteria themselves takes place
slowly, and that the difference in numbers between
the top layer and the bottom layer of water in tall
jars in laboratory experiments of a few days' duration
is very slight or quite within the limits of experimental
error (Tiemann and Gartner, 1889). Different species
may, of course, be differently affected (Scheurlen,
1891). It must be remembered, however, that in
natural streams bacteria are to a great extent attached
to larger solid particles upon which the action of gravity
is more important. Spitta (1903) found that from one-
fifth to one-half of the bacteria in canal water may
be attached to gross particles, as evidenced by their
sedimentation in a few hours. Jordan (Jordan, 1900)
is firmly of the opinion that in the lower part of the
Illinois River, where there is a fall of but 30 feet in
225 miles, the influences summed up by the term
sedimentation are sufficiently powerful to obviate the
necessity for summoning another cause " to explain
the diminution n\ numbers of bacteria," and he further

THE BACTERIA IN NATURAL WATERS 15

adds: " It is noteworthy that all the instances recorded
in the literature where a marked bacterial purification
has been observed are precisely those where the con-
ditions have been most favorable for sedimentation."

Little is known as to the share of other organisms in
hastening the decrease of bacteria in stored water.
Doubtless predatory Protozoa play some part in the
process. Huntemiiller (1905) after infecting water
containing flagellate Protozoa with typhoid bacilli,
found the Protozoa crowded with bacteria; and he
observed under the microscope the actual ingestion
of the living and motile bacilli. Korschun (1907) and
others have obtained similar results and consider the
activity of Protozoa to be an important factor in self-
purification. Fehrs (1906) found that typhoid bacilli
would live for 7 days in unsterilized Goltingen tap water,
for 46 days in the same water sterilized, and for 13
days in water inoculated with a culture of flagellate
Protozoa after sterilization. Water bacteria were of
course added with the Protozoa. Stokvis and Swel-
lengrebel (1911) have shown that ciliated infusoria
may also consume considerable quantities of bacteria
under favorable conditions as to oxygen and temperature,
and Horhammer (1911) reports that certain Crustacea
such as Cyclops may devour considerable quantities
of typhoid bacilli when present in masses from cultures?
stained with methylene blue, and suspended in water.

Certain bacteriologists have held that the toxic waste
products of the bacteria themselves may render water
unfit for their own development. Horrocks (Horrocks,
1901), Garre (Garre, 1887), Zagari (Zagari, 1887) and

16 ELEMENTS OF WATER BACTERIOLOGY

Freudenreich (Freudenreich, 1888) have shown that an
" antagonism " exists when bacteria are grown in
artificial culture media, such that the substratum which
has supported the growth of one form may be rendered
antiseptic to another. Frost (1904) has exhaustively
studied the phenomenon of antagonism by exposing
typhoid bacilli in collodion sacs to the action of certain
soil and water bacteria growing in broth. Artificial
culture media, however, offer conditions for bacterial
development which are scarcely paralleled in natural
waters. It is difficult to believe that under ordinary
conditions poisons are produced of such power as to
render a stream or lake specifically toxic for any par-
ticular type of bacteria. It does appear indeed from
the experiments of Jordan, Russell and Zeit (1904), and
Russell and Fuller (1906), which will shortly be referred
to more fully, that the life of typhoid germs is shorter
in water containing large numbers of other bacteria
than in that of greater purity. Horrocks (1899), too,
found freshly isolated typhoid bacilli alive in sterile
sewage after 60 days; while they disappeared in 5 days
when B. coli was also present. These phenomena may
be due, however, to a struggle for oxygen, or for food,
rather than to the assumed presence of highly toxic
bacterial products, of which there is no independent
evidence.

Many investigations conducted since the pioneer
researches of Downes and Blunt (Downes and Blunt,
1877) have confirmed the results reported by them,
which showed that direct sunlight is fatal to most
bacteria in the vegetative state and even to spores if

THE BACTEEIA IN NATURAL WATERS 17

the exposure be sufficiently long, while diffused light
is harmful in a less degree. Opinions vary as to the
degree to which light is active in destroying the bacteria
in natural waters. Buchner (Buchner, 1893) found
by experiment that the bactericidal power of light
extends to a depth of about three meters before it
becomes imperceptible. On the other hand, Procaccini
(Procaccini, 1893) found that when sunlight was passed
vertically through 60 cm. of drain- water the lower
layers contained nearly as many bacteria after 3 hours'
treatment as before the exposure. The middle and
upper portions showed a great falling off in numbers,
however.

But few studies have been made of the effect of light
on bacteria in flowing water. Jordan (Jordan, 1900)
has investigated several Illinois streams and arrived
at the conclusion that in moderately turbid water, at
least, the sun's rays are virtually without action. On
the other hand, Rapp has observed a considerable
reduction of the bacteria in the Isar at Pullach after the
period of diurnal insolation, as shown by the table on
the following page. Clemesha (191 2a) attributes very
great importance to the action of light in the self-
purification which takes place in Indian lakes and
rivers; his opinion is apparently not based on com-
parative experiments including and excluding this
factor, but chiefly on the greater numbers of intestinal
bacteria at the bottom as compared with the superficial
layers of water.

It is unnecessary to dwell in detail upon the effect
which the lack of nutritive elements must exert upon

ELEMENTS OF WATER BACTERIOLOGY

intestinal bacteria and soil bacteria in waters of ordinary
purity. Comparative studies of culture media, to be
quoted in the succeeding chapter, will show how del-
icately the bacteria respond to comparatively slight
changes in their food-supply. Wheeler (1906) found
that typhoid bacilli would persist in almost undimin-
ished numbers in sterilized water from a polluted well
containing considerable organic matter and kept in
the dark at 20 degrees, while in purer water or in the
light they died out in from 2 to 6 weeks.

EXAMINATIONS OF THE ISAR AT PULLACH

(RAPP, 1903)
(A} Carried out September 26. i8p8, no rain having fallen for three weeks

Temperature.

Time o the
Experiment.

Bacteria per c.c.

of the Water.

of the Air.

i3.o°C.

8Q O /"^
. 0 C.

7.3op.M.

146

I2.I°C.

7.o°C.

9.30P.M.

270

10. 5° C.

6.2°C.

5.00 A.M.

370

10.2° C.

8.2°C.

8.00 A.M.

320

(B) Carried out November 28, 1898, no rain having fallen for some time

5-5° C.

3-o°C.

6.OO P.M.

266

5-5°C.

2-5°C.

8.00P.M.

402

5.5°C.

2.0°C.

10. 00 P.M.

482

5.o°C.

2.0°C.

3-OO A.M.

532

4.5°C.

2-5°C.

7.30A.M.

400

Whipple and Mayer (1906) have called attention to
another important factor in the general problem. They
find that the presence of oxygen is essential to the per-

THE BACTERIA IN NATURAL WATERS

sistence of typhoid and colon bacilli in water, although
in nutrient media both forms may thrive under anaerobic
conditions.

EFFECT OF OXYGEN ON VIABILITY OF TYPHOID BACILLI
IN STERILE TAP WATER

WHTPPLE AND MAYER, 1906

Tubes Kept in Air.

Tubes Kept in Hydrogen.

Period in Days

Bacteria
per c.c.

Per Cent.

Bacteria
per c.c.

Per Cent.

600,000

IOO.O

6oo,000

IOO.O

455;°°°

76.0

2,400

0.4

190,000

32.0

0.004

120,000

20. O

O.O

67,000

II. 0

O.O

25,000

4-2

O.O

9,250

i-5

O.O

2,150

0.6

0.0

132

O.O2

O.O

O.OOI

O.O

o.ooo

0.0

Various inorganic constituents of the medium undoubt-
edly exercise an important influence upon the life of
bacteria in water; and the mutual interaction of the
different substances present is a highly complex one.
Thus Winslow and Lochridge (1906) report that five
parts of dissociated hydrogen per million parts of tap
water (0.005 normal HC1) is fatal to typhoid bacilli,
while ten times as much acid is required for sterilization
when i per cent of peptone is present to check the
dissociation of the hydrogen. In Hazen and Whipple's
study of the Allegheny, Monongahela and Ohio rivers

20 ELEMENTS OF WATER BACTERIOLOGY

at Pittsburgh the antiseptic effect of acid wastes was
strikingly shown. (Engineering News, 1912.)

Although it is hard to estimate the exact importance
of each factor, the general phenomena of the self-
purification of streams are easy to comprehend. A
small brook, immediately after the entrance of polluting
material from the surface of the ground, contains many
bacteria from a diversity of sources. Gradually those
organisms adapted to life in the earth or in the bodies
of plants and animals die out, and the forms for which
water furnishes ideal conditions survive and multiply.
It is no single agent which brings this about, but that
complex of little-understood conditions which we call
the environment. If any one thing is of prime impor-
tance it is probably the food-supply, for only certain
bacteria are able to multiply in the presence of the
small amount of organic matter present in ordinary
potable waters. As Jordan (Jordan, 1900) has said:
" In the causes connected with the insufficiency or
unsuitability of the food-supply is to be found, I believe,
the main reason for the bacterial self-purification of
streams."

Effect of Temperature upon Bacteria in Water.
The effect of temperature upon the survival of bac-
teria in water varies according to this primary con-
dition of food-supply which has just been discussed.
When bacteria are in a medium in which they are able
to grow and multiply, warmth, within reasonable limits
of course, favors their development, At times this
may be true even of certain intestinal bacteria in water.
Thus at Harrisburg, Pa., a series of B. coli examinations

THE BACTERIA IN NATURAL WATERS 21

made in the midsummer of 1906 showed positive
results in 7 per cent of the samples of water entering the
storage reservoir and in 27 per cent of the samples
leaving it. The storage period in this case was about
two days and the temperature of the water in the
reservoir was nearly at blood heat (Harrisburg, 1907).
Clemesha (1912*) has recently made an exhaustive study
of this multiplication of coli-like microbes in warm
waters and has shown that it is confined to certain
particular types within the colon group. For most
intestinal bacteria the conditions necessary for growth
and multiplication are not realized in water and an
entirely different temperature effect is manifest. When
a bacterium cannot multiply, the only vital activity
which can take place is a katabolic wasting away,
which soon proves destructive, and the higher the
temperature the more rapidly the fatal result is reached.
A frog in winter lives at the bottom of a pond breath-
ing only through its skin and eating not at all, but as
soon as the temperature rises it must eat and breathe
through its lungs or perish. It is quite true that even
in ice 40 per cent of typhoid bacilli perish in 3 hours
and 98 per cent in 2 weeks (Sedgwick and Winslow,
1902). Recent work has shown, however, that they
die in spite of the cold, not on account of it, and that
the decrease is more rapid at higher temperatures,
unless of course food-supply and other conditions admit
of multiplication. Houston (1911) has furnished a
very clear demonstration of this temperature rela-
tion by storing typhoid bacilli in water with the results
tabulated on page 22.

ELEMENTS OF WATER BACTERIOLOGY

EFFECT OF TEMPERATURE ON SURVIVAL OF TYPHOID
BACTERIA IN WATER

(HOUSTON, 1911)

Temperature C.

Percentage of Typhoid
Bacilli Surviving after
One Week.

Period of Final
Disappearance of BacilH.

9 weeks

7 weeks

O O7

5 weeks

O.O4

4 weeks

Ruediger (1911) has shown that colon bacilli are
far more abundant in the Red Lake River during the
winter when the river is covered with ice than in sum-
mer, although the volume of the river and the amount
of sewage pollution are about the same. Typhoid
bacilli in celloidin dialyzers floated down the river
showed only 2.5 and 3.5 per cent surviving in 2 days
and 0.51, 0.89, 2.2 and 3.2 per cent surviving in 3 days
when the river was not frozen, while dialyzers suspended
through the ice in colder weather showed 6.1, 10.5, 17.7,
46.8 and 62.9 per cent surviving in five different experi-
ments after 2 days, 31 per cent in 3 days, 19 per cent
in 7 days, and 2.5 per cent in 14 days. Ruediger
attributes this greater persistence at low temperatures
to the absence of poisonous waste products of other
organisms and to protection from the light; but there
can be little doubt that it is mainly a result of the general
preservative effect of cold. From an epidemiological
standpoint the conclusion that disease germs perish
quickly in warm waters is amply confirmed. Almost
without exception outbreaks of typhoid fever due to

THE BACTERIA IN NATURAL WATERS 23

polluted water occur in cold weather and this is, in
part at least, due to the greater persistence of typhoid
bacilli at low temperatures.

Relation between Time of Storage and Self-purifica-
tion. It is obvious that the efficiency of all the agencies
which tend to decrease the number of bacteria in sur-
face waters will increase with the prolongation of the
period for which they act. Time is the great measure
of self -purification.

The longer the storage the greater the improvement,
and after a certain period even a fairly polluted water
may be safe and potable. The absolute time necessary
to produce this result varies of course according to many
conditions. Food supply, light, temperature and the
activity of other living forms vary widely and in depos-
ited material conditions are different from those which
obtain in the water itself. Jordan, Russell and Zeit
(1904), in an important series of experiments, added
typhoid bacilli to the unsterilized waters of Lake
Michigan, the Chicago River and Drainage Canal and
the Illinois River, in collodion sacs suspended in the
respective bodies of water. From the relatively pure
Lake Michigan water the specific organisms could be
isolated for at least a week, but in the polluted waters
of the rivers and the Drainage Canal they were not
found after 3 days except in a single instance. Russell
and Fuller, (1906) confirmed these general results,
finding that typhoid bacilli would live for 10 days in
the unsterilized water of Lake Mendota, while they
could be isolated only after 5 days when immersed
in sewage. Other observers record much greater

24 ELEMENTS OF WATER BACTERIOLOGY

viability for the typhoid bacillus. Savage (1905)
added a heavy dose of the organism to unsterilized
tidal mud and found it living after 5 weeks. Hoffmann
(1905), after inoculating a large aquarium with a rich
typhoid culture, was able to isolate the germ from the
water after four weeks and from the mud at the bottom
after two months. Konradi (1904) reports the per-
sistence of typhoid bacilli in unsterilized tap water
for over a year.

These last experiments deal only with the maximum
survival period for a few out of great numbers of germs
introduced into the water or mud, and entirely ignore
the quantitative aspects of the case. When one con-
siders the proportion of the original bacteria surviving,
the period necessary to bring about a reasonably safe
condition is found to be much shorter. Houston
(1908) has shown that when water is artificially infected
with .typhoid bacilli and stored, 99.9 per cent of the
disease germs perish in one week, although some may
persist for from i to 9 weeks.

In later experiments (Houston, 1911) he finds that
" uncultivated " typhoid bacilli added to the water
directly from the urinary sediment of a disease carrier
perish much more rapidly than the laboratory strains,
usually disappearing entirely after one week and always
after three. On a number of occasions Dr. Houston
gave dramatic expression to his confidence in these
negative laboratory findings by drinking half pint
portions of water which a few weeks previously had
contained millions of typhoid bacilli. We have plenty
of practical epidemiological evidence, such as that

THE BACTERIA IN NATURAL WATERS 25

offered in the Chicago Drainage Canal case and in the
lawsuit over the condition of the water supply of Jersey
City, to confirm the general conclusion that any water
which has been stored for 4 weeks is practically safe.

Bacteria in Ground-waters. In general we have
seen that surface-waters tend continually to decrease
in bacterial content after their first period of contact
with the humus layer of the soil. In that other portion
of the meteoric water which penetrates below the
surface of the earth to join the reservoir of ground-
water, later to reappear as the flow of springs and wells,
this diminution is still more marked, since the filtering
action of the earth removes not only most of the bac-
teria, but much of their food material as well. The
numbers of bacteria in the soil itself decrease rapidly
as one passes downward. Kabrhel (1906) found several
million per c.c. in surface samples of woodland soil,
a few thousands or tens of thousands half a meter
below, and usually only hundreds in centimeter
samples collected at depths greater than a meter.
Many observers formerly believed that all ground-
waters were nearly free from bacteria, because often
no colonies appeared on plates counted after the
ordinary short periods of time. If, however, a longer
period of incubation be adopted considerable numbers
may be obtained.

For convenience we may divide ground-waters into
three groups, namely: shallow open wells, springs and
" tubular " (driven) or deep wells. This division is
important because ordinary shallow wells form a group
by themselves in respect to the possibility of aerial and

26 ELEMENTS OF WATER BACTERIOLOGY

surface contamination, their water often being fairly
rich in bacterial life. Egger (Wolff hugel, 1886) examined
60 wells in Mainz and found that 17 of them contained
over 200 bacteria to the cubic centimeter. Maschek
(Maschek, 1887) found 36 wells out of 48 examined in
Leitmeritz which had a bacterial content of over 500
per c,c. Fischer (Horrocks, 1901) reported 120 wells
in Kiel which gave over 500 bacteria per c.c. and only
51 with less than that number.

In the examination of 147 shallow farmyard wells
by one of us (S. C. P.) it was found that 124 of the
wells which contained no B. coli, and were therefore
probably free from fecal pollution, averaged 190 bacilli
per c.c. while 23 which gave positive tests for B. coli
averaged 570 per c.c. The distribution of the two series
of samples according to the number of bacteria present
is indicated in the table below.

BACTERIA IN SHALLOW FARMYARD WELLS
PERCENTAGE OF SAMPLES IN EACH GROUP

Bacteria per c c

ii-

21-

5i-

101-

501-

IOOI-

2OOI-

IOO

500

IOOO

2OOO

3000

Series I. B. coli absent.

Series II. B. coli present .

Very similar results are reported for shallow wells
used as farm water-supplies in Minnesota by Kellerman
and Whittaker (1909), although the general quality
of the wells examined was considerably below that
of the series tabulated above.

THE BACTERIA IN NATURAL WATERS

In the ordinary standard 48-hour period very few
bacteria develop from normal spring- waters. Thus
in an examination of spring-waters made by the Mas-
sachusetts State Board of Health in 1900 (Massachusetts
State Board of Health, 1901), of 37 springs which were
practically unpolluted and had less than o.io part
per 100,000 excess of chlorine over the normal, 54 sam-
ples were examined and gave an average of 41 bacteria
per c.c. Only 6 samples showed figures over 50.

It now remains to consider the other great division
of ground-waters, namely, deep, " driven," or " tubular "
wells, which, if carefully constructed, should ordinarily
be free from all surface-water contamination, and should
show low bacterial counts. The results tabulated below
obtained by Houston in the examination of a series of
deep wells of high quality at Tunbridge Wells are
fairly typical.

BACTERIAL CONTENT OF DEEP WELL WATERS

(HOUSTON, 1903)
Bacteria per c.c.

Fifteen driven wells in the neighborhood of Boston,
examined in 1903, showed at the end of 48 hours an
average of only 18 colonies per c.c.; and the results
of certain examinations of other wells and springs,
recently made by the authors, are given in the table
on page 28.

28 ELEMENTS OF WATER BACTERIOLOGY
BACTERIA IN DEEP WELL AND SPRING WATERS

Town

Bacteria
per c.c.

Town.

Bacteria
per c.c.

Worcester, Mass
Waltham, Mass
Newport, R.I

Saranac Lake, N. Y. .
Ellenville, N. Y
Hyde Park, Mass

It is plain that water absolutely free from bacteria
is not ordinarily obtained from any source. In deep
wells, however, their number is small; and the peculiar
character of the organisms present is manifested in
many cases by the slow development at room tem-
perature (frequently no growth until the third day),
the entire absence of liquefying colonies, and the
abundance of chromogenic species.

CHAPTER II

THE QUANTITATIVE BACTERIOLOGICAL EXAMINATION
OF WATER

Relation of the Medium to the Number of Bacteria
Obtained. The customary methods for determining
the number of bacteria in water do not reveal the total
bacterial content, but only a very small fraction of it,
as becomes apparent when we consider the large num-
ber of organisms, nitrifying bacteria, strict anaerobes,
etc., which refuse to grow, or grow only very slowly in
ordinary culture media, and which, therefore, escape
detection. On the one hand, certain obligate parasites
cannot thrive in the absence of the rich fluids of the
animal body; on the other hand, the prototrophic
bacteria, adapted to the task of wrenching energy from
nitrites and ammonium compounds are unable to develop
in the presence of so much organic matter. Winslow
(1905) in the examination of sewage and sewage effluents,
found 20-70 times as many bacteria by microscopic
enumeration as by the gelatin plate count. Certain
special media enable us to obtain much larger counts
than those yielded by the ordinary gelatin method.
The Nahrstoff Heyden agar, for example, has been
strongly advocated by Hesse (Hesse and Niedner,
1898) and other German bacteriologists upon this

30 ELEMENTS OF WATER BACTERIOLOGY

ground. In this country Gage and Phelps (Gage and
Phelps, 1902) showed that the numbers obtained by
the ordinary procedure were only from 5 to 50 per cent
of those obtained by the use of Heyden's Nahrstoff
agar. For practical sanitary purposes, however, our
methods are fairly satisfactory. Within limits, it is
of no great importance that one method allows the
growth of more bacteria than another. When we are
using the quantitative analysis as a measure of sewage
pollution the essential thing is that the section of the
total bacterial flora which we obtain should be thor-
oughly representative of that portion of it in which
we are most interested- — the group of the quickly
growing, rich-food-loving sewage forms. In this respect
meat-gelatin-peptone appears to be unrivalled; and it
is in this respect that such media as Nahrstoff agar fail.
Miiller (1900) showed that the larger counts obtained
by plating on the Nahrstoff medium are due to the
fact that it specially favors the more prototrophic
forms, among the water bacteria themselves. Intestinal
organisms and even the ordinary putrefactive germs,
when plated in pure culture, show no higher counts on
Nahrstoff agar than on gelatin. Gage and Adams
(1904) found by plating pure cultures of the common
laboratory bacteria, saprophytes and parasites, that
Nahrstoff counts were actually lower than those obtained
by the use of gelatin. When sewage and highly polluted
waters are examined counts are slightly higher on
Nahrstoff media, while with purer waters the Nahrstoff
numbers are far in excess of those obtained with gelatin.
Winslow (1905) found the ratio of Nahrstoff agar

QUANTITATIVE EXAMINATION OF WATER 31

to gelatin count to be 1.7 to i.o for sewage, and 4.8
to i.o for sand filter effluent. With waters of still
better quality the ratio goes up higher, reaching a
maximum when the bacteria which increase and
multiply in water are most abundant. Miiller (1900)
found, for example, that water which normally showed
six times as many bacteria on Nahrstoff agar as on
gelatin might give a Nahrstoff-gelatin ratio of 20-30
after it had been standing for some time in the supply
pipes. The table below, taken from the valuable
paper by Gage and Phelps (1902), shows strikingly
the different Nahrstoff-agar ratios for waters of

TABLE SHOWING PERCENTAGES OF BACTERIA DEVELOP-
ING ON REGULAR AGAR AND NAHRSTOFF AGAR
FOR DIFFERENT CLASSES OF WATERS

(GAGE AND PHELPS, 1902)

Regular Agar

Days' Count.

Class of Water.

Ground water

Filtered water

Merrimac River. . .

Filtered sewage. . . .

Sewage

Narhstoff Agar

Ground water

IOO

Filtered water

IOO

Merrimac River . . .

IOO

Filtered sewage. . . .

IOO

Sewage

IOO

32 ELEMENTS OF WATER BACTERIOLOGY

various grades of purity. It is obvious from all these
facts that the effect of using the Nahrstoff medium
is to increase disproportionately the bacterial counts
obtained from purer waters and thus to diminish the
difference in bacterial content between normal and
contaminated sources. The ordinary agar and gelatin
media, on the other hand, are adapted to the growth
of intestinal and putrefactive forms and, therefore,
serve best the prime object of bacteriological water
examination.

The first requisite in a procedure for water analysis
is, then, that it should be adapted to the end in view,
the differentiation of pure and contaminated waters.
The second and equally important requirement is that
the procedure should be a standard one, so that results
obtained at different times and by different observers
may be comparable. In this respect the work of G. W.
Fuller, G. C. Whipple, and other members of the
Committee on Standard Methods of the American Public
Health Association has placed the art of quantitative
water analysis in this country in a very satisfactory
state by contrast with the varying practices which
prevail in England and Germany. The first report on
this question was made in 1897 (Committee of Bac-
teriologists, 1898). A permanent Committee on Stand-
ard Methods was then formed which reported in 1901
(Fuller, 1902), in 1904 (Committee on Standard Methods
of Water Analysis, 1905), and again in 1911 (Committee
on Standard Methods for the Examination of Water
and Sewage, 1912), recommending in considerable
detail a standard routine procedure for the quantitative

QUANTITATIVE EXAMINATION OF WATER 33

and qualitative bacteriological examination of water
for sanitary purposes. These reports have had a far-
reaching effect in simplifying and unifying the methods
of water analysis. Similar results have followed from
the work of the English Committee appointed to con-
sider the Standardization of Methods for the Bac-
terioscopic Examination of Water which reported in
1904, although this committee unfortunately did not
consider the process of media making in great detail.
The last report of the American Committee on Standard
Methods (1912) will be adhered to in this and succeed-
ing chapters unless otherwise specifically stated; and
that portion of its report which deals with methods of
making media will be found in full in the appendix.

Standard Procedure for Quantitative Determination of
Bacteria in Water. The procedure for the quantitative
determination of bacteria in water consists, in brief,
in mixing a definite amount of a suitably collected
specimen of the water with a sterile, solidifiable culture
medium and incubating it for a sufficiently long time
to permit reproduction of the bacteria and the forma-
tion of visible colonies which may be counted. The
process is divided naturally into four stages — sampling,
plating, incubating, and counting.

Sampling. All samples of water for bacteriological
examination should be collected in clean, sterile bottles
with wide mouths and glass stoppers, preferably of the
flat mushroom type. It is desirable that these bottles
should have a capacity of at least 100 c.c.

They should be cleaned thoroughly before using, by
treatment with sulphuric acid and potassium bichromate

34 ELEMENTS OF WATEE BACTERIOLOGY

or with alkaline permanganate of potash followed by
sulphuric acid, dried by draining, and sterilized by
dry heat at 160° C. for at least i hour, or by steam at
115-120° for 15 minutes. If not to be used immediately
the neck and stopper should be protected against dust
or other contamination by wrapping with lead-foil.
For transportation the bottle should be enclosed in a
suitable case or box.

The greatest care must be taken that the fingers do
not touch the inside of the neck of the bottle or the cone
of the stopper, as the water thereby would become
seriously contaminated and rendered unfit for examina-
tion. It is well known that bacteria are found abun-
dantly upon the skin, and Winslow (Winslow, 1903)
has shown that even B. coli is present upon the hands
in a considerable number of cases.

In order to obtain a fair sample, great precautions
must be taken, and these will vary with the different
classes of waters to be examined and with local condi-
tions. If a sample is to be taken from a tap, the water
should be allowed to flow at least five minutes (if from
a tap in regular use) or for a longer period in case the
water has been standing in the house-service system.
In the small pipes, changes in bacterial content are
liable to occur, certain species dying and others mul-
tiplying.

If a sample is to be taken from a pump similar pre-
cautions are necessary. The pump should be in con-
tinuous operation for 5 minutes at least, and preferably
for half an hour before the sample is taken, in order to
avoid excessively high numbers due to the growth of

QUANTITATIVE EXAMINATION OF WATER 35

bacteria within the well and pump, the bacterial con-
dition of the water as it passes through the ground being
what we wish to determine. Thus Heraeus (Heraeus,
1886) in a well-water which had been but little used
during the preceding 36 hours found 5000 organisms
per c.c.; when the well was emptied by continuous
pumping, a second sample, after an interval of half an
hour, gave only 35. Maschek (Tiemann and Gartner,
1889) obtained similar results, shown in the following
table:

EFFECT OF PUMPING ON THE BACTERIAL CONTENT OF
WELL-WATER

Well-water after continuous pumping for fifteen minutes . . 458

many hours 140

later 68

after continuous pumping for fifteen minutes . . 578

many hours 1 79

later 73

After a proper interval of pumping the sample of a
well-water may be collected from the pet-cock of the
pump or from a near-by tap. With a hand-pump,
such as is found in domestic shallow wells, the water is,
of course, pumped directly into the sample bottle.
The difficulties in securing an average sample from
this latter source are often great, since if the flooring
about the pump is not tight, as is usually the case, con-
tinued pumping may wash in an unusual amount of
surface pollution.

In sampling surface-waters, the greatest precautions
must be observed to prevent contamination from the
fingers. In still waters the fairest sample is one taken

36 ELEMENTS OF WATER BACTERIOLOGY

from several inches down, as the surface itself is likely
to have dust particles floating upon it. The method
most frequently recommended is to plunge the bottle
mouth downward to a depth of a foot or so, then
invert and allow the bottle to fill.

Whenever any current exists, the mouth of the bottle
should be directed against it in order to carry away any
bacteria from the fingers. If there is no current, a
similar effect can be produced by turning the bottle
under water and giving it a quick forward motion. In
rapidly flowing streams it is only necessary to hold
the bottle at the surface with the mouth pointed
up-stream.

For taking samples of water at greater depths, a
number of devices have been employed, all of which
are fairly satisfactory. The essentials are, first, a weight
to carry the bottle down to the desired depth, and,
second, some method of removing the stopper when
that depth is reached. The student will find one good
form of apparatus described in Abbott's " Principles
of Bacteriology" (Abbott, 1899); an admirable one
was devised by Hill and Ellms (Hill and Ellms, 1898);
and Thresh (1904) figures an ingenious device for the
same purpose. Miquel and Cambier (Miquel and
Cambier, 1902) and other authors recommend the use
of a sealed glass bulb with a capillary tube which can
be broken off at the desired moment. Drew (1912)
has devised an interesting sampling apparatus for use
at great depths in the sea.

Changes in Bacterial Numbers after Sampling. As
soon as a sample of water is collected its conditions

QUANTITATIVE EXAMINATION OF WATER 37

of equilibrium are upset and a change in the bacterial
content begins. Even in the purest spring-waters,
which contain but few bacteria when collected, and in
which the amount of organic matter is infinitesimal,
enormous numbers may be found after storage under
laboratory conditions for a few days or even a few hours.
In some cases the rise in numbers is gradual, in others
very rapid. The Franklands (Frankland, 1894) record
the case of a deep-well water in which the bacteria
increased from 7 to 495,000 in 3 days. Miquel (Miquel,
1891) from his researches, arrived at the conclusion
that in surface-waters the rise is less rapid than in waters
from deep wells or springs, and that in the latter case
the decrease, after reaching a maximum, is likewise
rapid and steady. Just how far protection from light,
increase in temperature, and a destruction of higher
micro-organisms is responsible for the increase, and
to what extent an exhaustion of food-supply or the
formation of toxic waste products causes the succeeding
decrease, we are not aware; but the facts are well
established.

Whipple has exhaustively studied the details of
this multiplication of bacteria in stored waters
and has shown in the table given below that
there is first a slight reduction in the number
present, lasting perhaps for 6 hours; followed by
the great increase noted by earlier observers. It
is probable that there is a constant increase of the
typical water bacilli, overbalanced at first by a
reduction in other forms, for which the environment
is unsuitable.

ELEMENTS OF WATER BACTERIOLOGY

BACTERIAL CHANGES IN WATER DURING STORAGE
(WHIPPLE, 1901)

Sample

Initial
Temper-
ature.

Temp,
of Incu-
bation of
Sample.

Number of Bacteria per c.c.

Initial.

After
3 Hours.

After
6 Hours.

After
24 Hours.

After
48 Hours.

7.6°

17.0°

260

215

230

900

27,000

7-6°

17.0°

260

245

255

720

10,850

7.6°

12.5°

260

270

231

600

2,790

7.6°

12.5°

260

270

245

710

1, 800

7.6°

2.4°

260

243

210

675

1,980

7.6°

2-4°

260

235

270

560

1,980

11.0°

12.8°

101

10,250

11.0°

12.8°

2,175

11.0°

23.6°

11,000

41,400

6-7°

20.0°

430

375

245

385,ooo1

6.7°

20.0°

430

345

405

75o,ooo1

23-2°

23.0°

510

340

230

8,000

20,000

23-2°

2-5°

525

300

220

380

2,200

1 0.0005 Per cent peptone added to the

water.

WolfThiigel and Riedel (Wolffhiigel and Riedel,
1886) noted the dependence of this multiplication
on the air-supply, vessels closed with rubber stoppers
showing lower numbers than those plugged with cotton.
Similarly, Whipple found that the multiplication of
bacteria was much greater when bottles were only
half full than when they were filled completely; and also,
as shown in the very striking table on page 39, that
the size of the bottle markedly influenced the growth.

An important series of investigations by Kohn (1906)
suggests that this phenomenon of multiplication dur-
ing storage may be due in part to the solution of certain
constituents of glass which favor bacterial life, since the
increase is notably greater in bottles of the more soluble
glasses.

QUANTITATIVE EXAMINATION OF WATER 39

EFFECT OF SIZE OF VESSEL UPON THE MULTIPLICATION
OF WATER BACTERIA DURING STORAGE

(WHIPPLE, 1901)

Sample

Bottle.

Temp,
of
Incuba-
tion.

Number of Bacteria per c.c.

Ini-
tial.1

After
3 Hrs.

After
6 Hrs.

After
12 Hrs

After
24 Hrs.

After
48 Hrs.

i -gallon

13°

175

2-qutirt

13°

690

i-quart

13°

325

i-pint

13°

630

2-ounce

13°

IOI

10.250

i -gallon

24°

275

290

300

2-quart

24°

1 80

250

i-quart

24°

340

900

i-pint

24°

100

2,950

7,O2O

2-ounce

24°

11,000

41,400

1 Average of five plates.

Whipple's table, quoted above, shows that the multi-
plication during storage was greater at a higher tem-
perature; and this is a well-recognized general rule.
In order to obviate the abnormal results of storage
increase it is therefore obvious that samples must be
examined shortly after collection and that they must
be kept cool during their necessary storage. If fairly
pure waters are placed upon ice and kept between o
degrees and 10 degrees, they will show no material
increase in 12 hours. With polluted water, however,
another danger is here introduced. Samples of such
water when packed in ice show a marked decrease
due to the large number of sensitive intestinal bacteria
present. Jordan (Jordan, 1900) found that three
samples of river-water packed in ice for 48 hours fell

40 ELEMENTS OF WATER BACTERIOLOGY

off from 535,000 to 54,500; from 412,000 to 50,500,
and from 329,000 to 73,000, respectively. It is, there-
fore, important that even iced samples should not be
kept too long; and it is desirable to adhere strictly
to the recommendations of the Standard Methods
Committee that the interval between sampling and
examination should not exceed 12 hours in the case of
relatively pure waters, 6 hours in the case of relatively
impure waters, and i hour in the case of sewage.

Plating. The bottle containing the sample of water
is first shaken at least twenty-five times in order to
get an equal distribution of the bacteria. If the num-
ber of bacteria present is probably not greater than 200,
i c.c. is then withdrawn with a sterile i c.c. pipette
and delivered into a sterile Petri dish of 10 cm. diameter.
To this is added 5 c.c. of standard 10 per cent gelatin
at a temperature of about 30° C., or standard agar
(7 c.c.) at 40-42° C. Should the number of bacteria
per c.c. probably exceed 200, dilution is necessary.
This is best accomplished by adding i c.c. of the water
in question to 9, 99 or 999, etc., c.c. of sterile tap water
according to the amount of dilution required. The
diluted sample is then shaken thoroughly and i c.c.
taken for enumeration. In order to determine the
number of bacteria originally present it is only neces-
sary to multiply by the factor 10, 100, or 1000, etc.

When a sample of water from an unknown source
is to be examined it is generally desirable to make
two check plates at each of the above dilutions, select-
ing those which give nearest to 200 colonies on the
plates after incubation as the ones on which to rely

QUANTITATIVE EXAMINATION OF WATER 41

for the count. A much smaller number will not give
average figures, and if more than 200 colonies are present
on a plate many bacteria will be checked by the waste
products of those which first develop and the count
obtained will be too low. After the addition of the
diluted sample and the nutrient medium, their thorough
mixture in an even layer on the bottom of the plate
is obtained by careful tipping and rotation.

It was formerly customary to mix the water with the
gelatin in the tube before pouring into the plate, but
this method is objectionable because there is always
a residuum of medium remaining in the tube which
will retain varying numbers of bacteria and thus
interfere with the accuracy of the count. Before pour-
ing the medium into the plate the mouth of the tube
should be flamed to remove any possibility of con-
tamination.

The usual method of determining the number of
bacteria in water for sanitary purposes in Germany,
England and the United States has always been by
the use of gelatin plates with a 2 -day incubation period
at 20 degrees. The 1905 Standard Methods Report
of the American Public Health Association Committee
recommended this procedure, which has been universally
adopted. The 1912 Report, however, suggests the use
of agar with a i-day period at 37 degrees, as yielding
quicker results and indicating the presence of bacteria
more nearly related to pathogenic types. The com-
parative value of the two methods has been well dis-
cussed by Whipple (1913). The use of gelatin is not
only more time-consuming, but requires the use of a

42 ELEMENTS OF WATER BACTERIOLOGY

special 2o-degree incubator which is difficult to regulate.
The 37-degree incubator must be provided in any case
for the isolation of B. coli. On the other hand, the
time seems hardly ripe for the abandonment of the
2o-degree count, which has been used for 20 years all
over the civilized world, and for the interpretation of
which we have very complete data. There is at present
no such sound basis for interpreting the 37-degree
count, and in many cases, as in the control of water
nitration plants, the 37-degree numbers are too small
to be of any practical value. Furthermore the 20-
degree count may furnish evidence of surface con-
tamination as distinguished from fecal pollution, which
is often of considerable value.

The authors have always urged the use of the 37-
degree count along with the 2o-degree count as furnish-
ing most valuable information; but this is very different
from the substitution of one count for the other.
The recommendation that the 2o-degree count be
abandoned, with no evidence to warrant such a revolu-
tionary change, and no experimental results on which
to base an interpretation of the 37-degree count, has
aroused vigorous opposition from a large majority
of practical water bacteriologists. At the Washington
meeting of the American Public Health Association
in September, 1912, it was resolved " that in the opinion
of the Laboratory Section of the American Public
Health Association, ordinary routine examination of
water for sanitary purposes, and in the control of
purification plants, for the present should include the
determination of the number of bacteria developing

QUANTITATIVE EXAMINATION OF WATER 43

at 20 degrees and at 37 degrees and a presumptive
test for B. coli in lactose bile."

This action of the section responsible for the appoint-
ment of the Standard Methods Committee appears to
supersede the report of the committee itself and makes
the combination of the 20- and 37-degree counts the
standard American procedure. The 2o-degree count
may be made on either gelatin or agar; but it is the
2o-degree count which will be discussed in this chapter,
leaving the body temperature count for consideration
in Chapter IV.

The exact composition of the medium is, of course,
of prime importance in controlling the number of
bacteria which will develop. The figures previously
cited in connection with the discussion of Hesse's
Nahrstoff agar show how bacterial counts may vary
with media of widely different composition. The
table quoted on page 44 from Gage and Phelps (1902),
shows the considerable differences which may be due
to the presence or absence of meat infusion, peptone,
etc., in media of generally similar character (compare
the figures for plain gelatin, peptone, gelatin, and meat
gelatin). Much slighter variations than this, however,
are significant. The reaction of the medium was found
as early as 1891 to be important, for Reinsch (Reinsch,
1891) showed in that year that the addition of one
one-hundredth of a gram of sodium carbonate to the
liter increased sixfold the number of bacteria develop-
ing. Fuller (Fuller, 1895) and Sedgwick and one of
us (Sedgwick and Fresco tt, 1895), working indepen-
dently, established the fact that an optimum reaction

44 ELEMENTS OF WATER BACTERIOLOGY

existed for most water bacteria and that a devia-
tion either way decreased the number of colonies
developing.

TABLE SHOWING PERCENTAGES OF BACTERIA DEVEL-
OPING ON MEDIA OF DIFFERENT COMPOSITIONS

(GAGE AND PHELPS, 1902)

Medium.

Days'

Cour

Nahrstoff agar

0 ^

IOO

Nahrstoff peptone agar
Peptone agar
Meat agar

II
8

16
13

28
23

I 7

30
24

I 7

3°

24
1 7

30
24

I ^

Plain agar. ...

Regular agar

1 1

Nahrstoff glycerin agar
Nahrstoff meat agar

II
IO

Meat gelatin

Peptone gelatine

Standard gelatin .

I 3

Plain gelatin

I ^

I?,

Nahrstoff gelatin

Whipple (Whipple, 1902) has shown that not only
the particular kind of gelatin used, but its exact physical
condition as affected by sterilization and other previous
treatments, will materially affect the results obtained.
Gage and Adams (1904) found marked differences in
counts as the result of the use of the two best-known
commercial peptones. A long series of waters plated
on agar made up with Merck's and Witte's peptones,
respectively, showed the average relative results in the
table on page 45.

QUANTITATIVE EXAMINATION OF WATER 45

AVERAGE RELATIVE NUMBER OF BACTERIA ON PEP-
TONE AGAR WITH DIFFERENT PEPTONES

(GAGE AND ADAMS, 1904)

DAYS

Merck's

•2-2

Witte's

IOO

JOO

IOO

The same authors showed that the composition of
the water used exercised a marked selective action upon
the development of bacteria. Agar made up with
sewage permitted a maximum growth of sewage bacteria
and showed no colonies when inoculated with filtered
city water. On the other hand agar made up with city
water showed 100 per cent of the bacteria present in
city water and river water, three-quarters of those
present in sewage and less than half of those present
in sewage effluents.

Hesse (1904) found that the number of bacteria
developing on Nahrstoff agar varied with the composi-
tion of the glass tubes in which the media had previously
been sterilized. The more soluble glasses yielded
sufficient alkali to the medium to inhibit four-fifths of
the bacteria present in certain cases.

All these facts make it evident that only the strictest
adherence to a standard method can ensure comparable
results; the ordinary nutrient gelatin or agar should
then in all practical sanitary work be made up from
distilled water, meat infusion, peptone and gelatin or
agar, in exact accordance with the directions of the
Standard Methods Committee.

46 ELEMENTS OF WATER BACTERIOLOGY

Even the standard procedure fails to ensure uniformity
in one important respect. The meat infusion which it
calls for is in itself a highly variable quantity. Gage
and Adams (1904), in the examination of fifteen lots
of beef infusion, found variations of nearly i per cent
in organic solids (calculated on the weight of the whole
infusions), after the final filtration. The organic
constituents of the meat infusion varied, therefore,
among themselves by nearly the total amount of pep-
tone added. It is to be hoped that the standard methods
may soon be so revised as to eliminate this necessarily
uncertain constituent of nutrient media. Criticisms
of detail must, however, give way to the importance
of securing fairly comparable results; and the con-
fusion which would follow the use by individual bac-
teriologists of media made without meat would out-
balance the errors inherent in the standard procedure.

Incubation. Incubation should take place in a dark,
well-ventilated chamber where the temperature is
kept substantially constant at 20 degrees and where
the atmosphere is practically saturated with moisture.
It has been shown by Whipple (Whipple, 1899) and others
that the number of bacteria developing in plate cultures
is to a certain extent dependent upon the presence of
abundant oxygen and moisture. Thus, reckoning the
number of bacteria developing in a moist chamber at
100, the percentage counts obtained in an ordinary
incubator were as follows: 75 when the relative humid-
ity of the incubator was 60 per cent of saturation; 82
when it was 75 per cent; 98 when it was 95 per cent.
This source of error may be avoided by the use of ven-

QUANTITATIVE EXAMINATION OF WATER 47

tilated dishes and by the presence of a pan of water in
the incubating chamber.

According to American and German practice, plates
made for sanitary water analysis are counted at the
end of 48 hours. The English Committee appointed
to consider the standardization of methods for the
Bacterioscopic Examination of Water (1904) fixed the
time at 72 hours. French bacteriologists, and some
Germans (Hesse and Niedner, 1906), still recommend
longer periods, and the following table from Miquel
and Cambier (Miquel and Cambier, 1902) shows that
many bacteria fail to appear in our ordinary procedure.
It is, however, in the main, the characteristic water
bacteria which develop slowly, sewage bacteria almost
without exception being rapid growers. The longer
period of incubation is, therefore, not only inconvenient,
but undesirable, since it obscures the difference between
good and bad waters.

EFFECT OF THE LENGTH OF INCUBATION OF WATER
BACTERIA IN GELATIN UPON THE NUMBER OF
COLONIES DEVELOPING

(MIQUEL AND CAMBIER, 1902)

Length of Incubation.

Colonies
Developed.

Length of Incubation.

Colonies
Developed.

I day

o days

821

2 days

136

10 days . .

8qo

7 days

2 CA

1 1 days

802

4 days. . .

187

12 days

Q2I

5 days

S^o

i 3 days . .

QCI

6 days .

6*7

Mdays

076

7 days

72CJ

i ^ days . .

IOOO

8 days . .

780

48 ELEMENTS OF WATER BACTERIOLOGY

Counting. The number of bacteria is determined by
counting the colonies developed upon the plate, with the
aid of a lens magnifying at least five diameters. For
convenience in counting the plate may be placed upon
a glass plate ruled in centimeter squares and set over a
black tile; or the tile itself may be ruled. As has
already been said, it is desirable that the number of
colonies should not exceed 200, for when the number
is very high the colonies grow only to a small size,
making counting laborious and inaccurate, and many
do not develop at all. The best results are obtained
with numbers ranging from 50 to 200.

When it is possible to do so, all the colonies on the
plate should be counted. When they exceed 400 or
500 it is often easier, and fully as accurate, to count
a fractional part of the plate and estimate the total
number therefrom. This should not be done, however,
except in case of necessity.

Ayers (1911) has suggested two counting devices
which will be found very useful where a great many
plates have to be handled. For getting the best possible
transmitted light, he places his plate on the ground-
glass top of a wooden box, 7 inches square, with one side
open to admit light, which is reflected upward by a
plane mirror set in the box at an angle of 45 degrees.
An ordinary graduated-glass counting plate may be
placed between the ground-glass and the Petri dish, and
the eyes are protected from direct light by a screen
rising from the open side of the box. For picking
colonies from a gelatin plate in a warm room, he places
between the ground glass and the Petri dish a copper

QUANTITATIVE EXAMINATION OF WATER 49

box with top and bottom of glass 7 inches square and
ij inches deep, through which cold water is allowed to
circulate.

Expression of Quantitative Results. It is customary
in determining numbers to make plates in duplicate,
thereby affording a check upon one's own work. Owing
to the lack of precision in the method, the limit of
experimental error is a wide one. It should be possible
for careful manipulators to obtain results within 10
per cent of each other, but a closer agreement than this
is hardly to be expected. It has been suggested by the
committee of the American Public Health Association
that the following mode of expressing results be adopted
in order to avoid the appearance of a degree of accuracy
which the methods do not warrant.

NUMBERS OF BACTERIA FROM

1-50 shall be recorded to the

51-100

101-250

251-500

501-1000

1001-10,000

10,001-50,000

50,001-100,000

100,001-500,000

500,001-1 ,000,000

i ,000,001-5,000,000

nearest unit

5
10

25
50

IOO

500

1,000
10,000
50,000

100,000

The determination of numbers of bacteria in water
in the field has frequently been attempted. Since
the laboratory method of "plating out" is difficult
to use in field work, the Esmarch tube process has often
been employed. This consists in introducing into a
tube of melted gelatin or agar i c.c. of the water and then

50 ELEMENTS OF WATER BACTERIOLOGY

rotating the tube until the medium has solidified in a
thin layer on the inner wall. Other bacteriologists
have devised ingenious field kits for adapting the plate
method to this purpose, of which one very good form
has recently been described by Van Buskirk (1912).
The opportunity for air infection in work done outside
a proper laboratory is, however, always great; and it
is almost impossible to secure proper conditions for
incubation in any makeshift establishment. On the
whole, the authors are of the opinion that laboratory
examinations are to be preferred to those made in the
field, if a laboratory can be reached within 12 hours
or so of the time of collection of the samples.

CHAPTER III

THE INTERPRETATION OF THE QUANTITATIVE
BACTERIOLOGICAL EXAMINATION

Standards for Potable Water. The information fur-
nished by quantitative bacteriology as to the antecedents
of a water is in the nature of circumstantial evidence
and requires judicial interpretation. No absolute stand-
ards of purity can be established which shall rigidly
separate the good from the bad. In this respect the
terms " test " and " analysis " so universally used
are in a sense inappropriate. Some scientific problems
are so simple that they can be definitely settled by a
test. The tensile strength of a given steel bar, for
example, is a property which can be determined.
In sanitary water examination, however, the factors
involved are so complex, and the evidence neces-
sarily so indirect, that the process of reasoning much
more resembles a doctor's diagnosis than an engineering
test.

The older experimenters attempted to establish
arbitrary standards, by which the sanitary quality of a
water could be fixed automatically by the number of
germs alone. Thus Miquel (Miquel, 1891) published a
table according to which water with less than 10 bac-
teria per c.c. was " excessively pure," with 10 to 100

52 ELEMENTS OF WATER BACTERIOLOGY

bacteria, " very pure/' with 100 to 1000 bacteria,
" pure/' with 1000 to 10,000 bacteria, " mediocre," with
10,000 to 100,000 bacteria, " impure," and with over
100,000 bacteria, " very impure." Few sanitarians
would care to dispute the appropriateness of the titles
applied to waters of the last two classes; but many bac-
teriologists have placed the standard of " purity " much
lower. The limits set by various German observers
range, for example, from 50 to 300. Dr. Sternberg
(Sternberg, 1892) in a more conservative fashion,
has stated that a water containing less than 100 bacteria,
is presumably from a deep source and uncontaminated
by surface drainage; that one with 500 bacteria is
open to suspicion; and that one with over 1000 bac-
teria is presumably contaminated by sewage or surface
drainage. This is probably as satisfactory an arbitrary
standard as could be devised, but any such standard
must be applied with great caution. The source of
the sample is of vital importance in the interpretation
of analyses; a bacterial count which would condemn
a spring might be quite normal for a river; only figures
in excess of those common to unpolluted waters of the
same character give an indication of danger. Fur-
thermore, the bacteriological tests are far more delicate
than any others at our command, very minute addi-
tions of food material causing an immense multiplica-
tion of the microscopic flora. This delicacy necessarily
requires, both in the process of analysis and the inter-
pretation of results, a high degree of caution. As
pointed out in the previous chapter, the touch of a
finger or the entrance of a particle of dust may wholly

QUANTITATIVE BACTERIOLOGICAL ANALYSIS 53

destroy the accuracy of an examination. Even the
slight disturbance of conditions incident upon the
storage of a sample after it has been taken may in a
few hours wholly alter the relations of the contained
microbic life. It is necessary, then, in the first place,
to exercise the greatest care in allowing for possible
error in the collection and handling of bacteriological
samples; and in the second place, only well-marked
differences in numbers should be considered significant.

In the early days of the science, discussion ran high
as to the interpretation of bacteriological analysis;
and particularly as to the relation of bacterial numbers
to the organic matter present in a water. Different
observers obtained inconsistent results, and Bolton
(Bolton, 1886) concluded that there was no relation
whatever between the organic pollution of a water and
its bacterial content. Tiemann and Gartner (Tiemann
and Gartner, 1889) furnished the key to the difficulty
in their statement that there are two classes of bacteria,
the great majority of species normally occurring in
the earth or in decomposing organic matter, which
require abundance of nutriment, and certain peculiar
water bacteria which can multiply in the presence of
such minute traces of ammonia as are present in ordi-
nary distilled water. Even these prototrophic or
semi-prototrophic forms, however, require a definite
amount of food of their own kind.

Kohn (1906) determined the minimal nutrient mate-
rial requisite for certain of them and found that they
could develop in the presence of 198 X io~10 to I98X io"13
per cent of dextrose, 66Xio~13 to 66Xio"17 per cent

54 ELEMENTS OF WATER BACTERIOLOGY

ammonium sulphate and 66Xio~13 to 66Xio~19 per
cent ammonium phosphate. Similar minute amounts
of organic matter are found in the purest of natural
waters and under exceptional conditions certain species
of bacteria may therefore multiply in bottled samples,
or, at times, in a well or the basin of a spring. In
normal surface-waters, such growths of the prototrophic
forms do not apparently occur. Here it is found as a
matter of practical experience that the number of bac-
teria present depends upon the extent to which the
water has been contaminated with decomposing organic
matter, either by pollution with sewage or by contact
with the surface of the ground. The bacterial content
varies as the extent and character of the contamination
varies. It measures not merely organic matter, but
organic matter in a state of active decay, and like the
ammonias and other features of the sanitary chemical
analysis, indicates fresh organic pollution, with the added
advantage that the presence of the stable nitrogenous
compounds often present in peaty waters introduces
no error in the bacteriological analysis.

Bacterial Content of Surface-waters. In judging of a
surface-water the student will be aided by reference to
the figures given for certain normal sources in Chapter
I; the Boston tap water with 50 to 200 bacteria per
c.c. (Philbrick, 1905) and the water of Lake Zurich
with an average of 71 in summer and 184 in winter
(Cramer, 1885) may be taken as typical of good potable
waters; and numbers much higher than these are open
to suspicion, since all contamination whether contributed ,
by sewage or by washings from the surface of the

QUANTITATIVE BACTERIOLOGICAL ANALYSIS 55

ground is a possible source of danger. The excess of
bacteria in surface-waters during the spring and winter
months is by no means an exception to the general
rule that high numbers are significant, since the peril
from supplies of this character is clearly shown by the
spring epidemics of typhoid fever which at the times of
melting snow visit communities making use of unpro-
tected surface-waters. Streams receiving direct con-
tributions of sewage exhibit a similar excess of bacteria
at all times, numbers rising to an extraordinary height
near the point of pollution and falling off below as the
stream suffers dilution and the sewage organisms perish.
Miquel (Miquel, 1886) records 300 bacteria per c.c.
in the water of the Seine at Choisy, above Paris; 1200
at Bercy in the vicinity of the city, and 200,000 at St.
Denis after the entrance of the drainage of Paris.
Prausnitz (Prausnitz, 1890) found 531 bacteria per c.c.
in the Isar above Munich, 227,369 near the entrance of
the principal sewer, 9111 at a place 13 kilometers
below the city, and 2378 at Freising, 20 kilometers
further down. Jordan (Jordan, 1900), in his study
of the fate of the sewage of Chicago, found 1,245,000
bacteria per c.c. in the drainage canal at Bridgeport,
650,000 at Lockport, 29 miles below, and numbers
steadily decreasing to 3660 at Averyville, 159 miles
below the point of original pollution. Below Avery-
ville the sewage of Peoria enters and the numbers
rise to 758,000 at Wesley City, decreasing to 4800 in
123 miles flow to Kampsville. Brezina (1906) found
1900 bacteria per c.c. in the Danube River above, and
1 10,000 at the north of the Danube canal. This number

56 ELEMENTS OF WATER BACTERIOLOGY

fell to 85,000 one kilometer below, 62,000 four kilo-
meters below, and 40,000 seven kilometers down the
stream. Vincent (1905) records from 1000 to 46,000
bacteria per c.c. in the waters of more or less polluted
French rivers. Mayer (1902), on the other side of the
world, found 21 and 35 bacteria per c.c. in the Shaho
River, near its source, in the vicinity of the great
Chinese Wall and from 100,000 to 600,000 in the highly
polluted Whangpo near its mouth.

Bacterial Content of Ground-waters. In ground-
waters we have seen that bacteria may occasionally
be present in considerable numbers, but if so they are
generally organisms of a peculiar character, incapable
of development on the ordinary nutrient media in the
standard time. Thus in 48 hours we often obtain
counts measured only in units or tens such as have
been recorded in Chapter I. When higher numbers
are present, the general character of the colonies must
be taken into account, since besides the slowly-growing
forms certain other, water bacteria, which require a
comparatively small amount of nutriment, may multiply
at times in a deep well or the basin of a spring. In
such a case, however, the appearance of the plates at
once reveals the peculiar conditions, for the colonies
are of one kind and that distinct from any of the sewage
species. Thus Dunham (Dunham, 1889) reports that the
mixed water from a series of driven wells gave 2 bacteria
per c.c., while another well, situated just like the others,
contained 5000, all belonging to a single species common
in the air. Except in such peculiar cases as this high
numbers in a ground-water mean contamination.

QUANTITATIVE BACTERIOLOGICAL ANALYSIS 57

Bacteria in Filtered Waters. The process of slow
sand filtration for the purification of unprotected
surface-water is essentially similar to the action which
takes place in nature when rain soaks through the
ground to appear in wells and springs; and it is in the
examination of the effluent from such municipal plants
that the quantitative bacteriological analysis finds,
perhaps, its most important application. The chemical
changes which occur in the passage of water through
sand at a rate of 1,000,000 or 2,000,000 gallons per acre
per day are so slight as to be negligible. The bacteria
present should, however, suffer a reduction of 98 or
99 per cent, and their numbers furnish the best standard
for measuring the efficiency of such filtration plants.
At Lawrence, in 1905, Clark found an average of 12,700
bacteria per c.c. in the raw water of the Merrimac
River, while the number present in the filtered water
was only 70 (Massachusetts State Board of Health,
1906). Where the number of bacteria in the applied
water is smaller it is difficult to obtain so high a per-
centage efficiency. At Washington, for example, pro-
longed sedimentation generally reduces the bacterial
numbers to less than a thousand and it is almost impos-
sible to secure a 99 per cent removal. The actual
numbers of bacteria in the effluent are, however, much
lower than at Lawrence. The monthly average results
obtained for a year at these two plants are tabulated
on page 58.

Mechanical filtration gives similar results. Fuller
at Cincinnati (Fuller, 1899) records 27,200 organisms
per c.c. in the water of the Ohio River between

ELEMENTS OF WATER BACTERIOLOGY

September 21, 1898, and January 25, 1899, while
the average content of the effluent from the Jewell
filter was 400. Data with regard to the operation of
mechanical niters are now abundant, since all over the
world the operation of these plants is controlled by
bacteriological methods. Recently Johnson (1907) has
reported some interesting results from the far East.
At Osaka, Japan, an average of 200 bacteria per c.c. in
the raw water of the Yodo River was reduced, in 1905,
to an average of 25 by slow sand niters; at Bethmangala,
India, in 1906, mechanical niters treated the water of the
'Palar River, containing 4350 bacteria per c.c., and
yielded an effluent with only 13 per c.c. (Johnson, 1907).
The average monthly results obtained with the new
mechanical filter plant at Harrisburg, Pa., are included
in the table below for comparison with the figures

REMOVAL OF BACTERIA BY NATURAL SAND FILTERS
AND MECHANICAL FILTERS. BACTERIA PER C.C. IN
APPLIED WATER AND EFFLUENT. MONTHLY AVER-
AGES

Month.

Washington, 1906.

Lawrence, 1905.

Harrisburg, 1906.

Applied
Water.

Effluent.

Applied
Water.

Effluent.

Applied
Water.

Effluent.

January ....

1500

14,200

9,510

104

February . . .

550

14,800

21,228

298

March

650

10,300

3^326

April

400

3,600

170

39-905

May

1,900

6,187

June

o
22O

9,600

2,903

July

1 60

3^0°

685

August

190

14 19,500

1,637

September . .

I30

13,500

836 | 12

October

275

39,800

110

7-575

November . .

22O

8.700

26,224

236

December.. . .

700

• • ! 37,525

163

QUANTITATIVE BACTERIOLOGICAL ANALYSIS 59

recorded at Washington and Lawrence; and these
may be taken as typical, since the Harrisburg plant is
the latest of its type, as the Washington plant is the
newest and most perfectly equipped of slow sand
niters.

In well-managed purification plants the bacteria
in the effluent are determined daily, and any deviation
from the normal value at once reveals disturbing factors
which may impair the efficiency of the process. In
Prussia official regulations demand such systematic
examinations and prescribe 50 as the maximum number
of bacteria allowable in the filtered water. In the
same way the condition of an unpurified surface supply
may be determined by daily bacteriological analyses
and warnings of danger issued to the public, as has
been done at Chicago and other cities. In general,
any regular determination of variations from a normal
standard furnishes ideal conditions for the bacteriological
methods; and the detection by Shuttleworth (Shuttle-
worth, 1895) of a break in a conduit under Lake Ontario
by a rise in the bacteria of the Toronto water-supply
may be cited as a classic example of its application.

Often, however, the expert is called to pass upon the
character of a water of which no series of analyses is
available. In such cases an inspection of the location
from which the water comes should be insisted on, as a
sound interpretation of a water analysis can only be made
with a reasonably full knowledge of the source of the
sample. After a careful sanitary inspection, however,
the comparison of the result of even a single examina-
tion with the normal range for waters of the same class

60 ELEMENTS OF WATER BACTERIOLOGY

may prove of great significance, as a few practical
examples may make clear (Winslow, 1901).

In the spring of 1900 the city of Hartford, Conn.,
was using a double supply, from the Connecticut River
and from a series of impounding reservoirs among the
hills. A single series of plates showed from 4000 to
7000 bacteria per c.c. in the water of the river, while
the reservoir water contained 300 to 900. The abandon-
ment of the river supply followed, and at once the
excessive amount of typhoid fever in the city was
curtailed.

In the fall of 1900, Newport, R. I., experienced an
outbreak of typhoid fever, and when suspicion was
thrown upon the surface water-supply, chemical analysis
of the latter was not wholly reassuring; but there were
only 334 bacteria per c.c. in the water from the taps,
while a well in the infected district gave 6100. It was
no surprise to find, on a further study of the epidemic,
that the well was largely at fault and the public supply
was not.

In the case of ground-water the evidence is usually
even more distinct. At Framingham, Mass., in 1903,
high chlorin content in the public supply, drawn from a
filter gallery beside a lake, had led to public anxiety.
Five samples from different parts of the system showed
averages of i, 2, 2, 2, and 4 bacteria per c.c.; and
taking this in conjunction with the other features of
the bacteriological analysis, it was possible to report
that any pollution introduced upon the gathering
ground had at the time of examination been entirely
removed.

CHAPTER IV

DETERMINATION OF THE NUMBER OF ORGANISMS
DEVELOPING AT THE BODY TEMPERATURE

Relation between Counts Made at 20° and 37°. The

count of colonies upon the gelatin plate measures,
as we have pointed out, the number of the metatrophic
bacteria in general; and the distribution of these forms
corresponds with the decomposition of organic matter
wherever it may occur. In this great class, there are
some species which will grow under a wide variety of
conditions. These are present in most waters in small
numbers, and in sources contaminated with wash from
decaying vegetable matter they occur in abundance.
Other metatrophic forms, however, through a semi-
parasitic mode of life, have become specially adapted
to the peculiar conditions characteristic of the animal
body; and these bacteria possess the property of develop-
ing most actively at the temperature of the human
body, 37° C., which altogether checks the growth of
the majority of normal earth and water forms. The
determination of the number of organisms growing
at the body temperature may throw light, then, on the
presence of direct sewage pollution, since the bacteria
from the alimentary canal flourish under such con-
ditions, while most of those derived from other sources
do not. Savage classifies the bacteria which may

62 ELEMENTS OF WATER BACTERIOLOGY

be found in water under three headings: normal inhab-
itants, like B. fluorescens; unobjectionable aliens
(from soil), like B. mycoides, and objectionable aliens
(from excreta), like B. coli. The first sort and many
of the second sort are generally unable to grow at 37
degrees. This criterion is not an absolute one. Savage,
(1906) reports an experiment in which unpolluted soil,
which had not been manured or cultivated for at least
3 years, was added to tap water, with the result that a
20° count of 76 was increased to 1970, and a 37° count
of 3 was raised to 1630. In this case most of the
bacteria in the soil were capable of development at
body temperature. Experience shows, however, that
the numbers of such bacteria which actually reach
natural waters from such sources are seldom large.
The count at 37°, therefore, helps to distinguish con-
tamination by wash of the soil of a virgin woodland
from pollution by excreta, since in the former case the
proportion of blood-temperature organisms is much
smaller than in the latter. Furthermore, this method
is free from much of the error introduced by the mul-
tiplication of bacteria after the collection of a sample,
as most of the forms which grow in water during storage
cannot endure the higher temperature and conse-
quently do not develop upon incubation. Recently,
for example, water from a spring of good quality was
shipped to the laboratory from a considerable distance.
Gelatin plates showed 4200 bacteria per c.c., but agar
plates at 37° were sterile.

Significance of the 37° Count. A majority of the
English Committee appointed to consider the standard-

DETERMINATION OF ORGANISMS 63

ization of methods for water examination (1904)
recommended the body-temperature count as a stand-
ard procedure. The American Committee on Standard
Methods in its 1905 Report did not recommend this
method even for alternative use. In its last report
(1912), however, it substituted the 37° for the 20° count,
which was dropped out entirely. As we have pointed
out in Chapter II, this course seems to us an unwise
one, and it was formally condemned at the meeting of
the Laboratory Section of the American Public Health
Association in September, 1912, by the passage of a vote
declaring that " ordinary routine examinations of
water for sanitary purposes, and in the control of
purification plants for the present, should include the
determination of the number of bacteria developing
at 20 degrees and 37 degrees." By this action the
body-temperature count is placed on a par with the
2o-degree count as an integral part of sanitary bac-
teriological water examination, a course which has
been strongly urged in earlier editions of this book.

The body-temperature count must, of course, be made
upon agar plates; but otherwise the procedure is
much the same as that already described for the routine
quantitative bacteriological examination in Chapter
II. A 1.5 per cent agar medium has generally been
used, but the Standard Methods Committee in its
recent report recommends only i per cent of agar.
Whipple (1913) points out that this i per cent agar
often gives trouble from the running together of the
colonies on the weaker medium. On the other hand,
a i per cent agar gives higher counts than 1.5 per cent

64 ELEMENTS OF WATER BACTERIOLOGY

agar. He emphasizes the recommendation of Jackson
that the agar used should be dried at 105° C. for
30 minutes, as commercial agar itself contains more or
less water.

The period of incubation ordinarily adopted for body-
temperature counts is 24 hours. Lederer and Bach-
mann (1911) find that with sewage effluents a 48-hour
period at 37° may yield counts from two to six
times as high as those obtained in 24 hours; it is ques-
tionable, however, whether the higher counts thus given
would compensate for the loss of time. The adoption
of a 24-hour period by the Standard Methods Com-
mittee in any case represents an almost universal
practice.

In using agar plates at 37° difficulty is some-
times caused by the spreading of colonies of certain
organisms over the surface of the plate in the water
of condensation which gathers; this may be avoided
by inverting the plates after the agar is once well set,
or still better by the use of plates provided with earthen-
ware tops, as suggested by Hill. The porous earthen-
ware absorbs the water which condenses on it, the
surface of the plate remains comparatively dry, and the
percentage of " spread " plates is reduced from 30
per cent to i per cent (Hill, 1904). Special pains must
be taken, however, to keep the atmosphere in the
incubator nearly saturated with moisture or errors will
be introduced by the excessive evaporation of the
medium used.

Use of Litmus Lactose Agar. Additional evidence
as to the character of a water sample may be

DETERMINATION OF ORGANISMS 65

obtained with little extra trouble by adding a sugar
and some sterile litmus to trie agar medium and
observing the fermenting powers of the organisms
present, as first suggested by Wurtz (Wurtz, 1892)
for the separation of B. coli from B. typhi. It hap-
pens that the most abundant intestinal organisms,
belonging to the groups of the colon bacilli and the
streptococci, decompose dextrose and lactose with the
formation of a large excess of acid. The decomposi-
tion of the latter sugar, on the other hand, is almost
entirely wanting among the commoner saprophytic
bacteria, and therefore lactose is most commonly used
in making sugar agar, i per cent being added to the
medium just before the final filtration (between steps
15 and 1 6 in the standard process of media-making
given on p. 102). In pouring the plate a cubic centi-
meter of sterile litmus solution should be added.
After incubation the colonies of the acid-forming
organisms will be clearly picked out by the redden-
ing of the adjacent agar. Only those colonies which
are sharply colored should be considered as significant,
since certain bacteria of the hay-bacillus group pro-
duce weak acid and faint coloring of the litmus.

When polluted waters are examined in this manner
the number of organisms developing on the lactose-
agar plate will be very high, almost equalling in some
cases the total count obtained on gelatin. Chick
(Chick, 1901), using a lactose-agar medium with the
addition of one-thousandth part of phenol, found, of
colon bacilli alone, 6100 per c.c. in the Manchester
ship canal; 55-190 in the polluted River Severn, and

ELEMENTS OF WATEE BACTERIOLOGY

numbers up to 65,000 per gram in roadside mud. In
an examination of water from the Charles River above
Boston, 37° counts ranging from 9800 to 16,900 have
been found. The average result of 56 examinations of
Boston sewage from July to December, 1903, showed
5,430,000 bacteria per c.c., at 20° and 3,760,000 per
c.c. at 37°, of which 1,670,000 were acid formers. The
average of 25 samples examined in July and August,
1904, showed 1,690,000 bacteria per c.c. at 20° and
1,400,000 at 37°; 429,000 per c.c. were acid formers
(Winslow, 1905).

In unpolluted waters not only the absolute number of
organisms developing at the body temperature is less, but
its ratio to the gelatin count is very different. Rideal
(Rideal, 1902) states that the proportion between the
two counts in the case of a London water in a year's
examination was on the average one to twelve. Mathews
(Mathews, 1893) in 1893 gave the following figures,
the contrast between the ponds and streams, which
were presumably exposed to pollution, on the one
hand, and the wells, springs, and taps, on the other,
being marked.

Source of Water.

Average Number of Colonies per c.c.

Gelatin, 20°.

Wurtz Agar, 37.5°.

Wells springs

1664

153
296
242
273

28
43
95
24

IOI

Reservoirs

Ponds

Taps

Streams

DETEEMINATION OF ORGANISMS 67

According to the English Committee appointed to
consider the Standardization of Methods for the Bac-
terioscopic Examination of Water (1904), the ratio
of the 20° count to the 37° count in good waters is
generally considerably higher than 10 to i. " With a
polluted water this ratio is approached, and frequently
becomes 10 to 2, 10 to 3 or even less."

In 1903 Nibecker and one of ourselves (Winslow
and Nibecker, 1903) made an examination of 259 samples
of water from presumably unpolluted sources in Eastern
Massachusetts, including public supplies, brooks, springs,
ponds, driven wells, and pools in the fields and woods,
with a view to testing the value of the body-temperature
examination. In many cases the samples showed
high gelatin counts, since some of the waters were
exposed to surface wash from vacant land, but the
average number of organisms developing on lactose
agar at 37 degrees was less than 8 per c.c., as will be
seen by reference to the table on the following page.
The highest individual counts obtained were 95 in a
meadow pool, 83 in a brook, and 74 in a barnyard well,
the latter probably actually polluted. Only two samples
in the whole series, one from the well above mentioned,
gave any red colonies on the agar plates.

For a series of shallow surface wells recently examined
by one of us (S. C. P.) a similar relation is indicated
in the table on page 69; 124 samples which showed no
colon bacilli and were apparently unpolluted, gave an
average of 190 bacteria per c.c. at 20° and 8 at 37°
with less than one red colony per c.c.; 23 samples
which did contain colon bacilli averaged 570 bacteria

ELEMENTS OF WATER BACTERIOLOGY

RELATION OF 20° AND 37° COUNTS IN SAMPLES OF
WATER FROM APPARENTLY UNPOLLUTED SOURCES

(WlNSLOW AND NlBECKER, 1903)

Source of Samples.

Number of Samples.

"3 <8
GJ3

Litmus-
lactose-
agar
Plates,

37°.

Dextrose-Broth

Tubes.

Average Number
of Colonies.

Plates Showing
Red Colonies.

<u
*§

3
18

9
18
18

9
15
6

45
95

66
65

3
6
18
3
9

*£
,0

ii§

H °
'o rt

0
0

0
2
0

0
0

0
0
0

2
0
0
0
0

0
0

Number of Tubes
with Gas 2-1.

Cambridge supply (tap)
Wakefield and Stoneham supply
(tao)

7
6
i

3
6

5
4
3
5

i
61

15
i

22
TO
I

i
3

59
1 6

35
141

36
232
13

524
4700

223
18
294
167

365
181
811

47
188

1235
269

3
18

21
I

46
8

0
12

9
4
3i

4
o

2
27

2
I

O
0

2
O
0
I 0
0
0

o
o

0
0

o
o

0
0
0
0
0
0
0

0
0

o
o

0
0
0
0

0
0

2
0

0
0

o
5

0
0
0
0
0
0
0
0
0

0
0
0
0
0

0
0
0

o
o

Lynn supply (tap)

Brookline supply (tap)

Plymouth supply (tap)
Peabody supply (tap)
Dedham supply (tap)
Newburyport supply (tap)
Salem supply (tap)
Taunton supply (tap)
Sharon (well) (tap)
Medford supply (tap)
Milton supply (tap)
Westerly, R. I., supply (tap)
Brooks

Driven wells
Springs
Ponds fed by brooks
Melted snow

Pools in fields
Pools in woods

Roadside pool^

Stream. Blue Hill Reservation. . . .
Flow from rocks
Ponds fed by springs
Drainage from manured pasture . .
Swamps
Rain-water after twelve hours'
heavy fall
Shallow well in Lynn woods

Totals

2 SO

. .. J4

775

DETERMINATION OF ORGANISMS

per c.c. at 20° and 55 at 37° with an average of 7 red
colonies.

BACTERIAL CONTENT OF 147 SHALLOW WELLS

PERCENTAGE OF SAMPLES IN EACH GROUP

Bacteria per c.c.

coli.

I-IO

11-20

21-50

51-

100

101-
500

501-

IOOO

1001-

2OOO

2001-
3OOO

Gelatin, 20°

—

Agar, 37°

—

Red colonies

3_o

Significance of High Temperature Counts. Important
data as to the distribution of bacteria which will develop
at high temperatures may be found in a paper by
Gage (1906), coupled with a suggestive discussion of the
general significance of bacterial ratios. The table on page
70 shows some of the most significant results obtained by
plating waters of various degrees of purity at 20°, 40° and
50°. We have rearranged the lines of the table so as to
make the progression from more to less polluted waters
a fairly regular one. The colony count at 50° shows an
even sharper differentiation than that at 40°. Gage
rightly concludes that " the information to be obtained
by counts of bacteria and acid-producing organisms at
any one of the above temperatures is greatly increased
by the combination of the results obtained from counts
at two or more temperatures."

ELEMENTS OF WATER BACTERIOLOGY

AVERAGE NUMBER OF BACTERIA AND ACID-PRODUCERS
DEVELOPING AT 20°, 40°, AND 50° C., WITH DIFFER-
ENT CLASSES OF WATERS

Bacteria per c.c.

Acid-producing Bacteria.

20° C.

4D.

40° C.
24 Hrs.

50° C.

24 Hrs.

20° C.

4D.

40° C.
24 Hrs.

50° C.

24 Hrs.

Sewage

2,990,000
1,676,000
485,000
146,600
389,000
306,000

15,500
23,300
16,400
16,900
2,800

1,640

35
1,300
670

32
7i5
62

150
64
i ,000

507

7i
49
80

557,500
360,000
126,500
26,IOO
59,300
89,600

1,730
2,030
112

207
212

i,375
4
130
170

3
170

8
o

7,700
29,500
410
8,300
8,000
485

154
54
5
4

2
0

I
2

0
I

0
0

1,940,000
1,032,000
241,000
112,400
292,000
193,000

15,200
I6.OOO
6,700
2,500
1,650

2,360
29
345
1,045

259
16

14
ii

82
8
30
6
8

346,000
283,000
90,000
22,700
45,000
46,000

1,360
1,180
87
134
66

i,i95

2
119
154

I
IOI
O

55
i

r
O
2
0

4,400
24,900
240
8,000
8,000
200

100

2
2

I
O
O

I
I
o

I
I

o
o

Septic effluent
Contact effluent . .

« (i

Trickling filter
effluent
Do

Canal water
River water
Settled canal water
Sand filter effluent
(sewage)
Do

Water filter efflu-
ent
Do

Do
Do

Shallow well

( ( a.

Pond

< i

Spring

i (

Driven well

In warm weather the interpretation of the body-
temperature count must be made less rigid than at
other seasons. Recent investigations have shown that

DETERMINATION OF ORGANISMS

in midsummer bacteria capable of growth at 37° are
more abundant in normal waters than in winter and
spring.

Winslow and Phelps examined 86 samples from
springs, wells, brooks and pools during the winter
and spring months and found only 12 which showed
more than 25 bacteria per c.c. and only 3 which showed
more than 100 per c.c. on lactose-agar. On the other
hand, of 58 samples from corresponding sources
examined in summer, 16 contained more than 100
bacteria per c.c. A series of 20 pools, ponds, and brooks
at Mt. Desert, Me., which were entirely free from
human or animal pollution, were examined in the late
summer of 1906. Only 4 of the 20 samples gave

20° AND 37° COUNTS OF RAW WATER AT WILMINGTON
FILTER PLANT

(WHIPPLE, 1913)

1908.

1909.

Month.

Bacteria per c.c.

Per
Cent

Bacteria per c.c.

Per
Cent.

Gelatin,

20°.

Bile-agar,
37°.

Gelatin,

20°.

Bile-agar,
37°.

January

4630
6830
8800
3170
2OIO
1640
3150
3140
3400
5180
6850
4IOO

124
358
350
149
119
241
432
451
644

439

78
203

2.7

5-3
4.0

4-7
5-9
14-7
13-7
14.4
18.9
8-5

I . 2

4-9

3880
4800
4620
5080
3340
2350
2940
1430
2620
1380
1650
4150

94
260

387
347
229
158
57
230
619
129

97
194

2-3
5-4
8-4
6-7
6.8

6-7
1.9
16. i

22.8

9-4
3-9

4-7

February

March

April

May

Tune

July

August

September

October

November.

December

ELEMENTS OF WATER BACTERIOLOGY

counts under 25 at 37°, and 7 of them gave counts
over 100, the highest figure being 425.

Whipple (1913) gives some figures for the raw water
at the Wilmington, Del., filter plant (page 71) which
bring out the seasonal variation very clearly.

Another special case in which the ratio between the
20° and the 37° count fails to be significant is that
of a water which has been treated with bleaching
powder. Most of the bacteria which survive chlorine
treatment are of course spore formers, many of them
belonging to the hay bacillus group, and it happens
that most of these spore formers can grow at body
temperature. Thus it is common to get counts as high
at 37° as at 20° with such waters, although the absolute
numbers are generally small. This point is illustrated
in the two tables below, showing the results of experi-
mental treatment of Merrimac River water at the
Lawrence Experiment Station and of swimming pool
water at the University of Wisconsin.

COMPARATIVE EFFECTS OF CHLORINE DISINFECTION
UPON 20° AND 37° COUNTS, MERRIMAC RIVER
WATER, AT LAWRENCE, MASS.

(CLARK AND GAGE, 1909)

Bacteria per c.c.

Untreated Water.

Treated Water.

Sample.

20°.

37°.

20°.

37°.

3,400

28,900

130

14,000

3,700

DETERMINATION OF ORGANISMS

COMPARATIVE EFFECTS OF CHLORINE DISINFECTION
UPON 20° AND 37° COUNTS, SWIMMING POOL WATER
AT UNIVERSITY OF WISCONSIN

(TULLY, 1912)

Bacteria per c.c.

Untreated Water.

Treated Water.

Sample.

20°.

37°

20°.

37°.

275

445

480

920

483

5)630

680

19,100

1,140

24,000

1,190

130

120

IO,OOO

1, 080

1,700

690

2,570

78o

2,800

560

Under ordinary conditions it is clear that organisms
growing at the body temperature and those fermenting
lactose are not numerous in normal waters. The
absolute count at 37° seldom exceeds 50, and is rarely
over 10 per cent of the 20° count, except after hot
periods in the late summer ; acid producers are generally
entirely absent. On the other hand, the numbers on
the litmus-lactose-agar plate will be likely to run into
hundreds with a good proportion of red colonies when
polluted waters are examined.

CHAPTER V

THE ISOLATION OF SPECIFIC PATHOGENES FROM
WATER

THE discovery of the organisms which specifically
cause infectious diseases naturally led to the hope
that their isolation from polluted water might become
the most convincing proof of its sanitary quality. The
typhoid bacillus and the spirillum of Asiatic cholera
were in this connection of paramount importance, and
to the search for them many investigators have devoted
themselves.

The Search for Typhoid Bacilli. In the earlier exam-
inations of water for the typhoid bacillus an attempt
was made to use media which especially favored the
growth of the microbe sought for, or to begin with
some process of " enrichment " in which the sample
was incubated under conditions which would favor
the growth of the pathogenic organisms while check-
ing the development of the common water bacteria.
It was apparent that the body temperature and the
presence of a slight excess of free acid furnished such
conditions, and most of the methods suggested rest
upon these principles. Among them, one of the earliest
was that of Parietti (Parietti, 1890), which consists
in the addition of the water to a series of broth tubes

ISOLATION OF SPECIFIC PATHOGENES 75

containing increasing amounts of a solution of 4 per
cent hydrochloric acid and 5 per cent phenol. From
tubes in which growth occurs after 24 hours at 37 degrees
the organisms present may be isolated in pure cultures
by some plating method and identified by subcultures.

The great difficulty with a majority of the enrichment
processes is that the conditions which favor the multipli-
cation of the typhoid bacillus are frequently suited in an
even higher degree to B. coli and other intestinal organ-
isms. Being present in almost all cases in much
higher numbers than B. typhi, these bacteria develop
more abundantly, and effectually mask any disease
germs originally present. In order to obviate this
difficulty, Hankin (Hankin, 1899), after adding suc-
cessively increasing portions of Parietti solution to
tubes inoculated with the water to be tested, selected
the second highest tube of the series in which growth
occurred for the inoculation of a new set, finally plating
as above. He believed that the chance for overgrowth
by this method is somewhat decreased; but in the hands
of other investigators it has not met with marked suc-
cess. Klein (Thomson, 1894) in his investigations,
made use of the Berkefeld filter to concentrate the
organisms in the sample. Some observers abandoned
the enrichment process altogether and recommended
direct plating upon solid media such as phenolated
gelatin or the Eisner (Eisner, 1896) medium, made by
adding 10 per cent of gelatin and i per cent of potas-
sium iodide to an infusion of potato whose reaction
has been adjusted to 30 on Fuller's scale.

In the last five years considerable progress has been

76 ELEMENTS OF WATER BACTERIOLOGY

made in the development of new methods for isolating
the typhoid bacillus. These fall in three distinct
groups: first, the direct isolation by differential, fre-
quently colored, solid media; second, isolation as
above, preceded by enrichment methods; third, isola-
tion, with or without enrichment, preceded by concen-
tration of the organisms by agglutination with typhoid
serum or concentration by chemical precipitation.

Isolation Methods, Using Solid Media. Drigalski
and Conradi (Drigalski and Conradi, 1902), prepared
a medium primarily for the isolation of typhoid bacilli
from excreta, which may also be applied in water
bacteriology. This consists of an agar medium
containing nutrose, sodium chloride, litmus, lactose,
and a dye, " crystal violet "; and it is used in the form
of plate cultures infected by smearing the surface with
a bent glass rod after thorough cooling. The culture
medium is a selective one, ordinary saprophytes fail-
ing to grow, while after 14 to 24 hours at 37°,
colon and typhoid colonies can be readily distinguished
from one another. The colon bacillus produces red,
non-transparent colonies, of variable size and depth
of color, while the typhoid colonies are blue or violet,
transparent and of smaller size, seldom exceeding three
millimeters in diameter.

Endo (Endo, 1904) has suggested the use of a fuchsin-
lactose-agar decolorized by sodium sulphite. Upon
this medium B. coli produces bright red, sharply defined
round colonies in 24 hours at 37°, while B. typhi
gives round, colorless, transparent colonies with thin
margins. This medium has been somewhat modified

ISOLATION OF SPECIFIC PATHOGENES 77

by Gaehtgens (Gaehtgens, 1905) by the addition of
caffein, and he found it of great service in isolating the
typhoid bacillus from stools of patients suffering with
the disease. No attempts were made by him to isolate
the organism from polluted water.

Loeffler (Loeffler, 1903 and 1906) and Lentz and
Tietz (Lentz and Tietz, 1903 and 1905) have made use
of an agar medium containing malachite green. This
medium is supposed to inhibit the growth of B. coli
while favoring B. typhi, and has been recommended
for the isolation of the organism from faeces. Dcebert
(Dcebert, 1900) has shown that certain varieties of
ma!achite green are not suited to this purpose. Nowack
(Nowack, 1905) has also pointed out the same fact,
and ascribed the difference to the presence of dextrin.
He also showed that a medium 0.8 per cent alkaline to
phenol-phthalein is more favorable to B. typhi and
less favorable to B. coli than one neutral to litmus.
With such a medium he found that about 20 per cent
of the typhoid bacilli present develop. Lemke (1911)
has recently reported good success in isolating typhoid
and para-typhoid bacilli from artificially infected
waters by the use of nutrient broth containing 3-5
per cent of sodium chloride and varying amounts of
malachite green as an enrichment medium.

The use of the inhibitive anilin dyes like crystal
violet and malachite green has the disadvantage of also
inhibiting to some extent the development of the weaker
typhoid organisms. Another principle is involved in
the media proposed by Hiss and Hesse. These are
both agar media of lower spissitude than ordinary

78 ELEMENTS OF WATER BACTERIOLOGY

agar and the separation of the typhoid and colon groups
of organisms depends on the greater motility of the
former and their tendency to swim out from the colonies
and form branch-like processes or turbid zones on a
semi-solid medium. The Hiss medium is made up as
follows (Hiss, 1902):

Agar 15 gm. NaCl 5 gm.

Gelatin 15 gm. Dextrose 10 gm.

Liebig's meat extract 5 gm. Distilled water 1000 c.c.
Reaction i per cent normal.

The Hesse medium has been used with great success
by Jackson and Melia (1909). Its general composition
is as follows:

Agar 5 gm. NaCl 8.5 gm.

Witte's pepton 10 gm. Distilled water 1000 c.c.

Liebig's meat extract 5 gm.
Reaction i per cent normal.

Jackson (1909) recommends that the agar used
should be dried for half an hour at 105°, and under
these circumstances 4.5 grams may be used in the
formula instead of the 5 grams recommended by Hesse.
The medium must be stored in an ice chest with saturated
atmosphere. Plates must be made in sufficient dilu-
tion to give a few colonies on the plate; and where
this is done the typhoid colonies are sharply dis-
tinguished from those of B. coli by the fact that they
grow to a considerable size, often several centimeters
in diameter and show a broad translucent or scarcely

ISOLATION OF SPECIFIC PATHOGENES 79

turbid zone between the white opaque centre or nucleus
and the perfectly circular narrow white seam or edge.

Stokes and Hachtel (1912) have suggested a modifica-
tion of the Hesse medium which consists in the increase
of the agar to 5.5 per cent and in the addition of 10
gm. of lactose and 50 gm. of glycerin to the formula
cited above. The agar is dried out at 105 degrees for
half an hour and dissolved in half a liter of water. The
meat extract is added to the other half liter and freed
from muscle sugar by inoculation with the colon bacillus
and incubation for 24 hours at 37 degrees. The sugar-
free broth thus prepared is filtered and to it is added
the peptone, lactose and salt. The two half liter por-
tions of the medium are then mixed and boiled for
30 minutes. The medium is filtered and adjusted to a
neutral reaction, the glycerin is added and the medium
is tinted with azolitmin solution before tubing and steril-
izing. Typhoid and para-typhoid organisms develop
medium-sized, pinkish colonies with concentric rings,
and may thus be distinguished from colonies of B.
alcaligenes, B. proteus and other motile forms. Organ-
isms of the B. subtilis group must be eliminated by
microscopic examination, using the Gram stain.

Preliminary Enrichment. In most cases plate isola-
tion is preceded by some sort of enrichment process
designed to favor the typhoid bacilli at the expense
of the members of the B. coli group. The original
use of carbol broth has been already discussed. In
Europe caffein media have been used for this purpose
and in the United Staes bile media have been strongly
recommended.

80 ELEMENTS OF WATER BACTERIOLOGY

The important fact that caffein has an inhibitory
action on colon bacilli, announced by Roth (Roth,
1903) has given rise to much investigation, and offers
one of the most promising methods for the isolation
of the typhoid bacillus from water. Hoffman and
Ficker (Hoffman and Ficker, 1904) developed methods
for the isolation of B . typhi from faeces and from infected
water by its use in connection with nutrose and crystal
violet. For the isolation from infected water solutions
were prepared as follows :

1. Ten grams of nutrose dissolved in 80 c.c. of steril-
ized distilled water.

2. Five grams caffein, in 20 c.c. sterilized distilled
water.

3. One-tenth gram of crystal violet in 100 c.c. water.
Solutions i and 2 were mixed by shaking together in a
flask, and the mixture poured into a flask containing
900 cubic centimeters of the water to be tested; 10
c.c. of solution 3 were gradually added, and the whole
thoroughly mixed by shaking and then incubated at
37 degrees for not over 12-13 hours. At the end of the
incubation period loopfuls of the solution were smeared
over Drigalski-Conradi plates.

By this method the B. typhi was isolated from
mixtures in river water containing one typhoid bacillus
to 51,867 water bacteria and colon bacilli.

A number of investigations have shown that the
action of the caffein is not as markedly selective as at
first claimed. Kloumann (Kloumann, 1904) obtained
no better results by this method than by the Drigalski-
Conradi medium alone, and Willson (Willson, 1905)

ISOLATION OF SPECIFIC PATHOGENES 81

found that certain strains of B. typhi were inhibited,
while strains of B. coli developed feebly in the presence
of 0.5 per cent of caffein.

In this country marked success in the isolation of
the typhoid bacillus has been attained by the use of
lactose bile as an enrichment medium. Jackson (Jack-
son and Melia, 1909), the principal exponent of this
procedure, recommends that sterilized undiluted fresh
ox gall (or an ii per cent solution of dry fresh ox gall)
containing i per cent of peptone and i per cent of
lactose be made up in 40 c.c. amounts in fermen-
tation tubes to which varying amounts of water, up
to 10 c.c. may be added. After incubation for 48
or 72 hours, he plates on Hesse a gar and he finds
that in the bile medium B. typhi tends to over-
grow B. coli while most other organisms are entirely
suppressed.

Concentration by Agglutination or Precipitation. A
physical concentration of the typhoid bacilli precedes
enrichment or isolation in the procedure recommended
by many authors. Klein, as noted above, accomplished
this by passing the water through a Berkefeld filter.
Other workers have made use of agglutination or chem-
ical precipitation for the same purpose.

The phenomenon of agglutination was made the
basis of a method of isolating B. typhi from water by
Adami and Chopin (Adami and Chopin, 1904). Two-
liter samples of the water were collected in sterilized
bottles (Winchester quarts), and to each was added
20 c.c. of i per cent glucose broth. The sample was
incubated for 18 to 24 hours at 37° C., after which

82 ELEMENTS OF WATER BACTERIOLOGY

10 c.c. portions were withdrawn and placed in long,
narrow test tubes. To each of these tubes enough
typhoid serum of known potency was added to make
a regularly graded series, 1-50, i-ioo, 1-150, and 1-200.
The probable presence of the typhoid bacillus was
manifest by the formation of flocculi within a quarter
of an hour, and agglutination was complete in from
2 to 5 hours.

The tube having the greatest dilution in which
agglutination was apparent was then examined by
breaking off the lower end, containing the precipitate,
washing the sediment two or three times with sterile
water after removing the clear supernatant liquid,
and allowing the bacteria to settle again. The organ-
isms remaining were plated upon various media, and
examined biochemically to determine the true character
of the suspected colonies. It was found that a dilution
of i to 60 was the highest which could be used with the
organisms examined, and it is therefore probable that
high dilutions (greater than 1-60) cannot be success-
fully used.

Investigation of an organism isolated by this method
was made by Klotz (1904), who found the culture
to be not a typical B. typhi, but a form showing
certain points of similarity to both B. typhi and to
B. coli, and probably intermediate between them.
Frost (1910) isolated a bacillus of the B. proteus
group from filtered Potomac water which agglutinated
with typhoid serum in high dilutions. As Klotz points
out, therefore, it is evident that even when a positive
result is obtained with a relatively high dilution of

ISOLATION OF SPECIFIC PATHOGENES 83

typhoid serum, the action may by no means be abso-
lutely specific.

Schepilewski (Schepilewski, 1903) and Altschuler
(Altschuler, 1903) have also used agglutination as a
means of precipitating the bacteria after enrichment
cultivation in broth. The former incubated the cul-
ture at 37° for 24 [hours, then added a serum of
high potency, allowed the mixture to stand for 2
to 3 hours, and then centrifuged. The supernatant
liquid was removed, and the mass of agglutinated cells
broken up by shaking with glass beads and salt solu-
tion. Upon plating upon litmus lactose agar the organ-
isms could be detected. In this way positive isolation
was made from water containing i loopful of a broth
culture in 50 liters of water. Altschuler's method of
enrichment was essentially like that of Schepilewski.
From the surface of the culture developed at 37°,
10 c.c. were removed to a tapering tube provided with a
rubber tube at the bottom. Serum was added in the
proportion of one part in 50, the culture agitated to
release entangled non-agglutinated bacilli and the
sediment run into a tube containing i per cent peptone
and 0.5 per cent salt. The agglutinated mass was broken
up by shaking with sand, and the culture incubated at
37° for 24 hours, then plated on Drigalski-Conradi
plates. The organism was isolated from dilute suspen-
sions in water (150 in i liter) and also from the faeces
of a typhoid patient from which other methods gave
negative results.

Precipitation Methods. A number of methods for
concentrating typhoid bacilli in water by chemical

84 ELEMENTS OF WATER BACTERIOLOGY

precipitation have been tested experimentally, with
some degree of promise. Vallet (Vallet, 1901) was
the first to employ this principle, and made use of
sodium hyposulphite and lead acetate. The mixture
was then centrifuged and the precipitate dissolved in
more hyposulphite. The clear solution was then plated.

Schlider (Schiider, 1903) observed that the lead
salt reacted harmfully upon the bacteria, and that
the hyposulphite should be in excess. In his experi-
ments water was allowed to stand in tall jars for 24
hours. To 2 liters of infected water, 20 c.c. of a 7.75
per cent solution of sodium hyposulphite was added,
and after thorough mixing 20 c.c. of a 10 per cent
solution of lead nitrate. The precipitate, after 20
to 24 hours, was treated with 14 c.c. of saturated sodium
hyposulphite solution and shaken. From the clear
solution 0.2 to 0.5 c.c. portions were streaked upon
Drigalski-Conradi plates which were then incubated
at 37° for 24 hours. Ficker (Ficker, 1904) modi-
fied the process still more by using ferric sulphate,
and dissolved the precipitate with neutral potassium
tartrate. The final solution was then plated on
Drigalski-Conradi medium. Ficker claimed that this
method gives excellent results, 97-98 per cent of the
typhoid bacteria being carried down with the precipitate.

Miiller (Miiller, 1905), in comparing different pre-
cipitation methods, adopted ferric oxychloride as the
most suitable precipitant, because of its quicker and
less destructive action. Willson (Willson, 1905) sug-
gested the use of alum as a precipitant. He added
0.5 gr. alum per liter of water examined. The mixture

ISOLATION OF SPECIFIC PATHOGENES 85

was then centrifuged, and the precipitate suspended in a
small amount of water and plated on Drigalski-Conradi
medium. Nieter (Nieter, 1906) made 20 parallel experi-
ments, using very pure water infected with typhoid
bacilli in varying numbers. By precipitating with ferric
sulphate and sodium hydrate, centrifuging, and then
filtering through a sterile filter he obtained small
numbers of bacteria. Using iron oxychloride as the
precipitant, he confirmed the results of Miiller. By
plating on malachite green agar he was often able to
get positive results when the Drigalski-Conradi medium
failed.

By use of a combination of enrichment and chemical
precipitation, Ditthorn and Gildemeister (Ditthorn
and Gildemeister, 1906) isolated the typhoid bacillus
from enormous artificial dilutions in water. In the
typhoid fever epidemic in Posen, in 1906, it was found
that the bile of those dying from the disease contained
nearly pure cultures of typhoid bacilli. This led the
authors mentioned to use bile and bile agar as enrich-
ment^media. After precipitating by Miiller's method,
the whole of the precipitate was added to 100
c.c. sterile ox bile and grown at 37° for 24 hours,
after which time i c.c. portions were plated. With
extreme dilutions it was found desirable to incubate
for 48 to 72 hours. The results were unsatisfactory
in the presence of large numbers of water bacteria.
It is also pointed out that the iron oxychloride is
bactericidal in 48 hours.

Separation on the Basis of Motility. Drigalski (Dri-
galski, 1906) has suggested the separation of B. typhi

ELEMENTS OF WATER BACTERIOLOGY

from other bacteria in water through its greater motility.
He succeeded in isolating typhoid bacilli from two
springs by the following method: 5 to 10 liters of water
were allowed to stand one to two days in tall milk cans
at room temperature. Samples were taken from the
surface and plated on litmus-lactose agar (Drigalski-
Conradi medium), the amount of water to be used
varying with the contamination.

Starkey (1906) has suggested the use of an apparatus
consisting of a piece of glass tubing bent so as to give
four successive connected loops. This is filled with phenol
broth, inoculated atone end, and incubated anaerobically.
The more actively motile bacilli find their way to the
fourth loop from which they may be isolated by plating.

Review of Suggested Procedures. The methods of
examining water for B. typhi may be conveniently
summarized as follows:

a. By filtration

b. By agglutination

f Schiider's process

c. By chemical I Fischer's process
precipitation 1 Willson's process

( M tiller's process
a. Hoffmann and Ficker's caffein

i. Physical
concentration

2. Enrichment

Examination

of water for

typhoid
bacilli

3. Isolation

4. Identifica- !

tion

process

b. Jackson's lactose bile

c. Parietti's carbol broth

a. Eisner's gelatin medium

b. Endo's medium

c. Loeffler's malachite green

medium

d. Drigalski-Conradi agar

e. Hiss's medium
/. Hesse's medium

a. Morphological and cultural

characters

b. Agglutination

ISOLATION OF SPECIFIC PATHOGENES 87

Of the comparative advantages of these methods
it is still too early to speak with finality. Up to the
present time the use of caffein and lactose bile has
apparently been followed by the best results, and it
seems likely that of the precipitation methods that
employing the oxychloride of iron is the best. Lubenau
(Lubenau, 1907) has made some interesting com-
parisons, using media containing malachite green and
caffein and caffein alone, in which the advantage is
decidedly in favor of the latter.

Identification of the Typhoid Bacillus. At the end
of the process the identification of the pure cultures
isolated is again subject to considerable uncertainty.
The typhoid bacillus belongs to a large group which
contains numerous varieties differing from each other
by minute degrees. The inability to reproduce the
disease by inoculation in available test animals owing
to their natural immunity is a serious drawback;
and the specific biochemical characters of the organism
are, as it happens, mostly negative ones, as shown by
comparison with B. coli, to which it is supposed to be
allied.

COMPARISON OF THE CHARACTERS OF B. COLI AND
B. TYPHI

(HORROCKS, IQOl)
B. COLI B. TYPHI

(i) Surface Colonies, Gelatin (i) Much thinner than those

Plates. — Thicker, and grow more of B. coli, and grow more slowly,

rapidly than those of B. typhi. After forty-eight hours' incuba-

After forty-eight hours' incubation tion at 22° C. they are hardly

at 22° C. they are usually large visible to the naked eye.
and characteristic.

ELEMENTS OF WATER BACTERIOLOGY

(2) Gelatin-stab. — Quick growth
on the surface and along the line
of inoculation.

(2) Slow growth on the surface
like the colonies; along the line of
inoculation, the growth is much
thinner, and often ends below in a
few white points consisting of dis-
crete colonies.

(3) Thin narrow grayish-white
growth, crenated margin not
marked as in B. coli.

(4) No formation of indol.

(5) Unchanged after a month.

(6) Very small amount of acid
produced, requiring not more than
6 per cent of N/io alkali to neu-
tralize it.

(7) No change.

(8) No gas formation.

(3) Gelatin-slope. — Thick, broad
grayish-white growth with a cre-
nated margin.

(4) Witte's Peptone and Salt
Solution. — Indol produced.

(5) Milk.— Coagulated.

(6) Litmus-whey, one week at
37° C. Acid produced usually
requiring from 20 to 40 per cent of
N/io alkali to neutralize it.

(7) Neutral-red Glucose-agar. —
Marked green fluorescence.

(8) Glucose-gelatin and Lac-
tose-gelatin Shake Cultures, and
Glucose-agar-stab. — Marked gas
formation.

(9) Gelatin, 25 oer cent, incu-
bated at 37° C. — Thick film
appears on the surface.

(10) Potato. — As a rule, a thick
yellowish-brown growth.

(n) Proskauer and Capaldi's
Media. No. I, after twenty hours'
growth, medium acid. No II,
Growth, medium neutral or faintly
alkaline.

(12) Nitrate-broth. — Nitrate re-
duced to nitrite.

(13) Microscopical Appear-
ances.— A small bacillus often
like a coccus, not motile as a rule.

(14) Flagella. — Usually i to 3,
short and brittle; sometimes 8 to
12, long and wavy.

(15) Agglutination. — As a rule,
no agglutination with a dilute anti-
typhoid serum.

In addition to the biochemical characters noted above
the typhoid bacillus is characterized by its failure to
produce gas and by its feeble acid production in lactose

(9) No film appears on the sur-
face, but a general growth takes
place throughout the tube.

(10) Thin transparent growth
hardly visible to the naked eye.

(u) No. I, no growth or change
in the reaction of the medium.
No. II, Growth, medium acid.

•(i 2) Reduction of nitrate not so
marked.

(13) Usually longer than B.
coli; highly motile, with a quick
serpent-like movement.

(14) Usually 8 to 12, long and
wavy.

(15) Marked agglutination with
dilute anti-typhoid serum.

ISOLATION OF SPECIFIC PATHOGENES 89

media and by its characteristic colonies on Drigalski-
Conradi agar, Endo medium, Hiss medium, and Hesse
agar. In studying its immunity reactions agglutina-
tion should in all important cases be supplemented
by the Pfeiffer reaction and the absorption test which
will be found described in standard text-books on
bacteriology.

Of the many observers who have reported the isola-
tion of the typhoid bacillus from water, all but the most
recent are quite discredited, on account of the insuf-
ficiency of the confirmatory tests, and even the latest
results should be received with caution. Since the
introduction of the Widal (Widal, 1896) reaction,
founded on the fact that typhoid bacilli examined
under the microscope in the diluted blood-serum of
a typhoid patient lose their motility and " agglutinate "
or clump together, an important aid has been fur-
nished in the diagnosis. Yet serum tests are notably
erratic, and insufficient to identify an organism with-
out an exhaustive study of biochemical reactions.
Many organisms are agglutinated by typhoid serum in
a more or less dilute solution, and agglutinations are
not significant unless obtained in dilutions as great as
1-500 or i-iooo. The discovery of the Bacillus dysen-
teriae of Shiga,1 which closely resembles the typhoid
bacillus, has made the identification of the latter more
dubious than ever. Hiss (1904) has shown that the
fermentation and agglutination reactions of the two
organisms are in many respects alike, and Park and his

* For an account of biology of B. dysenteriae the student is
referred to an article by Dombrowsky; 1903.

90 ELEMENTS OF WATER BACTERIOLOGY

associates (1904) have shown that there are not less
than three distinct types of dysentery bacilli forming
that group.

Isolation of Typhoid Bacilli from Water. The methods
we have been discussing have many of them been used
chiefly for obtaining the typhoid bacillus from fasces,
which is much easier than its isolation from polluted
water. There are, however, a number of cases in which
the organism has undoubtedly been isolated from
polluted water, as by Kubler and Neufeld (Klibler and
Neufeld, 1899), who examined a farmhouse well at
Neumark in 1899, and Fischer and Flatau (Fischer and
Flatau, 1901), who discovered an organism responding
to a most exhaustive series of tests for the typhoid
bacillus in a well at Rellingen in 1901. In these cases
the water was directly plated upon Eisner's medium or
phenolated gelatin with no preliminary process of
enrichment. Willson (Willson, 1905) summarized the
instances in which the typhoid bacillus had been
isolated from infected drinking water, up to 1905, and
included, in addition to the above-mentioned cases, the
following :

1. By Losener, in 1895, from the Berlin water supply.

2. By Conradi, in 1902, from a well at Pecs in Hun-
gary, by use of carbol gelatin plates.

3. By Jaksch and Rau, in 1904, from the water
supply. of Prague, and also from the river Moldau, by
caffein-nutrose crystal violet agar.

4. By Stroszner, in 1904, from a well near Budapest,
by the same method.

Several other instances in which the isolated organ-

ISOLATION OF SPECIFIC PATHOGENES 91

isms gave positive agglutination tests, as well as the
usual cultural reactions, are also cited by Willson.

During the last 5 years a number of successful isola-
tions of the typhoid bacillus have been reported in
America. An organism obtained from the water-supply
of Scran ton, Pa., in 1907, by simple enrichment in
Parietti bouillon, was identified as the typhoid bacillus
by Prof. Fox after a very careful series of tests with
immune sera (Pennsylvania, 1908). The most impor-
tant results have been achieved, however, by Jackson
with lactose bile enrichment and subsequent plating on
Hesse agar. He reports the isolation of B. typhi from
10 c.c. samples of the Grass River at Canton, N. Y.,
and of a pond and stream at Hastings, N. Y., (both
used as sources of water-supply) and from two i c.c.
samples of the Hudson River near Hastings at the
time of the typhoid epidemic there (Jackson and Melia,
1909). Stokes and Hachtel (1910) by the same method
found organisms corresponding to typhoid_in their general
cultural reactions in four samples of surface-waters
(two of them from an impounding reservoir of the
Baltimore supply), in the sediment of a school well
supposed to have caused typhoid fever, in a sewage-
polluted stream and in two samples of market oysters.
These organisms agglutinated with the blood of typhoid
patients in 1/50 and i/ioo dilutions, but with an
immune serum producing agglutination with a standard
laboratory typhoid culture in dilution of 1/25,000
these water organisms would only agglutinate in
dilutions of 1/250 or 1/500. Their identity must
therefore be regarded as somewhat doubtful. The

92 ELEMENTS OF WATER BACTERIOLOGY

same authors (Stokes and Hachtel, 1912) have more
recently reported the isolation of the typhoid bacillus
from the water in the neighborhood of a polluted
oyster bed.

The search for the typhoid bacillus is usually sug-
gested when an outbreak of the disease has cast strong
suspicion upon some definite source of water-supply.
By the time an epidemic manifests itself, however,
the period of the original infection is long past, and the
chances are good that any of the specific bacilli once
present will have disappeared. While elaborate exper-
iments have shown that B. typhi may persist in
sterilized water for upwards of 2 months and in unster-
ilized water from 3 days to several weeks, the number
of the organisms present is always very rapidly reduced.
Even in highly polluted water their number is propor-
tionately small; as is well shown by the experiments
of Laws and Andrewes (Laws and Andrewes, 1894)
who entirely failed to isolate the typhoid bacillus from
the sewage of London and found only two colonies of
the organism on a long series of plates made from the
sewage of a hospital containing forty typhoid patients.
So Wathelet (Wathelet,i895) found that of 600 colonies
isolated from typhoid stools and having the appearance
characteristic of B. coli and B. typhi only 10 belonged
to the latter species.

Epidemiological evidence confirms these results and
indicates that the number of typhoid bacilli even in
polluted water probably is never very great, while the
fate of Lowell and Lawrence in 1890-91 and the more
recent epidemics at Butler, Pa., and Ithaca, N. Y.,

ISOLATION OF SPECIFIC PATHOGENES 93

demonstrate that even a small number of virulent
organisms can bring about an almost wholesale infection.
Indeed, if the virulent organism were as abundant as
some results would indicate (Remlinger and Schneider,
1897), the human race would long since have been
exterminated. A negative result in testing for typhoid
bacilli has no significance and there is danger that it
may be misinterpreted if the fact that it has been made
comes to public knowledge. In spite of this danger,
however, and in spite of the laborious and time-con-
suming nature of the process, the increasingly large
number of positive isolations in recent years indicate
that it is well worth trying in cases of special importance.
The search for the typhoid bacillus should of course
never supersede the examination for colon bacilli,
since the latter are so much more numerous in water
and so much more easily identified. Because of these
facts, colon bacilli will continue to be our best index
of pollution, while the positive isolations of the typhoid
bacillus will supply additional proof of the deadly
character of a water containing it.

Other Bacteria of the Typhoid Group Related to
Intestinal Disease. The typhoid bacillus and the colon
bacillus (which will be fully discussed in succeeding chap-
ters) stand at the opposite ends of a series of many dif-
ferent varieties of organisms which are intermediate in
their properties between B. typhi and B. coli, all being
non-spore-forming, non-liquefying rods, which produce
a more or less characteristic growth on solid media.
Durham (1898) divided these forms into three main
divisions, grouped, respectively, about B. typhi, B.

ELEMENTS OF WATER BACTERIOLOGY

enteritidis and B. coli. Organisms of the first division
ferment neither dextrose, lactose nor saccharose; those
of the second ferment dextrose but not lactose; and
those of the B. coli division form gas in both these
sugars. The relationship of the commonest species
is indicated in tabular form below:

BACTERIA OF THE COLON-TYPHOID GROUP

Species.

Dextrose.

Lactose.

Gas For-
mation.

Acid Pro-
duction.

Gas For-
mation.

Acid Pro-
duction.

B. alcaligenes

None
None
None

Active
Active
Active

Active

None
Slight
Distinct

Strong
Strong
Strong

Strong

None
None
None

Active

None
None
None

Slight
Slight
Slight

Strong

B typhi

B dysenterise ... ...

B enteritidis

Paratyphoid bacilli

Hog cholera bacillus

B coli

In the typhoid division, B. alcaligenes and B. dysen-
teriae are the best-known forms, besides B. typhi itself.
B. alcaligenes stands at the lower end of the whole
series in fermentative power. B. typhi forms a slight
initial acidity in milk and a slight acidity in dextrose
broth, while the reaction of B. alcaligenes in sugar
media is always alkaline. B. dysenteriae, on the other
hand, differs from B. typhi in the direction of the B.
enteritidis group, producing a well-marked acid reac-
tion, but no gas in dextrose media. B. typhi and B.
dysenteriae are, of course, also distinguished by their
specific serum reactions.

ISOLATION OF SPECIFIC PATHOGENES 95

The second great division of the colon-typhoid
bacteria is the hog cholera group, or the Gartner group,
as Durham (1898) called it. As defined by him, it
differed from the typhoid group by gas formation in
dextrose, and from the colon group by the production
of a final alkaline reaction in milk. It includes the
Gartner bacillus (B. enteritidis), the hog cholera bacillus
(B. cholerae suis), and the paratyphoid bacilli. Some
of these forms, the paratyphoid bacilli, for example,
and B. enteritidis (isolated in cases of meat poisoning),
produce intestinal disease in man.

There is no doubt that water is sometimes the
means of distributing the germs of dysentery and
diarrhoea, as shown by the decrease of these diseases
in Burlington, Vt, (Sedgwick, 1902), and other com-
munities where pure water-supplies have been sub-
stituted for polluted ones. Thresh (Thresh, 1903)
described an epidemic of over 1000 cases of diarrhoea
with 14 deaths, which occurred in England at Chelmsford
and Widford, and was undoubtedly spread by the
public water-supply. A somewhat similar epidemic
of dysentery occurred in Warren and Kittanning, in
Pennsylvania, in 1906, which was unquestionably due
to contamination of the water, in this case a river-
supply. It is possible that the examination of water
for the B. dysenteriae may in the future ."! help to
throw important light on the sanitary condition of
a water.

Starkey (1909 and 1911) believes that all organisms
giving the general reactions of the Gartner and para-
typhoid groups are significant and warrant the con-

96 ELEMENTS OF WATER BACTERIOLOGY

demnation of a water-supply. The difficulty, however,
is that while non-acid-forming bacteria of this general
type are sometimes found in faeces, they are also
found in other habitats, and they are less abundant
proportionately, in polluted than in stored and safer
waters. If true dysentery and paratyphoid bacilli
can be isolated and identified by serum reactions it is,
of course, highly important. Houston' (1911), however,
has recently tested the method suggested by Starkey
(1906) for isolating these forms and found that it gave
negative results even with a water artificially infected
with about 14 typhoid bacilli and 21 Gartner bacilli
per c.c. In his own studies Houston reports that in
the examination of 13,442 microbes from polluted
river water he found only one member of the Gartner
group; and in another study of 20,771 colonies he
found only 2 typhoid-like forms.

Isolation of the Cholera spirillum. The isolation of
the cholera spirillum from water can probably be accom-
plished with somewhat less difficulty than is encoun-
tered in the case of B. typhi. Schottelius (Schottelius,
1885) was the first to point out the necessity for grow-
ing this organism in an alkaline medium, and Loeffler
(Loeffler, 1893) found that its isolation from water
could be successfully accomplished by_adding 10 c.c.
of alkaline pepton broth to 200 c.c. of the infected
water and incubating for 24 hours at 37 degrees, when
the organism could be found at the surface of the
medium.

Somewhat earlier than this Dunham (Dunham, 1887)
had made a special study of the chemical reactions of

ISOLATION OF SPECIFIC PATHOGENES 97

the cholera spirillum and found that the organism would
grow abundantly in a solution containing i per cent
peptone and 0.5 per cent salt .(Dunham's solution),
producing the " cholera-red or nitroso-indol reaction."
This medium was brought into practical use by Dunbar
(Dunbar, 1892), who succeeded in isolating the organisms
from the water of the Elbe in 1892, during the cholera
epidemic at Hamburg.

Koch (Koch, 1893) prescribed the following method
for the isolation of the organism from water:

To 100 c.c. of the water to be examined is added i
per cent pepton and i per cent salt. The mixture is
then incubated at 37 degrees. After intervals of 10,
15, and 20 hours the solution is examined microscopically
for comma-shaped organisms, and agar plate cultures
are made which are likewise incubated at 37 degrees.
If any colonies showing the characteristic appearance
of the cholera spirillum are found, these are examined
microscopically, and if comma-shaped organisms are
present, inoculations are made into fresh tubes to be
further tested by means of the indol reaction and by
inoculation into animals.

Stokes and Hachtel (1912) have suggested the use
of a modified Hesse agar containing starch for the
isolation of the cholera spirilla, which produce acid on
such a medium, while the colon-typhoid organisms do
not. The glycerin and lactose are omitted from the
medium described on p. 79 and 10 gms. of soluble
starch are added. The intestinal spirilla as a class
form round, spreading, pinkish colonies on the starch
medium, while colonies of other intestinal bacilli remain

98 ELEMENTS OF WATER BACTERIOLOGY

blue. The medium is best used after the Koch enrich-
ment method described above.

Other pathogenic organisms have been isolated from
waters, according to the accounts of numerous investi-
gators, but from the sanitary point of view the typhoid
and cholera bacilli are of most importance, since these
are manifestly the germs of disease most likely to be
disseminated through this medium. For the detection
of B. anthracis and other spore-forming pathogenic
bacteria which may at times gain access to water from
stockyards, slaughter-houses, etc., the method suggested
by Frankland (Frankland, 1894) may be adopted.
The water to be examined is heated to 90 degrees for
2 minutes and then plated, the characteristic colonies
of the anthrax organism being much more easily dis-
cerned after the destruction of tne numerous non-
sporing water bacteria.

CHAPTER VI

THE COLON GROUP OF BACILLI AND METHODS FOR
THEIR ISOLATION

The Colon Group of Bacilli. The Bacillus coli was
first isolated by Escherich (Escherich, 1884) from the
faeces of a cholera patient. It was subsequently found
to be a normal inhabitant of the intestinal tract of man
and many other animals, and to occur regularly in
their excreta, and on this account it became of the
highest interest and importance to sanitarians, since
its presence in water-supplies was regarded as direct
evidence of sewage pollution.

Specific disease germs are difficult to isolate even
when they are present; and water may of course be
grossly polluted with sewage without any specific
disease germs being there at all. All sewage-polluted
water, however, is potentially dangerous, since where
faecal matter exists, disease germs are at any time likely
to appear. A test for faecal material as distinguished
from infected material is, therefore, essential; and
for such a test the colon group of bacilli are specially
well suited. They are not dangerous in themselves,
but they are significant as indices 'of the probable
presence of disease germs.

The so-called Bacillus coli may be described as a short,

100 ELEMENTS OF WATER BACTERIOLOGY

usually motile rod, with diameter generally less than one
micron and exhibiting no spore formation. It often ap-
pears in pairs of rods so short as to suggest a diplococcus.
It decolorizes by the Gram stain. It forms thin, irregu-
lar translucent films upon the surface of gelatin, called
" grape-leaf colonies " by the Germans, produces
no liquefaction, and gives a wire-nail-like growth in
stick cultures. It forms a white translucent layer of
characteristic appearance upon agar, produces a more
or less abundant, moist, yellowish growth on potato,
and turbidity and some sediment in broth; it ferments
dextrose and lactose with the formation of gas of which

the ratio is approximately, - - = - , as ordinarily deter-

C(J2 I

mined; a strong acid reaction and gas are produced
in many other sugar-containing media. The organism
generally gives a characteristic reaction in esculin
media and typically reduces neutral red, changing
its color to canary yellow with a greenish fluorescence.
It grows in the Capaldi-Proskauer media, forming
acid in the albumin-free medium, No. i, and giving
a neutral or alkaline reaction in the peptone-mannite
medium No. 2. It coagulates casein in litmus milk,
and reduces the litmus with subsequent slow return
of the color (red), and generally forms indol in peptone
solution. Many cultures of this organism are fatal
to guinea pigs when the latter are inoculated sub-
cutaneously with one-half c.c. of a 24-hour bouillon
culture, and most cultures produce death when this
amount is inoculated intraperitoneally. Although not
a spore-forming bacillus, and in general not possessing

THE COLON GROUP OF BACILLI 101

great resistance against antiseptic substances, B. coli
is less susceptible to phenol than are many other forms,
especially certain water-bacteria.

We have spoken as if Bacillus coli were a single defi-
nite organism. As a matter of fact it is a name applied
to a considerable group of distinct forms which may
be split up almost as far as one wishes by the applica-
tion of various biochemical tests. The " colon bacillus,"
as we have pointed out, usually does not liquefy gelatin
and reduces neutral red and coagulates milk, and
produces indol; but there are closely allied forms which
differ from the type in one or more of these respects.
The colon group, as Smith (1893) long ago pointed
out, may first be divided into two distinct subtypes
according to the action of the organisms upon saccharose.
One subtype forms gas and acid in saccharose media
and the other does not. Winslow and Walker (1907)
have found that those strains which ferment saccharose
attack raffinose also, and point out that these two
sugars which behave alike are those which lack the
aldehyde grouping characteristic of dextrose and lactose.
The application of tests in other carbohydrate media,
such as dulcite, adonite, inulin, etc., make it possible
to recognize perhaps a hundred distinct types each
characterized by a particular combination of reactions.

The results obtained by the " colon test " will of
course depend largely upon the definition of what a
colon bacillus is; and there is marked disagreement
upon this point among different observers. Konrich
(1910) tabulates the tests used by 34 different workers,
All of them defined the colon bacillus as a Gram-

102 ELEMENTS OF WATER BACTERIOLOGY

negative non-spore-forming rod, but there was unanimity
in no other respect; 26 of the 34 included the formation
of acid and gas in dextrose media, 22 the coagulation
of milk, 21 the formation of indol, 18 the formation
of acid and gas in lactose media, and 18 the failure to
liquefy gelatin. No other test was used by more than
13 out of the 34 observers. Konrich (1910) himself
found that of over 600 colon-like organisms from faeces,
all produced gas in dextrose broth, 79 per cent formed
acid and 77 per cent gas in lactose broth, 65 per cent
coagulated milk, 59 per cent fermented dextrose at
46°, 54 per cent reduced neutral red, 38 per cent formed
indol.

Ferreira, Horta and Paredes (i9o8a) studied 117
strains of lactose-fermenting bacilli from human faeces.
All proved to be motile and Gram negative, all coag-
ulated milk and produced fluorescence in neutral red
media, none liquefied gelatin in 15 days, all but one
formed indol. Dextrose, lactose, maltose, galactose,
and mannit were fermented by almost all strains,
while gas was formed in saccharose by 38 per cent,
in dulcite by 69 per cent and in inulin by 12 per cent
of the strains studied. The " Proskauer reaction "
(apparently the Voges-Proskauer reaction, though it
is not quite clearly stated) was positive only 8 times
out of 117 strains in dextrose, only 7 times (out of 48
strains tested) in galactose, and not at all in dulcite or
inulin; lactose and maltose, on the other hand, showed
it in almost every case. Copeland and Hoover (1911)
report that out of 3000 colon-like organisms fermenting
-lactose bile 65 per cent gave positive tests for B. coli

THE COLON GROUP OF BACILLI 103

in milk, nitrate solution, pepton solution and gelatin;
28 per cent failed to produce indol and 5 per cent did
not reduce nitrates. Numerous other results indicat-
ing similar variations are cited in Chapters VII and
VIII.

Where shall the line be drawn? The English bac-
teriologists usually require in addition to the morpholog-
ical characters mentioned above, motility, non-liquefac-
tion of gelatin, fermentation of dextrose and lactose
media, coagulation of milk, production of indol, and
reduction of neutral red. The usual American pro-
cedure has included reactions in dextrose broth, milk,
peptone solution (for indol) , gelatin (absence of liquefac-
tion), and reduction of nitrates. Of late years, how-
ever, there has been a growing feeling that such
arbitrary definitions went either too far or not far
enough. The whole group of lactose-fermenting bacilli
is characteristically of intestinal origin. That we be-
lieve to have been clearly established by results to
be cited later in this chapter and in the succeeding one.
A differentiation between various sub-types of this
group can only be properly justified by the fact that
some of them are less resistant in water than others,
and hence are indicative of fresh and recent pollution.
There is some evidence that such is the case which will
be discussed in Chapter VIII, but whatever may be
concluded from the somewhat conflicting opinions
on this point it appears certain that, in temperate
climates at least, the whole class of lactose fermenters
should be absent from safe water supplies. As the
Committee on Standard Methods of Water Analysis

104 ELEMENTS OF WATER BACTERIOLOGY

(1912) wisely concludes, " The entire group is typical
of the presence of faecal matter when water or sewage
examinations are to be considered." The Committee
defines the group as a whole by the following character-
istics: " Fermentation of dextrose and lactose with
gas production, short bacillus with rounded ends,
non-spore-forming, facultative anaerobe, gives positive
test with esculin, grows at 20° on gelatin and at 37°
on agar, non-liquefying in 14 days on gelatin. Gram-
staining negative." This definition again, however,
includes debatable elements. It seems to us very
doubtful whether there is sufficient evidence to warrant
making the esculin reaction a general criterion of the
colon group; and bacteria which liquefy gelatin more
or less slowly grade into otherwise identical non-
liquefying forms by almost imperceptible degrees.

We hesitate to add another to the long list of arbitrary
definitions of the much-defined colon bacillus; but it
does seem to us important to get down to rock bottom.
From this standpoint we believe that the colon group
may be defined as including all aerobic non-spore-
forming bacilli which produce acid and gas in dextrose
and lactose media. For practical purposes the test
may be further reduced to positive reactions in a lac-
tose fermentation medium, growth on an aerobic agar
streak, and microscopic examination, since almost all
forms which ferment lactose ferment dextrose as well.
The relative importance of the various subdivisions of
this general group will be discussed in Chapter VIII.

Isolation of Colon Bacilli by Direct Plating. The Wurtz
litmus-lactose-agar plate (Wurtz, 1892), as noted in

THE COLON GROUP OF BACILLI 105

Chapter IV, furnishes one ready method for the isola-
tion of B. coli from water, and it was used by Sedgwick
and Mathews for the purpose as early as 1892 (Mathews,
1893). The process is based upon the fact already
alluded to, that B. coli readily ferments lactose with
the formation of acid. If, therefore, plates are made
with agar containing both lactose and litmus, the colon
colonies develop as red spots in a blue field. Since
organisms other than B. coli (notably the streptococci)
may also develop red colonies, it is necessary to examine
them further. This is done by fishing from isolated
colonies, replating and inoculating into other media
for identification.

The plate method of isolation is recommended by the
Committee on Standard Methods of Water Analysis
(1912) for sewages and polluted waters, in which colon
bacilli are present in i c.c. or less. They recommend
that Petri dishes with porous covers be used and that
incubation be carried out at 40° instead of 37°. For
success in the use of this method it is necessary to
get a sufficient dilution so that colonies may be well
isolated, and to this end it is advisable that a number
of different dilutions be employed, a series of plates
being prepared from each. Under any conditions the
detection of the colon bacillus is seriously hampered
by the development of other forms. Certain observers
have therefore added phenol to the agar medium, com-
bining the effect of high temperature and an antiseptic
to check the growth of water-bacteria. Copeland for
this purpose added to his tubes .2 c.c. of a 2 per cent
solution of phenol (Copeland, 1901). Chick (Chick,

106 ELEMENTS OF WATER BACTERIOLOGY

igoo) found that 1.33 parts of phenol in 1000 materially
decreased the number of colon bacilli which would
develop, while i part gave very satisfactory results,
the plates showing pure cultures of B. coli. The addi-
tion of antiseptics in this way is always open to the
objection that weaker strains may be killed and lost.

In Germany the Endo medium and the Conradi-
Drigalski medium have been extensively used for the
direct isolation of colon bacilli with excellent results.
The composition and use of these media have been
discussed in Chapter V. It does not appear that they
have sufficient advantages to compensate for the dif-
ficulty in preparing them.

The Use of Preliminary Enrichment Media in the
Isolation of Colon Bacilli. The test for the colon
bacillus may be made more delicate by a preliminary
cultivation of the sample in a liquid medium for 24
hours at 37°, thus greatly increasing the propor-
tion of these organisms present before plating. As
suggested in the classic researches of Theobald Smith
(Smith, 1892), this method may be made approx-
imately quantitative by the inoculation of a series
of tubes with measured portions of the water. If,
for example, of ten tubes inoculated each with yiir
of a cubic centimeter, four show B. coli, we may
assume that some 40 of these organisms were present
in the cubic centimeter. Irons (Irons, 1901), in a
comparative study of various methods for the isola-
tion of B. coli, showed that the preliminary enrichment
frequently gave positive results when the results of
the direct use of the agar plate were negative, and

THE COLON GROUP OF BACILLI 107

concluded that " where the amount of B. coli is small
and the colony count large, the lactose plate for plating
water direct is inferior to the dextrose fermentation-
tube." Gage came to a similar conclusion (Gage, 1902).

The medium most commonly used in the United
States prior to 1906 for preliminary enrichment was
ordinary broth to which i.o per cent of dextrose had
been added, and the reaction brought to the neutral
point. Into each of a number of fermentation-tubes
of this medium a measured quantity of the water to be
examined is inoculated, and the culture is incubated for
24 hours at 37.5° C. It used to be customary to incu-
bate for 48 hours. Recent experience has, however,
shown that a 24-hour period gives approximately the
same results if the production of gas rather than any
specified amount of gas is the criterion of a positive
test. Longley and Baton (1907) found that of 1091
enrichment tubes giving positive tests after 48 hours
only 173 showed no gas in 24 hours; of these latter
only two contained B. coli. The advantage of saving a
day is so great as to warrant the adoption of the shorter
period. At the end of 24 hours at least, the tubes
are examined for gas formation. If gas is found, a
small amount of the culture should be added, after
suitable dilution, to litmus lactose agar and plated.

With polluted waters it will be found advantageous to
plate out on the first appearance of gas (4-8 hours).
It has been shown by one of us (Prescott, i902b) that a
very rapid development of B. coli takes place in the
first few hours after dextrose solutions are inoculated
with intestinal material, and a nearly pure growth of

108 ELEMENTS OF WATER BACTERIOLOGY

colon bacilli often results, while other bacteria multiply
more slowly. With highly polluted waters gas forma-
tion will probably begin within 12 hours, but with fewer
colon bacilli present the duration must be increased.
If the period of incubation be too long continued,
trouble in the subsequent steps of the isolation may be
encountered because of overgrowths by the sewage
streptococci, or other forms which check the growth
of the colon bacilli in the later stages of fermentation
and finally kill them out. Even with pure cultures of
colon bacilli Clemesha (i9i2b) has shown that sugar-
broth tubes may be almost sterile after 4 days.

When it is desired to examine samples larger than
i c.c. it becomes necessary to modify the enrichment
process by adding the nutrient material to the water
instead of the reverse. For this purpose dextrose
broth or phenol-dextrose broth (consisting of broth
with 10 per cent dextrose, 5 per cent peptone, and
.25 per cent phenol) may be added to the sample of
water to be enriched as suggested by Gage (Gage, 1901).
Generally 10 c.c. of the broth is added to 100 c.c. of the
water. The sample is then incubated at 37° for 24 hours,
and if at the end of that time growth has taken place, a
cubic centimeter is inoculated into a dextrose tube.

Advantages and Disadvantages of the Dextrose Broth
Fermentation Tube. Experience with the dextrose
broth fermentation tube as a first step in the isola-
tion of colon bacilli soon led to the conclusion that a
fair idea of the sanitary quality of water could be
obtained from the results of this test taken by them-
selves and without the further process of isolating

THE COLON GROUP OF BACILLI 109

specific cultures. It appeared that a rather definite
proportion of tubes showing a characteristic fermenta-
tion proved on further examination to contain bacilli
of the colon group; and it was therefore suggested
that the dextrose broth test alone might be used as a
rapid " presumptive " test. The underlying principle
of this method is that B. coli develops rapidly* in dex-
trose broth with gas formation of from 25 to 70 per
cent of the capacity of the closed arm of the fermenta-
tion tube. Of this gas approximately one-third is
carbon dioxide and two-thirds hydrogen, that is, as the

gas formula is generally expressed, — — = — .

C(J2 I

In testing a water by this method a series of samples,
in suitable dilution, .001, .01, .1, i.o, or 10 c.c., are added
directly to the dextrose-broth tubes and incubated for
24 hours at 37°.

On measurement of the gas, if the results above given
are obtained, the reaction is considered typical. If the
amount of gas is between 10 and 25 percent or more than
70 per cent, or the percentage of carbon dioxide is greater
than 40, the reaction is considered atypical. If no gas
forms, or less than 10 per cent, the test is called negative.

In recent years. Irons (Irons, 1901) was perhaps the
first to call attention to the value of this method,
stating that " when the dextrose tube yields approx-
imately 33 per cent of CO2, Bacillus coli communis is
almost invariably present." In the next year the
reliability of the fermentation test as an indication of
B. coli was worked out by Gage (Gage, 1902) as given
in the table on p. no:

110 ELEMENTS OF WATER BACTERIOLOGY

I C.C.

too c.c.

Number of samples tested .

ei 72

I 27 C

Number giving preliminary fermentation
Per cent of latter proved to contain coli

1036
7O

474

Whipple (Whipple, 1903) examined a large number
of surface-water supplies by this " presumptive test "
and obtained striking results, shown in the following
table. The waters are arranged in six groups according
to the results of sanitary inspection, group I including
waters collected from almost uninhabited watersheds,
and group VI waters too much polluted to be safely
used for domestic purposes.

PERCENTAGE OF SAMPLES OF WATERS OF VARIOUS
SANITARY GRADES GIVING POSITIVE TESTS FOR B.
COLI WHEN DIFFERENT AMOUNTS WERE EXAMINED

(WHIPPLE, 1903)

Group.

O.I C.C.

1.0 C.C.

10 C.C.

ioo c.c.

500 c.c.

I
II

0.0

r o

3-5

7 3

20.8
ICO

50.0
60 o

50.0

60 o

III

o o

7 o

cJO O

so o

60 o

4.0

6.8

41 . 7

67 .0

75 -°

5.0

13.0

7<.o

IOO.O

? o

2O 2

7S O

80 o

IOO O

In view of these results Whipple suggested the fol-
lowing provisional scheme of interpretation:

Presumptive Test for Bacillus Coli.

O.OI C.C.

O.I C.C.

I.O C.C.

10. 0 C.C.

100 C.C.

Safe

-f

Reasonably safe
Questionable

o
o

+
+

Probably unsafe.

Unsafe

-f

THE COLON GROUP OF BACILLI 111

It is undoubtedly true that a negative presumptive
test is generally obtained with unpolluted waters. For
example, in a study previously cited, Winslow and
Nibecker (1903) reported that of 775 dextrose-broth
tubes inoculated from 259 unpolluted sources only 41
showed gas. On the other hand, it is equally true
that in a large proportion of cases colon bacilli are
isolated from positive dextrose-broth tubes. Longley
and Baton (1907) in the examination of 3553 samples
of Potomac water obtained positive tests 794 times,
while B. coli was actually present 529 times; 67 per
cent of the presumptive tests were therefore correct.
Gage (1902), in the Massachusetts work cited above,
found that 70 per cent of his fermented dextrose tubes
contained B. coli.

The work of recent years has made it clear, however,
that both the coincidence of negative presumptive tests
with the absence of B. coli and the general coinci-
dence of positive presumptive tests with the presence
of B. coli, are open to disastrous exceptions.

The errors in the dextrose broth test are both positive
and negative ; it may lead to the inference that bacteria
of the colon group are present when they are not,
and it may fail to show them when they are really there.
In the first place, with some waters, positive presump-
tive tests may be obtained when colon bacilli are not
present. According to Clark and Gage (1903) there
are 58 well-described species of bacteria which give
the presumptive test in dextrose-broth, of which 23
are widely separated from the B. coli group. An
unpublished investigation by Winslow and Phelps

112 ELEMENTS OF WATER BACTERIOLOGY

indicates that the result of the dextrose broth test is
markedly influenced by the factor of temperature. Their
work consisted in the examination of 185 samples of
water from 90 different sources, ponds, brooks, pools,
wells and springs in five different States, Maine, New
Hampshire, Massachusetts, Michigan and Virginia, at
three different seasons of the year. All the waters
examined were, as far as could be determined, free
from specific pollution, although washings from roads
or pastureland might have had access to some of them.
Most of the sources were undoubtedly unpolluted and
the examination of 119 samples for B. coli yielded only
12 positive results. The presumptive test, however,
was obtained in a large proportion of the cases, and
much more often in summer than in winter or spring,
as indicated in the table below.

DEXTROSE BROTH FERMENTATION IN 185 SAMPLES OF
NORMAL WATERS AT DIFFERENT SEASONS

(WINSLOW AND PHELPS)
Percentage of Positive Results

Summer,
1906.

Winter.

Spring.

Summer,
1907.

Framingham, Mass. . . .

Ann Arbor, Mich

Exeter, N. H

Richmond, Va

Mt. Desert, Me

All stations

-27

The Ann Arbor waters in this series included a number
of driven wells and the Mt. Desert sources were mountain
brooks and ponds of the highest sanitary quality.

THE COLON GROUP OF BACILLI 113

Fromme (1910) in this connection reports the results
of 673 colon tests made on the water of the Elbe during
a period of a year and a half. We have calculated
from his figures the average results for the winter
months and the summer months in the table below.
It is evident that typical colon bacilli are nearly twice
as numerous in the cold weather (for reasons discussed
in Chapter I) while organisms fermenting dextrose
broth but proving not to be B. coli are absolutely
more abundant and relatively much more abundant
in summer.

GAS PRODUCERS AND B. COLI IN ELBE WATER

(AFTER FROMME, 1910)

Positive Dextrose
Broth Tests.

B. Coli
Isolations.

Per Cent of Dex-
trose Broth Tests
Showing B. Coli.

October-March
April-September. . . .

415

258

363
170

8?
66

Phelps and Hammond (1909) cite a very interesting
case of the same phenomenon in the case of a ground
water. A deep well at a hospital in Trenton, N. J.
was temporarily polluted from a leaking sewer and after
the source of pollution had been removed the condi-
tion of the water was carefully studied for a period of
two months. During the period between Sept. 10 and
Oct. 12 (the pollution being removed on Sept. 19)
of 107 dextrose-fermenting microbes isolated 40 failed
to produce gas in lactose broth; during the period
between Oct. 12 and Nov. 9, 52 out of 64 dextrose-fer-
menting microbes failed to give gas in lactose broth.
All through the investigation organisms of low fer-

114 ELEMENTS OF WATER BACTERIOLOGY

mentative power, many of them liquefying gelatin,
were , present, but their numbers relatively increased
during the period after the original pollution had been
removed.

Much valuable light is thrown upon the significance
of these positive dextrose broth results in the absence
of the colon group by the investigations of Clemesha
(igi2a). In a careful study of 46 samples of human
faeces and 25 different samples of cow dung, includ-
ing about 3500 different colonies isolated by various
methods, only about 5 per cent belonged to the class
fermenting dextrose but not lactose. In rivers and
ground waters, on the other hand, this group made
up 24 per cent of the colon-like organisms present,
and in lakes with long storage, 58 per cent. The table
below shows the relative increase of the lactose negative
forms with the natural purification of rivers in four
dry months following rain and also their relative
increase in settled and filtered water as compared with
the raw river water. The numbers dealt with are too
small to give entirely uniform results, but the general
trend is clear.

PERCENTAGE OF DEXTROSE-FERMENTING ORGANISMS
FAILING TO FERMENT LACTOSE. CALCUTTA WATER
SUPPLY

(CLEMESHA, 1912*)

Month

Oct.

Nov.

Dec.

Jan.

Feb.

Condition of river

Heavy

Very

Clear-

Clear

ram

muddy

ing

Raw river water (80 colonies)

Settled water (20 colonies) . .

Filtered water (60 colonies) .

THE COLON GROUP OF BACILLI 115

Clemesha confirmed these results by a long series
of examinations of naturally and artificially polluted
waters, all tending to show that with fresh pollution
most of the dextrose fermenters ferment lactose as
well, while with storage there is a relative increase in
the lactose-negative forms. Careful studies of the
history of faecal mixtures in water showed that the
resistant form was a particular type, called by Clemesha
Bacillus P. Houston (1911), reports similar results
for London waters. Of 12,744 specimens of raw river
water containing dextrose-fermenting organisms 81
per cent gave positive results in lactose and formed
indol as well, thus indicating the presence of the colon
group. Of 18,960 specimens of filtered water contain-
ing dextrose-fermenting organisms only 51 per cent
gave positive results in lactose and formed indol.
Clemesha (191 2b), in an analysis of Houston's results,
shows that the preponderance of dextrose-positive
lactose-negative forms is here not due to the Bacillus P,
which Clemesha found in India, but to two different
forms.

From all these investigations it is clear that the dex-
trose broth test does not bear a constant relation to
the presence of the colon group, since another type of
organisms fermenting dextrose but not lactose is rela-
tively much more abundant in stored and relatively
pure water, particularly in warm weather.

Phelps and Hammond (1909) have pointed out a
rather serious error in the routine isolation of B. coli,
as it used to be practised in this country, due to the
presence of this group of organisms which fail to form

116 ELEMENTS OF WATER BACTERIOLOGY

gas in lactose broth. The five standard tests for B.
coli which have been most generally adopted in the
United States included gas production in dextrose
broth and coagulation of milk, but not gas production
in lactose broth. It was supposed that types pro-
ducing gas and acid in dextrose media and coagulating
milk, but failing to form gas in lactose broth, would
be rare. In the particular polluted well studied by
Phelps and Hammond, however, such forms were
very common, outnumbering true colon bacilli four
to one during the latter part of the investigation when
the pollution was less recent. Two workers following
the standard methods but using dextrose broth for
enrichment on the one hand and lactose broth on the
other would in 50 per cent of the samples tested have
reached opposite conclusions as to the presence or absence
of B. coli, the isolations begun with dextrose broth
being apparently positive and those begun with lactose
broth being negative.

It is also clear on the other hand that the dextrose
test as ordinarily used may be negative when colon
bacilli are present. This is due to the interaction of
various bacteria in the fermentation tube and to the
solution and escape of gases which often prevents the
production of the typical gas formula. Of 43 cultures
isolated by Fuller and Ferguson (1905) at Indianapolis,
1 8 showed less than 20 per cent of gas after 48 hours
in the enrichment tube, and n showed less than 10
per cent. Hale and Melia (1910), working with a pure
culture of B. coli in unsterilized water (containing no
other gas former) report that of 818 tubes showing gas

THE COLON GROUP OF BACILLI 117

only 474 or 58 per cent gave between 25 and 70 per
cent of gas in the closed arm with 25-40 per cent carbon
dioxide.

Stamm (1906) and others have pointed out that the
ratio of carbon dioxide to hydrogen changes with
the age of the culture. At first the proportion of the
former to the latter is as two to one, and later, in the
same tube, the ratio is reversed. More recently,
Longley and Baton (1907), in one of the ablest and most
fruitful of recent contributions to water bacteriology,
have made it clear that neither of these quantitative
determinations is of importance if made in an ordinary
open tube. They show, first, that the total amount
of gas formed by B. coli varies widely, from 10 to 80
per cent, the mode of the curve being found, not at 50,
but at 35 per cent. Secondly, they show that the
proportion of carbon dioxide present is a function of
the total amount of gas. They find that when grown
in an atmosphere of CCb, B. coli produces a gas which
consists of about 3 parts of carbon dioxide to one of
hydrogen. Assuming that the gas originally formed
by B. coli has always about this composition, and that
the absorption of CCb by the medium is the chief cause
of the differences observed in the gas which collects
in the closed arm, the gas ratio would vary directly
with the amount of total gas; the more rapidly gas is
formed, the greater the proportion of CO2 remaining
unabsorbed. Calculation on this basis gives a curve
very close to the observed data.

These criticisms apply only to the fermentation
test made in an open tube and uncorrected for the

118 ELEMENTS OF WATER BACTERIOLOGY

absorption of CCb. Keyes (1909) and others have
introduced more exact methods based on the collec-
tion and analysis of all gases formed in a vacuum
and in a paper shortly to be published by L. A. Rogers,
W. M. Clark and B. J. Davis of the Bureau of Animal
Industry (kindly loaned to us by Mr. Rogers) it is
shown that the gas ratio when accurately determined
is highly characteristic for certain members of the
colon group.

Not only is it true that little reliance can be placed
on the exact gas formula in the open dextrose broth
tube, but cultures of the colon group may be actually
overgrown and lost by the multiplication of other
forms. This is particularly true when the waters are
heavily polluted or when large samples are examined.
Hunnewell and one of us (Winslow and Hunnewell,
i902b) found that of 48 samples of certain polluted
river waters 18 showed B. coli when i c.c. was inoculated
directly into dextrose broth, while in only 4 cases was
a positive result obtained after preliminary treatment
of 100 c.c. in carbol broth. In 153 samples from
presumably unpolluted water B. coli was found 5
times in i c.c. and n times by the examination of
the larger sample. It wrill be noted that these results
were obtained with carbol broth for the enrichment of
the larger samples and carbol broth is less liable to
overgrowths than dextrose broth.

Whipple (Whipple, 1903) notes that 2.9 per cent of
some samples of water examined by him gave positive
tests with .1 c.c. but not with i c.c., while 4.3 per
cent gave positive tests with .1 c.c. or i c.c. and negative

THE COLON GROUP OF BACILLI 119

tests with 10 c.c. Again, in another series of samples
examined, of those which gave positive tests in smaller
portions 5.3 per cent were negative in 10 c.c., 4.7 per
cent in 100 c.c., and 7.7 per cent in 500 c.c.

Fromme (1910) has made an interesting study of
this point and reports that of 59 samples of water of
good quality which showed B. coli in small portions
25 per cent gave negative results in larger portions;
while of 654 samples of polluted waters 33 per cent
gave negative results in large portions and positive
results in smaller ones. These results are of value as
indicating the greater liability to loss by overgrowth
in polluted waters; but the absolute figures are much
higher than workers in this country obtain when the
enrichment cultures are carefully watched and plat-
ings made from them at an early period.

The Use of Phenol Broth as an Enrichment Medium
to Check Overgrowths. As has already been stated,
phenol has less inhibitory action upon B. coli than
upon normal water-bacteria, and it was hoped that a
broth containing this substance might be employed
for preliminary enrichment with advantage, its inhibitory
power checking the overgrowing forms, but not B. coli.
This medium was used in place of dextrose broth for
many of the studies made in connection with the Chicago
drainage canal (Reynolds, 1902). Phenol broth con-
sists of ordinary broth to which o.i per cent phenol
is added, and the method of procedure is to add i c.c.
of the water to 10 c.c. of the sterilized phenol broth
and incubate at body temperature for 24 hours. Litmus-
lactose-agar plates are then made and the examination

120 ELEMENTS OF WATER BACTERIOLOGY

of the red colonies carried out as described for the
dextrose-broth method. It has unfortunately proved,
however, that with waters of fairly good quality the
phenol interferes with the colon bacilli themselves to
a serious extent. The dextrose broth furnishes a more
delicate test than the carbol broth when the number of
colon bacilli present is small, as is clearly shown by the
following table from Irons:

PROPORTION OF POSITIVE RESULTS IN TESTS OF POL-
LUTED AND UNPOLLUTED WATERS BY DEXTROSE
FERMENTATION-TUBE AND CARBOL-BROTH METHODS

(IRONS, 1901)

Dextrose

Fermentation-
tube.

Carbol-broth
Method.

+ - ?

4 ?

Polluted waters

33 3i

38 30 i

Relatively unpolluted waters

56 38 2

37 61 21

The Eijkman Test. Another enrichment test, which
has been extensively used in Germany for checking
the development of overgrowing forms and limiting
the bacteria in the fermentation tube to the colon
group, is the Eijkman test, which depends on the use
of a high temperature (46°) (Eijkman, 1904). There
is no doubt that such a procedure cuts out the water
bacteria, and Christian (1905), Neumann (1906), and
Thomann (1907) have reported good results from its
use. Hilgermann (1909), too, urges the value of the
Eijkman test and concludes that the colon-like bacilli,
which fail to grow at 46°, are characteristic of compara-
tively unpolluted waters. Other observers maintain,

THE COLON GROUP OF BACILLI 121

and apparently with good reason, that the conditions
of this test are too severe and eliminate many intestinal
bacilli of undoubted significance. Nowack (1907) found
that laboratory cultures of B. coli often fail to produce
gas in Eijkman's medium at 46°, unless large numbers
are introduced. With some strains an inoculation
of over a million bacteria was necessary to cause gas
formation.

Konrich (1910) compared the Eijkman enrichment
method (dextrose peptone water at 46°) and that of
MacConkey (dextrose-bile-salt peptone water at 42°)
with 57 water samples and obtained only about
70 per cent as many positive results with the former
as with the latter method. With artificial emulsions
of pure cultures and of faeces even greater differences
were manifest. These studies showed conclusively
that incubation at 46° prevents the development of
great numbers of colon bacilli and is unsuitable for an
enrichment process. In comparing the two media
(dextrose peptone water with and without bile salts)
at the same temperature, 37°, he obtained essentially
similar results.

Fromme (1910) has shown that during the first 5
hours in various enrichment media colon bacilli multiply
more rapidly at 46° than at 37°. After that time,
however, their development is checked. At 12 hours
the numbers at the two temperatures are about equal
and between 12 and 24 hours the numbers increase
much more rapidly at 37°.

The Neutral Red Reaction. Other special media
have been suggested for rapid routine water analysis,

122 ELEMENTS OF WATER BACTERIOLOGY

of which those containing " neutral red," one of the
safranine dyes, have been somewhat fully studied.
Rothberger (Rothberger, 1898) first pointed out that B.
coli reduces solutions of this substance, the color chang-
ing to canary-yellow accompanied by green fluorescence.
Makgill (Makgill, 1901), Savage (Savage, 1901), and
other English observers, as well as Braun (1906), in
France, report favorable results from the use of this
test; but according to American standards, Irons
(Irons, 1902) and Gage and Phelps (Gage and Phelps,
1903) conclude that the group of organisms giving a
positive neutral red reaction is too large a one to give
very valuable sanitary information.

Stokes (1904) urged the use of lactose broth with
the addition of neutral red, and believed that the pro-
duction in this medium of 30-50 per cent of gas with a

- gas formula and the change of neutral red to canary

yellow in the closed arm of the fermentation tube was
characteristic for B. coli.

The Lactose Bile Test for the Colon Group. On the
whole, by far the most satisfactory results in making a
rapid test for the colon group have been obtained by
the use of media containing bile salts, a procedure
the development of which in this country we owe
principally to Jackson (1906).

MacConkey (1900) long ago suggested the use of
media containing bile salts (sodium taurocholate)
for the differentiation of B. coli and B. typhi, and bile-
salts media have been used by various English observers
(MacConkey, 1901; MacConkey and Hill, 1901) for

THE COLON GROUP OF BACILLI

123

the isolation of sewage bacteria. Jackson studied
the action of various bile media and showed their
selective inhibitory action in the striking table quoted
below. His important contribution to the subject,
however, was the discovery that ox bile itself could
be used as a culture medium, and that it was easier
to prepare, cheaper and more effective than combina-
tions of meat infusion with the purified bile salts.

SELECTIVE ACTION OF BILE SALTS

(JACKSON, 1906)

Bacteria

per c.c.

Uncon-
taminated
Well.

Contami-
nated Pond.

Suspension
of
Faeces.

Gelatin 20°

O2O

2700

3 50 ooo

900,000

Agar 37°

170

450,000

900,000

Bile agar,* 37°

300,000

900,000

Lactose bile agar,* 37° .

250,000

675,000

Lactose bile agar,* 37° .

250,000

6oo,OOO

Bile agar, 37°

60,000

900,000

* Bile diluted, i.i.

Jackson suggested the use of fresh ox bile containing
i per cent of lactose as a presumptive test instead of
dextrose broth. In particular he hoped that this medium
would be free to a great degree from the negative results
due to overgrowths in polluted waters. He reported
275 examinations of badly contaminated waters, in
which 65 per cent of the samples failed to give the
dextrose-presumptive test, and only 10 per cent failed
to show gas in lactose bile. In a more recent communica-
tion, Jackson (1907) reports that in the examination of

124 ELEMENTS OF WATER BACTERIOLOGY

5000 samples of water at the Mt. Prospect Laboratory,
the bile medium has proved uniformly satisfactory.
He recommends incubation for 72 hours, results being
commonly obtained, however, after 48 hours; and he
considers any tube showing 25 per cent gas as positive.
In a series of examinations carried out at the Institute
of Technology, 16 per cent of the positive tubes showed

COMPARATIVE PRESUMPTIVE TESTS WITH DEXTROSE
BROTH AND LACTOSE BILE

(SAWIN, 1907)

Source.

Percentage of Samples Giving Positive Tests for
B. Coli.

Dextrose Broth.

Lactose Bile.

O.I C.C.

I.OC.C.

IO.O C.C.

O.I C.C.

I.OC.C.

lo.oc.c.

I Deep wells

O
0
T5-°

IO.O
IO.O

IO.O

1.0
IO.O

15-7

5-0
10.5

IO.O

15-0

IO.O

15-0

21 .0
4O.O
26.0

5-o
35-o

O
O

5-o

5-2
IO.O
IO.O

o
o

0
5-2

5-o
iQ-5

5-o

O
6.0
15.0
31.0
iS-o
50.0

i5-o
30.0

2 Shallow wells

3 Lake
4 Lake

5 Lake
6 Lake
7 Lake
8 Lake

Average, Nos. 3, 4, 5,
6, 7, 8

8-3
47.0

26.3
36.8

II .0

72.2

37-6
55-i

23.6

55-5

73-7
68.9

5-0

50.0
30.0
40.0

4-3

75-0
90.0
73-5

26.0

84.2
85.0
78.1

9 River
10 River
ii River.

Average. Nos. 9, 10, n
1 2 Brook

36.7

47-7
50.0
25.0

55-i

63-2
73-7
25.0

66.0

72.2
78.9

8.2

40.0

60.0

84.2
87-5

79-4

90.0
90.0
93-7

82.4

84.2
90.0
81.2

13 Drainage

14 Sewage

Average, Nos. 12, 13, 14

40.9

53-9

53-1

77-2

91.2

85-1

THE COLON GROUP OF BACILLI 125

gas in 24 hours, 73 per cent after 48 hours, and the
remaining 27 per cent only after 72 hours; but the
Committee on Standard Methods (1912) believes that
the forms which fail to develop in 48 hours are attenuated
forms of little practical significance. Sawin (1907)
reports comparative results with dextrose broth and
bile on different classes of waters, the most striking
of which are tabulated on p. 124.

Like other enrichment methods which eliminate com-
peting forms it is no doubt true that the lactose bile
test cuts out some weak colon bacilli. As a presumptive
method, however, it is far superior to dextrose broth,
giving a higher proportion of positive tests with polluted
waters and a lower proportion of erroneous positive
tests with waters of good quality. In an examination
of 176 surface waters in eastern Massachusetts, carried
out under our direction, B. coli was isolated 70 times.
The dextrose-broth test was positive 120 times, an error
of 70 per cent; while the bile test, alone, was positive
78 times, an error of only n per cent. The tabulated
results of these experiments indicates fairly the merits
of the bile medium for preliminary enrichment and as a
presumptive test.

PRELIMINARY AND COMPLETE RESULTS OF DEXTROSE
BROTH AND BILE TESTS. 176 SURFACE-WATERS

Preliminary Positive
Results.
(Gas Formation.)

Final Positive Results.
(B. Coli).

Dextrose broth

1 2O

Lactose bile

126 ELEMENTS OF WATER BACTERIOLOGY

Hale and Melia (1910) have also made a valuable
comparison of a number of presumptive tests as applied
during a period of 2 years to 85 samples of Manhattan
water (surface) and 160 samples of Brooklyn water
(largely ground-water). Their principal results are
tabulated below.

RESULTS OF VARIOUS PRESUMPTIVE TESTS

(HALE AND MELIA, 1910)
245 Samples of New York Water

Medium.

Percentage of Positive Results in

O.I C.C.

I.O C.C.

10.0 c.c.

Dextrose broth (standard gas formula) . . .
Lactose bile

0.4
0.8
0.8

0.8
6-3

7.0

8.2

15-5

12.0
31-4

30-7
38.4
51.2

57-6
73-5

Lactose-peptone bile
Dextrose broth, all tubes showing 5% gas
transplanted to bile

Dextrose broth, 5% gas and over, called
positive

This table indicates very clearly the fallacies of the
dextrose broth tube. Counting all gas formers in this
medium as positive indicates much too high a value
and including only the tubes showing the standard gas
formula gives much too low a value. This conclusion
is based on the assumption, warranted by the results
of many workers, that incubation in dextrose broth
followed by reinoculation into lactose bile (fourth line
of the table) gives with reasonable accuracy the real
number of colon bacilli present. By this standard
the use of plain lactose bile with these waters is seen
to give results which are also much too low; but lac-
tose peptone bile approximates closely to the truth.

. THE COLON GROUP OF BACILLI 127

Of course it must be remembered that the advantages
of lactose-bile over dextrose broth are partly due to
the inhibiting effect of the bile salts and partly to the
use of lactose instead of dextrose which cuts out the
dextrose-positive lactose-negative group to which allusion
has been made earlier in the chapter. The relative
importance of these two factors, lactose and bile, is
well brought out in a study by Stokes and S toner (1909).
These authors have compared a considerable series
of preliminary enrichment tests followed by final
isolation in dextrose broth, lactose broth and lactose
bile. Of 567 colonies from positive dextrose broth
tubes only 52 per cent were colon bacilli; of 3752
colonies from positive lactose broth and lactose bile
tubes, 88 per cent of the lactose broth colonies and 95
per cent of the lactose bile colonies were B. coli.

With sewages and heavily polluted waters in par-
ticular the lactose-bile medium has proved of the
greatest value. When a large proportion of sewage
is present the colon bacilli are fresh from the intestine
and apparently able to resist the antiseptic salts. On
the other hand, the high numbers of other bacteria
present make the danger of overgrowths particularly
great. With waters of fair quality, such as those
with which we ordinarily deal in sanitary water analysis,
lactose bile is open to the same objection as phenol
broth and the Eijkman test though in less degree.
It inhibits not only the overgrowing forms but the
weaker representatives of the B. coli group itself, and
the net effect is to diminish positive results.

Hale and Melia (1910) inoculated unsterilized water

128 ELEMENTS OF WATER BACTERIOLOGY

(shown to contain no gas formers) with a pure culture
of B. coli and stored it for different periods and under
different conditions, testing at intervals by various
presumptive tests. The colon bacilli lived for 8-10
days at 37°, for 38-75 days at 20°, and 77-84 days
at 8° C. Comparison of presumptive tests with plate
counts on litmus-lactose agar showed that a gas test
in dextrose broth corresponded to an average of 4
bacteria and a positive test in lactose bile to 39 bacteria.
In general the dextrose broth showed gas in one dilu-
tion higher than the lactose bile; and the difference
increased with the attenuation due to prolonged sojourn
in water.

These results of Hale and Melia make it clear that
the selective action of bile salts upon the various mem-
bers of the colon group may perhaps be an advantage
rather than a disadvantage for practical sanitary
purposes. The Committee on Standard Methods (1912)
recognizes that lactose bile does inhibit certain weaker
members of the colon group, but believes that these
attenuated organisms indicate only remote pollution
and are of little significance. They say, " In the
interpretation of the sanitary quality of a water,
it. is best to discount the presence of attenuated
B. coli and to be sure to obtain all vigorous types.
The lactose bile medium accomplishes both of these
objects."

The advantages of the dextrose broth enrichment of
weak colon bacilli and of the elimination of gas-forming
organisms other than B. coli by bile may both be obtained
as pointed out above by inoculating lactose bile from

THE COLON GROUP OF BACILLI . 129

the dextrose broth. Hale and Melia (1910) find that
this gives results in as high a dilution as by the use of
dextrose broth and with the clear-cut results of lactose
bile.

The Aesculin Test. A test for the colon group which
has attracted much interest during the last few years
is the fermentation of the glucoside aesculin. B.
coli effects a hydrolytic splitting of this substance,
producing sugar and a substance called aesculetin,
which reacts with iron citrate to produce a dark brown
salt. Harrison and van der Leek (1909) have used
broth or agar made up with i or 2 per cent Witte's
peptone, .5 per cent sodium taurocholate, .1 per cent
aesculin and .05 per cent citrate, and find that the
black colonies with a black halo produced by the colon
group of organisms are highly characteristic. Of
60 samples of water which showed blackening of
aesculin broth, all proved to contain B. coli. Hale
and Melia (1911) have shown that two species of
streptococci, the aurococcus, and the bacillus of
pneumonia, also give the aesculin reaction and that
in the absence of the bile salt which the aesculin
medium contains a number of other forms may ferment
this glucoside.

On the whole we do not think it has been shown
that aesculin has sufficient differential value to war-
rant its inclusion in enrichment media to be used in
the colon test. It appears that the anaerobic B. welchii
is practically the only form outside the colon group
which produces a characteristic reaction in lactose
bile and not in aesculin-bile salt media. It may prove

130 ELEMENTS OF WATER BACTERIOLOGY

worth while to use aesculin to exclude this form, but
its occurrence in ordinary water work is so rare that the
extra complication seems hardly justified.

The Use of Synthetic Media for the Isolation of the
Colon Group. Dolt (1908), working in Prof. Gorham's
laboratory at Brown University, has attempted with
success to substitute synthetic media of simple and
known composition for the usual meat-infusion-peptone
media used in the isolation of B. coli. He first found
that colon bacilli will grow readily on a medium con-
taining asparagin and sodium or ammonium phosphate.
He then attempted to substitute for asparagin various
simple organic substances similar in their structure
to the cholic acid of the bile which exerts a selective
action in favor of the colon group. He finally succeeded
in preparing two media which promise to be of con-
siderable value in permitting the growth of B. coli
while checking other forms. The first of these media
is made up as follows: 500 c.c. of a 3 per cent solution
of purified agar is mixed with an equal portion of a
solution of i per cent glycerin and 0.2 per cent
(NH^HPCU. It is neutralized with sodium hydroxide
and i per cent of lactose is added before sterilization.
In the second medium 5 gm. of ammonium lactate is
substituted for the glycerin and i gm. of Na2HPO4
for the ammonium phosphate. These media proved
to have a considerable selective value, cutting out most
of the water bacteria; but like all such selective media
they cut out a good many of the colon bacilli too. The
results of a single test are shown in the table below.
The procedure well merits further study, however.

THE COLON GROUP OF BACILLI

131

COMPARATIVE RESULTS OF ORDINARY AND SYNTHETIC
AGAR MEDIA

(DOLT, 1908)

Total Colonies.

Red Colonies.

B. Coli.

Standard agar

Glycerin agar

Ammonium lactate agar

The Use of Liver Broth for the Isolation of a Maximum
Proportion of Gas-forming Bacteria. The media we
have been discussing, phenol broth, dextrose broth
incubated at 46°, and bile, are designed to cut down
the gas producers which appear in ordinary dextrose
broth so that only vigorous typical members of the
colon group are able to develop. For special research
purposes when it is desired to get the largest possible
proportion of gas formers of all kinds, there are other
media which give even more positive fermentation
results than dextrose broth itself. The most important
of these is the liver broth of Jackson and Muer (1911)
made up with beef liver, peptone, dextrose and potas-
sium-acid-phosphate. The Committee on Standard
Methods of Water Analysis (1912) recommends that
if " a study of all gas-forming bacteria, including
attenuated forms, is desirable, then liver broth should
be employed in preference to the usual dextrose broth,
as it gives a larger number of attenuated forms, has
better rejuvenating power, and gives fewer anomalies
and greater and more rapid gas production." In
order to avoid attenuation or inhibition transplants

132 ELEMENTS OF WATER BACTERIOLOGY

should be made from this enrichment medium after
6-12 hours at 37°.

Isolation of Pure Cultures from the Enrichment Tube.
In case one does not rely upon a " presumptive " test
alone but desires to study the organisms present in
detail the isolation upon a solid medium, usually
litmus-lactose-agar in this country, must follow the
enrichment process. Since the enrichment tube was
inoculated with a known amount of water all further
work is purely qualitative, and it is only necessary to
obtain such a number of colonies upon the lactose plate
that the isolation of a pure culture shall be easy. In
practice the following procedure has been found gen-
erally successful. After the enrichment tubes have
been incubated for 12 to 24 hours at 37°, from those
which show gas, one loopful is carried over to a tube
containing 10 c.c. of sterile water, and of this water one
loopful is taken for the inoculation of the plate.
Ordinarily this will give colonies which are sufficiently
well separated, but a second plate, inoculated from the
dilution water with a straight needle instead of a loop,
furnishes a desirable safeguard. With practice it is
possible to effect a proper seeding more rapidly by
barely touching the tip of a straight needle to the broth
in the fermentation tube and transferring this directly
to the agar. The touch must be a very light one, how-
ever, or the colonies on the plate will be too thick for
proper isolation.

The litmus-lactose-agar plates made in this manner
should be incubated for from 12 to 24 hours at the body
temperature (37°), at the end of which time, if B. coli is

THE COLON GROUP OF BACILLI 133

present, red colonies upon a blue field will be visible.
The Htmus-lactose-agar plate may become blue again
after 48 hours, owing, presumably, to the formation of
amines and ammonia by the action of the bacteria
upon the nitrogenous matter present. If the dilution
is too low, the resulting colonies will be small and
imperfectly developed, making it difficult to be sure
of pure cultures for the subsequent tests. A great
number of colonies will also prevent the change of
reaction from acid back to alkaline.

In the selection of those red colonies which are to be
fished from the litmus-lactose-agar plate the appearance
of the growths must be closely noted. A colony of
irregular contour, surrounded by a very faint area of
reddening, will probably belong to some member of the
B. mycoides group (Winslow and Nibecker, 1903);
small, compact, bright-red colonies are characteristic
of the streptococci, and Gage and Phelps (Gage and
Phelps, 1903) have pointed out that of these there are
two types, one of a brick-red color, and of such con-
sistency as to be readily picked up by the needle-point,
and the other smaller and of an intense vermilion
color. The colonies of the colon bacillus are usually
well formed, pulvinate on the surface and fusiform
when growing deeper down.

If no red colonies appear on the litmus-lactose-agar
plate after a positive result in dextrose broth one of
four things has occurred: There may be an organism
present which forms gas in dextrose but no acid in
lactose; there may be present forms which individually
fail to attack lactose but growing together, symbiotically,

134 ELEMENTS OF WATER BACTERIOLOGY

produce gas in dextrose; B. welchii or some other
form which will not grow on aerobic plates may have
produced the gas; or an organism orginally present
and capable of fermenting both sugars may have been
overgrown and lost in the enrichment tube. If plates
are made on the first appearance of gas the likelihood
of the latter possibility will be reduced to a minimum.
Neither of the first two contingencies has any sanitary
significance; as we have seen, bacteria which ferment
dextrose and not lactose are not specially characteristic
of pollution. In general, therefore, the absence of red
colonies on the agar plate may be considered a negative
result. If red colonies are present they must be sub-
cultured and examined further.

The agar streak made from the litmus-lactose-agar
plate shows after 24 hours certain marked character-
istics. The most distinct types are two, the abundant,
first translucent, later whitish and cheesy growth,
covering nearly the whole surface of the agar, character-
istic of B. coli and its allies, and a very faint growth,
either confined strictly to the streak or made up of faint
isolated colonies, dotted here and there over the surface.
The latter cultures are typical of the sewage streptococci,
and a microscopic examination will generally settle
their status at once. Of the more luxuriant growths,
some of which are stringy to the needle, many will
generally prove to be atypical, and if any of the weakly
fermenting forms (B. mycoides) are present a dull
wrinkled growth will be produced.

The various tests which may be applied to the cul-
tures after they have been isolated, the subgroups

THE COLON GROUP OF BACILLI 135

into which the colon group may be divided by their
use, and the significance of the results obtained will
be discussed in Chapter VIII.
Practical Routine Test for the Colon Group. As has

been pointed out above the aggregation of lactose-
fermenting bacilli which we call the colon group may
be almost indefinitely subdivided by the application
of a more or less elaborate series of diagnostic tests.
Each observer in the past drew up a scheme of what
he believed to be essential tests and called all the
bacteria which failed to conform to them " atypical."
The more of such " atypical " forms a particular
worker includes the greater will be the number of positive
isolations. The definition of this or any other bacterial
species is more or less arbitrary; we consider as true
colon bacilli those which fulfil a particular set of tests,
and class as pseudo-colon organisms those which do
not. If we find, having established such an arbitrary
standard, that the colon bacillus, as determined by it,
is found in waters known to be polluted, and not, as
a rule, in those known to be free from pollution, the
sanitarian can afford to ignore the theoretical question
of specific values and make confident use of the practical
test. In order that results may rest on a sound basis
of comparable data for various waters, it is of course
essential, however, that a standard set of reactions
should be agreed upon by sanitary bacteriologists.

After a considerable period of uncertainty, in which
each observer used the procedure which happened to
appeal to him, the attainment of comparative results
was made possible by the establishment of standard

136 ELEMENTS OF WATER BACTERIOLOGY

methods of procedure by bodies of authoritative posi-
tion, both in England and America. In 1904 an
English Committee, appointed to consider the Stand-
ardization of Methods for the Bacterioscopic Examina-
tion of Water, presented a series of obligatory tests
and optional tests; and in 1905 the Committee on
Standard Methods of Water Analysis of the American
Public Health Association drew up a set of diagnostic
characters for B. colL The latter corresponded in
general with the plan developed by the Massachusetts
State Board of Health (Massachusetts State Board of
Health, 1899) and long in use at the Massachusetts
Institute of Technology, and involved the determina-
tion of morphology, motility, fermentation of dex-
trose broth, coagulation of milk, production of indol
and reduction of nitrates. The English standard pro-
cedure corresponded quite closely to this (Committee
appointed to consider the Standardization of Methods
for the Bacterioscopic Examination of Water, 1904),
although it differed from the American method in cer-
tain respects.

Since these standards were formulated 8 years ago
their artificial nature has been made more and more
manifest to those who have used them. A still more
fundamental question, however, has pressed itself
upon the practical analyst. Each lactose bile presump-
tive test involves the use of a single tube of medium
and the work is complete in 48 hours. The " com-
plete " test as used in America required the use of
seven different media and took 9 days to complete,
since the gelatin subculture must be incubated for

THE COLON GEOUP OF BACILLI 137

at least a week. Granting that the lactose bile test
gave us the whole colon group and that the " typical
B. coli " giving characteristic reactions in milk, peptone,
gelatin and the nitrate solution, only the more sensitive
members of the group, indicative of recent pollution,
was the extra information gained worth the additional
trouble? Under the conditions which generally obtain,
in the United States at least, it appears not. The colon
test at best is an approximate one, and its results
are usually only expressed in decimal fractions, positive
in 10 c.c., i c.c., or .1 c.c. for example. From 70 to
90 per cent of the bacteria which give the lactose bile
test prove to be " typical " B. coli on any of the defini-
tions ordinarily used. This makes a difference so slight
as to be almost negligible. We cannot condemn a
water because it contains 10 rather than 7 colon bacilli
in a given proportion. It may be that under tropical
conditions such as those described by Clemesha (to
be discussed in a following chapter) certain forms of
the colon group persist for a long time in stored waters
from which disease germs have disappeared. These
resistant forms must, however, be studied by a much
more elaborate procedure than the 7 tests in the old
American Standard method; and it seems clear that
they do not occur in large numbers in temperate cli-
mates. For our conditions the whole group of forms
which produce gas in lactose bile should be absent from
safe waters.

The Committee on Standard Methods of Water
Analysis in its last report (1912) apparently takes
this ground, although its discussion of the problem

138 ELEMENTS OF WATER BACTERIOLOGY

is distinctly ambiguous. In one section of the report
" Recommended Procedures for Treating Samples "
complete isolation and the use of the old confirmatory
tests in fermentation tubes, milk, gelatin tube, peptone
solution and nitrate broth are discussed. In another
place it is pointed out that the entire colon group is
typical of the presence of faecal matter and the follow-
ing " Quantitative Test for the B. coli Group " is
recommended :

" Add the quantities of water or sewage to be tested
in dilutions by tenths, sufficient in number to obtain
a negative test, to fermentation tubes holding at least
40 c.c. of lactose bile, incubate at 37° C. and note
the production of gas. Gas often forms in a few hours
when large numbers of B. coli are present, but the
standard time for observing gas production is 48 hours.
Small numbers of somewhat attenuated B. coli may
require 3 days to form gas. Attenuated B. coli does
not represent recent contamination and all B. coli
not attenuated grows readily in lactose bile. No
other organism except B. welchii gives such a test in
lactose bile. B. welchii is of rather rare occurrence
in water, is of faecal origin, is almost invariably accom-
panied by B. coli, and while the sanitary significance
is the same it may if desired be distinguished from B.
coli by a microscopical examination of the bile solution
when long strings of much larger bacilli than B. coli
are seen."

So far as we can judge from the report this appears
to constitute the preferred procedure of the Committee.
In any case the matter was passed upon by the Labora-

THE COLON GROUP OF CACILLI 139

tory Section of the American Public Health Associa-
tion at its Washington meeting in September, 1912,
by the adoption of a resolution recommending deter-
minations of numbers at 20° and 37° and the lactose
bile presumptive test as the standard routine procedure
in water examinations. The detailed study of particular
types of the colon group may, of course, be important,
in special cases; but the lactose bile test is sufficient
for general sanitary purposes.

CHAPTER VII

SIGNIFICANCE OF THE PRESENCE OF THE COLON GROUP
IN WATER

Colon Bacilli in the Intestines of the Lower Animals.

The Bacillus coli i » by no means confined to the human
intestine. Dyar and Keith (Dyar and Keith, 1893) found
it to be the prevailing intestinal form in the cat, dog, hog,
and cow. About the same time, Fremlin (Fremlin, 1893)
found colon bacilli in the faeces of dogs, mice, and rab-
bits, but not in those of rats, guinea pigs, and pigeons.
Smith (Smith, 1895) recorded the presence of the
organism, in almost pure cultures, in the intestines of
dogs, cats, swine, and cattle; and he also found it in
the organs of fowls and turkeys after death. Brotzu
(Brotzu, 1895) reported B. coli and allied forms as very
abundant in the intestine of the dog; and Belitzer
(Belitzer, 1899) isolated typical colon bacilli from the
intestinal contents of horses, cattle, swine, and goats.
Moore and Wright (Moore and Wright, 1900) recorded
the finding of the colon bacillus in the horse, cow,
dog, sheep, and hen; and in a later report (Moore
and Wright, 1902) they noted its occurrence in swine and
in some, but not all, the specimens of rabbits examined.
In frogs it was not found. Eyre (1904) has more
recently isolated typical B. coli from the intestines

140

SIGNIFICANCE OF COLON GROUP IN WATER 141

of mice, rats, guinea pigs, rabbits, cats, dogs, sheep,
goats, horses, cows, hens, ducks, pigeons, sparrows,
divers, gulls, and fish of various sorts. Houston (1904)
found B. coli abundant in the faeces of gulls, as might be
expected from their feeding habits. Houston (1905)
and other recent observers have found it impossible,
even by the use of elaborate series of fermentation tests,
to distinguish human B. coli from those found in animals.
Savage (1906) compared colon-like organisms isolated
from the intestines of swine, cattle, horses, and sheep
with those of human origin in respect to their action
upon lactose, dulcite, mannite, raffinose, glycerine,
maltose, galactose, lamilose, saccharose, starch and
cellulose; but he failed to find any general correlations
between habitat and biochemical powers.

Ferreira, Horta and Paredes (i9o8b) have made
perhaps the most elaborate study of the distribution
of colon bacilli in the lower animals. They isolated
8 1 lactose-fermenting bacilli from 38 species of mammals
and 8 species of birds, including monkeys, bears, wolves,
foxes, hyenas, lions, panthers, tapirs, a camel, deer,
and ostriches from the Zoological Gardens. These
cultures were studied by an elaborate series of tests
and 93 per cent of them proved to be typical B. coli.
Bettencourt and Borges (i9o8b) working in the same
laboratory showed that there were no specific differences
in agglutination with immune sera and in complement
fixation between the colon bacilli of human and of
animal origin. Konrich (1910) reports the examina-
tion of 170 samples of faeces from men, horses, swine,
sheep, cows, goats, dogs, cats, guinea pigs, mice, rabbits,

142 ELEMENTS OF WATER BACTERIOLOGY

rats, earthworms, moles, fowls, swallows, sparrows,
ducks, pigeons, geese, a jackdaw, a redstart, a blackbird,
an adder, and a trout. Three out of 5 guinea pig
samples, 4 out of 20 horse samples, 2 out of 3 mouse
samples, 3 out of 8 rabbit samples, and 2 out of 8
earthworm samples, 14 in all, were negative; while
all the rest showed B. coli.

In cold-blooded animals the occurrence of B. coli is
less constant. Negative results in the frog and positive
results in certain fishes, an adder and earthworms have
just been quoted. Amyot (1902) failed to find the
organism in the intestines of 23 fish representing 14
species. Johnson, on the other hand (Johnson, 1904),
in the examination of the stomach and intestines of
67 fish caught in the polluted Illinois and Mississippi
Rivers, isolated B. coli 47 times. He concluded from
these results that the migration of fish from a con-
taminated stream or lake to an unpolluted one may
explain the occasional finding of B. coli in small samples,
or the more regular detection of it in large volumes
of the water.

Bettencourt and Borges (i9o8b) isolated 29 cultures
of colon-like microbes from the intestines of 17 types
of fishes, reptiles and amphibia. Only 8 of the 29
formed gas in lactose broth and only 2 (from an eel
and an adder) proved to be typical B. coli. It should
be noted, however, that the samples of faecal material
were plated directly on Endo medium instead of being
subjected to the more sensitive process of preliminary
enrichment.

Fromme (1910) reviews the work of many observers

SIGNIFICANCE OF COLON GROUP IN WATER 143

in regard to the presence of colon bacilli in the intes-
tines of cold-blooded animals (particularly fish of
various sorts and oysters) and concludes that while
they are regularly found in warm-blooded animals
they are found often, but not regularly, in cold-blooded
animals. The lower the zoological type the rarer are
the colon bacilli.

Alleged Ubiquity of the Colon Bacillus. Many bacte-
riologists have gone further and affirmed that the
colon bacillus was not a form characteristic of the
intestine at all, but a saprophyte having a wide dis-
tribution in nature. The first of this school, perhaps,
was Kruse (Kruse, 1894), who in 1894 protested against
the arbitrary conclusions drawn from the colon test
as then applied. He pointed out that the characters
usually observed marked, not a single species, but a
large group of organisms. As ordinarily defined, he
added, " the Bacterium coli is in no way characteristic
of the faeces of men or animals. Such bacteria occur
everywhere, in air, in earth, and in the water, from the
most different sources." Even if the relations to milk
and sugar media be considered, " micro-organisms
with these characteristics are also widespread." Dr.
Kruse gave no experimental data on which his opinion
was based. In the same year Beckmann (Beckmann,
1894) isolated a bacillus which he identified by pretty
thorough tests as B. coli from the city water of S trass-
burg, a ground-water which he believed could by no
possibility be subject to faecal contamination. Large
quantities of water were used for the isolation.

Refik (Refik, 1896) recorded the constant presence

144 ELEMENTS OF WATER BACTERIOLOGY

of colon bacilli in water of all sorts, public supplies,
wells, cisterns, and springs in the neighborhood of
Constantinople, and Poujol in the succeeding year
reported (Poujol, 1897) the isolation of B. coli from
22 out of 34 waters studied by him in relation to their
use as public supplies. The waters were from various
sources — springs, wells, and rivers — but all were of fair
quality and many quite free from any possibility of
contamination. Samples of 100 c.c. were used for
analysis.

Certain Italian observers appear to have come to
even less conservative conclusions. Abba (Abba, 1895)
found colon bacilli constantly present in unpolluted
waters near Turin. Moroni (Moroni, 1898; Moroni,
1899) reported the examination of numerous deep and
shallow wells and unpolluted springs about Parma,
as well as of the public water-supply of the city, and
concluded that the colon bacillus was a water form and
had no sanitary significance. The characters used
for the identification of the species in this case were
fairly exhaustive, but both Abba and Moroni used
liter samples for analysis.

Levy and Bruns (Levy and Bruns, 1899) gave a new
turn to the discussion by emphasizing the importance of
animal inoculation, already suggested by Blachstein
(Blachstein, 1893) and others. They claimed that the
existence of numerous para-colon and para-typhoid
organisms in air, in dust, and in unpolluted water
made it impossible to decide by ordinary bacteriological
methods whether true colon bacilli were present in
water or not. In no case, however, did representatives

SIGNIFICANCE OF COLON GROUP IN WATER 145

of the colon group isolated by them from water kill
a guinea pig, even when i or 2 c.c. were injected intra-
peritonally. The authors, therefore, considered patho-
genicity as an attribute belonging only to the true
B. coli of the intestine. This paper aroused Professor
Kruse's pupil, Weissenfeld, to a publication, in which
the position of the Bonn school was carried to an
extreme. Weissenfeld reported (Weissenfeld, 1900) the
analysis of 30 samples of water supposedly pure,
and of 26 samples considered to be contaminated. In
each case a single centimeter sample was first incubated
in Parietti broth, and if no growth occurred, larger
samples of half a liter or a liter were examined. Colon
bacilli were found in all the samples; and the patho-
genicity varied independently of the source of the water.
The author concluded that " the so-called Bacterium
coli may be found in waters from any source, good or
bad, if only a sufficiently large quantity of the water
be taken for analysis."

With regard to the question of pathogenicity as a
diagnostic test for intestinal B. coli, there is little doubt
of the correctness of Weissenfeld's conclusions. This
property is so variable as to have no important value.
Colon bacilli freshly isolated from the intestine are
frequently non- virulent, and Savage (1903*) and others
have shown that there is in general no correlation
between pathogenic power and direct or indirect intes-
tinal origin. On the other hand Weissenfeld's work
entirely fails to show that the colon bacillus, pathogenic
or non-pathogenic, is a normal inhabitant of unpolluted
waters. Even his own results, if the quantitative rela-

146 ELEMENTS OF WATER BACTERIOLOGY

tions be considered, furnish evidence to the contrary.
In 24 of the 26 samples from bad sources, he isolated
his imperfectly defined colon bacilli from i c.c. of the
water, while in only 8 of the 30 samples of good waters
could he find such organisms in that quantity.

Colon Bacilli on Plants and Plant Products. The
work of certain recent observers has suggested the
possibility that the colon bacillus may live in a semi-
parasitic fashion on plants as well as on animals. Of
a series of 47 cultures of lactic-acid bacteria, recently
examined by one of ourselves (Prescott, i902a; Prescott,
1903, Prescott, 1906), 25 were found to give the reactions
of B. coli. These organisms were isolated chiefly from
cereals and products of milling, such as flour, bran,
cornmeal, oats, barley, etc., while others were in technical
use for producing the lactic fermentation. There is
no evidence that any of these organisms were of intesti-
nal origin, and yet they possess all the characters of typi-
cal colon bacilli, even to the pathogenic action when
inoculated into guinea pigs. In Germany, Papasotiriu
(Papasotiriu, 1901) was meanwhile carrying on almost
exactly similar investigations to Prescott's, with identi-
cal results.

Other testimony is somewhat conflicting with regard
to the occurrence of B. coli on plants. Klein and
Houston (1900) reported the finding of typical colon
bacilli in only 3 out of 24 samples of wheat and oats
obtained from a wholesale house; rice, flour, and oat-
meal bought at two different retail shops gave B. coli
in all three cereals in one case and on none in the other.
Clark and Gage (1903) were unable to isolate B. coli

SIGNIFICANCE OF COLON GEOUP IN WATER 147

from standing grains. Gordan (1904) could not find
B. coli in .1 and .01 mg. samples of clean bran, but
isolated it easily from that of poor quality. Winslow
and Walker (1907) have recently reported the examina-
tion of 178 samples of grain and 40 samples of grasses
for B. coli without success. On the other hand, Diiggeli
(1904) found B. coli among the bacteria occurring on
the leaves of growing plants, although it was not one
of the most abundant species. Barthel, too (Barthel,
1906), found B. coli widely distributed on plants from
both cultivated and uncultivated regions. Bettencourt
and Borges (1908^ examined 35 samples of vegetables
and cereals purchased in open market and found 12
lactose-fermenting forms, of which only 6 proved to be
B. coli. It should be noted, however, that the method
of isolation used was direct plating on Endo-medium,
which is of course less sensitive than the enrichment
processes used by other workers.

Neumann (1910) has recently studied the distribu-
tion of colon bacilli on and in various food substances
such as bread, milk, butter and fruit. From fresh
fruits immediately after picking he never isolated
them, but they were present in a certain proportion of
all the foods which had been exposed to human con-
tamination and the author concludes that wherever
human hands have been, there will B. coli be found.
Konrich (1910) in a similar series of investigations
obtained positive results from 46 out of 100 .1 to .5 gm.
samples of cultivated plants while leaves of trees and
grasses and herbs on waste places gave about 6 per
cent positive results. Hay showed colon bacilli in

148 ELEMENTS OF WATER BACTERIOLOGY

91 per cent of the 135 samples examined and grains
in 55 per cent of 300 samples.

Colon Bacilli in Dust and Soil. Winslow and Kligler
(1912) have shown that colon bacilli may be very
abundant in the dust of city streets and houses, as
might naturally be expected from the fact that such
dust is largely made up of horse droppings. They
examined 24 samples of street dust and 72 samples of
house dust (all in New York City). All of the street
dusts and 63 of the 72 house dusts contained colon
bacilli in at least one of three duplicate i<U gram
portions. In two street samples the numbers rose to
330,000 and 660,000 per gram respectively, while the
largest indoor result was 60,000. The average for the
indoor dusts was between 1000 and 2000 per gram and
for the street dusts over 50,000 per gram. This dust
was dust deposited on surfaces and would only be
carried up into the air by currents of some force. It
is well known that colon bacilli are, as a matter of fact,
rarely present in street or house air. Konrich (1910)
exposed open Petri dishes of dextrose broth to the air
of Jena streets for 24-hour periods, daily, for 3 months
and found colon bacilli only n times. The colon
bacilli in street dust may, however, perhaps account
for the anomalous positive results sometimes obtained
in reservoirs bordered by roadways.

Konrich (1910) has also made important contribu-
tions to the study of colon bacilli in the earth. Out
of 547 samples of soil, 65% showed B. coli in por-
tions of between .1 and .5 gm. The farther removed
from cultivation a sample was, the less were the chances

SIGNIFICANCE OF COLON GROUP IN WATER 149

of positive results. He concludes that B. coli is widely
distributed in the outer world. It is almost always
found in soil from cultivated fields or from traveled
places. The farther a source is removed from travel
and from cultivation the more rarely is the colon
bacillus -found; but it is never altogether absent. On
plants or parts of plants it is frequently found when they
come from cultivated land; on plants from waste places
it is rarely found. It seems probable that colon bacilli
may be even more widely distributed in a thickly
settled and intensively cultivated country like Ger-
many than in the United States.

The Number of Colon Bacilli, not Their Mere Pres-
ence, as an Index of Water Pollution. The more
important practical conclusions to be drawn from these
various investigations seem to be as follows:

1. Bacteria corresponding in every way to B. coli are
by no means confined to animal intestines, but are
widely distributed elsewhere in nature.

2. The finding of a few colon bacilli in large samples
of water, or its occasional discovery in small samples,
does not necessarily have any special significance.

3. The detection of B. coli in a large proportion of
small samples (i c.c. or less) examined is imperatively
required as an indication of recent sewage pollution.

4. The number of colon bacilli in water rather than
their presence should be used as a criterion of recent
sewage pollution.

With these qualifications the value of the colon test
was never more firmly established than it is to-day.
Whether or not originally a domesticated form, it is

150 ELEMENTS OF WATER BACTERIOLOGY

clear that the colon bacillus finds in the intestine of the
higher vertebrates an environment better suited to its
growth and multiplication than any other which occurs
in nature. Houston (igof) records the number of B.
coli per gram of normal human faeces as between
100,000,000 and 1,000,000,000. It is almost certain
that the only way in which large numbers of these
organisms gain access to natural waters is by pollution
with the domestic, industrial, and agricultural wastes
of human life. If pollution has been recent, colon
bacilli will be found in comparative abundance. If
pollution has been remote the number of colon bacilli
will be small, since there is good evidence that the
majority of intestinal bacteria die out in water. If
derived from cereals or the intestines of wild animals, the
number will be insignificant except perhaps in the vicinity
of great grain-fields or where the water receives refuse
from grist-mills, tanneries, dairies, or lactic-acid factories.
The first recognition of the necessity for a quantita-
tive estimation of colon bacilli in water we owe to Dr.
Smith, who in 1892 (Smith, 1893^ outlined a plan for a
study to be made by the New York Board of Health on
the Mohawk and Hudson Rivers. Burri (Burri, 1895)
pointed out that the use of so large a sample as a liter
for examination would lead to the condemnation of
many good waters. Freudenreich (Freudenreich, 1895)
at the same time indicated the necessity for taking into
account the number of colon bacilli present. He recorded
the isolation of the organisms from unpolluted wells,
when as large a quantity of water as 100 c.c. was used,
and concluded that it was entirely absent only from

SIGNIFICANCE OF COLON GROUP IN WATER 151

waters of great purity and present in large numbers
only in cases of high pollution. This author also
quoted Miquel as having found colon bacilli in almost
every sample of drinking-water if only a sufficient por-
tion were taken for analysis.

The practical results of the application of the colon
test from this standpoint have proved of the highest
value. As originally outlined by Dr. Smith, it con-
sisted in the inoculation of a series of dextrose tubes
with small portions of water, tenths or hundredths of
the cubic centimeter. It was first used by Brown
(Brown, 1893) in 1892 for the New York State Board
of Health, and it showed from 22 to 92 faecal bacteria
per c.c. in the water of the Hudson River at the Albany
intake, and from 3 to 49 at various points in the Mohawk
River between Amsterdam and Schenectady. In some
previous work at St. Louis, the colon bacilli in the
Mississippi River were found to vary from 3 to 7 per c.c.

Hammerl (Hammerl, 1897) used the presence of
Bacillus coli as a criterion of self-purification in the
river Mur. He considered, in spite of the position taken
by Kruse, that when a water contained large numbers
of colon bacilli, as well as an excess of bacteria in
general, it might be considered to be contaminated by
human or animal excrement. As, however, the organism
would naturally be present in large quantities of such
a water as that of the Mur, he used no enrichment
process, but made plate cultures direct; he defined the
B. coli as a small bacillus, non-motile or but feebly
motile, growing rapidly at 37° C., coagulating milk
and forming gas in sugar media. In general, Hammerl

152 ELEMENTS OF WATEE BACTEEIOLOGY

failed to find colon bacilli in the river by this method,
except immediately below the various towns situated
upon it; at these points of pollution he discovered a
few colon colonies upon his plates, not more than 4 to
6 per c.c. of the water. He concluded that " the
Bacterium coli, even when it is added to a stream in
great numbers, under certain circumstances disappears
very rapidly, so that it can no longer be detected in
the examination of small portions of the water."

The most important work upon the distribution of B.
coli has been that carried out in England by the bacteri-
ologists of the local government board, by Dr. Houston
in particular. This investigator (Houston, 1898; Hous-
ton, 1899*; Houston, igooa) made an elaborate series
of examinations of soils from various sources to see
whether the microbes considered to be characteristic
of sewage could gain access to water from surface wash-
ings free from human contamination. In the three
papers published on this subject the examination of 46
soils was recorded. In only 10 of the samples was B. coli
found, and of these 10, 9 were obviously polluted, being
derived from sewage fields, freshly manured land, or
the mud-banks of sewage-polluted rivers. The author
finally concluded that " as a matter of actual observa-
tion the relative abundance of B. coli in pure and impure
substances is so amazingly different as to lead us to
suspect that not only does B. coli not flourish in nature
under ordinary conditions, but that it tends to even
lose its vitality and die." " In brief, I am strongly
of opinion that the presence of B. coli in any number,
whether in soil or in water, implies recent pollution of

SIGNIFICANCE OF COLON GROUP IN WATER 153

animal sort." Pakes (Pakes, 1900) stated on the
strength of an examination of " about 300 different
samples of water," no particulars being published, that
water from a deep well should not contain B. coli at
all, but that water from other sources need not be con-
demned unless the organism was found in 20 c.c. or less.
When colon bacilli were found only in greater quantities
than 100 c.c. the water might be considered as probably
safe. Horrocks (Horrocks, 1910), after a general
review of English practice, concluded that " when a
water-supply has been recently polluted with sewage,
even in a dilution of one in one hundred thousand, it
is quite easy to isolate the B. coli from i c.c. of the
water." " I would say that a water which contained
B. coli so sparingly that 200 c.c. required to be tested
in order to find it had probably been polluted with
sewage, but the contamination was not of recent date."
Chick (Chick, 1900) found 6100 colon bacilli per c.c.
in the Manchester ship canal, 55-190 in the polluted
River Severn, and numbers up to 65,000 per gram
in roadside mud. On the other hand, of 38 unpolluted
streams and rivulets, 31 gave no Bacillus coli and the
other 7 gave i per c.c. or less. The Liverpool tap water,
snow, rain, and hail showed no colon bacilli.

One of the first elaborate applications of the colon
test was made by Jordan in the examination of the fate
of the Chicago sewage in the Desplaines and Illinois
Rivers. In these studies of self-purification (Jordan,
1901) the analyses were made quantitative by the
examination of numerous measured samples, fractions
of the cubic centimeter; and the method employed

154 ELEMENTS OF WATER BACTERIOLOGY

was enrichment, either in dextrose-broth fermentation
tubes or in phenol broth, with subsequent plating
on litmus lactose agar. The cultures isolated were
tested as to their behavior in dextrose broth, pep-
tone solution, milk, and gelatin; of the dextrose
tubes made directly from the water all were con-
sidered positive which gave more than 20 per cent
gas in the closed arm, with an appreciable excess of
hydrogen. The results were very significant. In fresh
sewage a positive result was obtained about one-third of
the time in one one-hundred-thousandth of a cubic cen-
timeter and almost constantly in one-ten-thousandth
of a cubic centimeter. The Illinois and Michigan
canal proved almost as bad, giving positive results on 7
days out of 28 in dilutions of one in a hundred thousand
and on 28 days out of 32 in a dilution of one in ten
thousand. At Morris, 27 miles below Lockport, where
the canal enters the bed of the Desplaines River, and
9 miles below the entrance of the Kankakee, the
principal diluting factor, the numbers were so reduced
that positive results were obtained only on n days out
of twenty in one-thousandth of a cubic centimeter, on
20 days out of thirty in one-hundredth of a cubic centi-
meter, and on 20 days out of 23 in one-tenth of a cubic
centimeter. At Averyville, 159 miles below Chicago,
colon bacilli were isolated on only 4 days out of 27 in
one- tenth of a cubic centimeter, and on 13 days out of
31 in one cubic centimeter. A comparison with certain
neighboring rivers showed this to be about the normal
value for waters of similar character, as the following
table extracted from Professor Jordan's paper will show :

SIGNIFICANCE OF COLON GEOUP IN WATER 155

NUMBER OF B. COLI PRESENT IN CERTAIN RIVER
WATERS

(JORDAN, 1901)

O.I

C.C.

I C

.c.

Source of Sample.

No. Days
Water
Examined.

No. Days
B. Coli
Found.

No. Days
Water
Examined.

No. Days
B. Coli
Found.

Illinois River, Averyville . . .
Mississippi River, Grafton . .
Fox River

27
34

4
10
2

31
35

13
23
6

Sangamon River

Missouri River

These results harmonize rather closely with those
previously recorded by Brown and Fuller and indicate
that in the larger rivers where the proportionate pollu-
tion is not extreme, colon bacilli may be isolated in about
half the i-c.c. samples examined. Such rivers are of
course inadmissible as sources of water-supply, accord-
ing to modern sanitary standards, unless subjected to
purification of some sort.

Hunnewell and one of us (Winslow and Hunnewell,
igo2b) examined a considerable series of normal waters
for B. coli, testing i c.c. from each by the dextrose-
broth method and a larger portion of 100 c.c. by incuba-
tion with phenol broth as described in Chapter VI.
The samples were obtained from the public supplies of
Taunton, Boston, Cambridge, Braintree, Brookline,
Needham, and Lynn in Massachusetts, and Newport,
R. I., from the Sudbury River, from the ocean, from
the waters of springs bottled for the market, from
ponds, pools of rain and melted snow, springs, brooks,
shallow wells, and driven wells in various towns near

156 ELEMENTS OF WATEE BACTERIOLOGY

the city of Boston. For comparison 50 samples of
polluted waters from the Charles, Mystic, Neponset,
and North Rivers were examined. The colon bacillus
was defined by gas production in dextrose broth,
coagulation of milk, reduction of nitrates, formation
of indol and failure to liquefy gelatin; and organisms
which lacked the power to reduce nitrates or to form
indol were classed in the " Paracolon group." The
results are summarized in the following table:

PRESENCE OF B. COLI IN POLLUTED AND UNPOLLUTED
WATERS

(WlNSLOW AND HUNNEWELL, IQO2b)

Unpolluted Waters

Samples examined 157

Dextrose broth positive 40

Lactose plate positive 13

Colon group 5

Paracolon group 5

B. cloacae group

Streptococcus group 3

Polluted Waters

I C.C.

Samples examined 50

Dextrose broth positive 50

Lactose plate positive 50

Colon group 18

Paracolon group 6

B. cloacae i

Streptococcus group 25

3i
ii

37
26

As the authors pointed out, these tables indicate that
bacteria capable of growth at the body temperature and
fermenting dextrose and lactose are infrequently found

SIGNIFICANCE OF COLON GROUP IN WATER 157

in unpolluted waters, and colon bacilli are very rarely
present. In 157 samples, typical colon bacilli were
found only 5 times out of 157, in i c.c. Lactose ferment-
ing organisms appeared in only 8 per cent of the nor-
mal samples and in 100 per cent of the polluted ones,
in i c.c. Incidentally it may be pointed out that these
tables well illustrated the dangers of overgrowths,
particularly in large samples. It is clear that the strep-
tococci had killed out colon bacilli, originally present,
in a large proportion of the loo-c.c. samples of polluted
waters and in some of the i-c.c. samples, since, in so
many cases, gas formation was followed by the isola-
tion of the streptococcus alone.

Colon Bacilli in Surface Waters. Clark and Gage
(1903) have published the results of certain studies of

DISTRIBUTION OF TOTAL BACTERIA
SURFACE-WATERS

(CLARK AND GAGE. 1903)

AND B. COLI IN

Lake.

Population of
Watershed per
Square Mile.

Bacteria
per c.c.

B. Coli
Per Cent Positive Tests.

I C.C.

100 C.C.

1400

612

13-3

33-o

356

319

3-5

17.2

116

103

0.0

o.o

9°

170

o.o

14.0

o.o

9.0

2-3

4-5

4-6

21 .0

133

0.0

9.0

J3i

0.0

6.7

10*

0.0

6.2

0.0

7-7

107

0.0

9-3

* Shores used for pleasure resorts.

158 ELEMENTS OF WATER BACTERIOLOGY

Massachusetts ponds which indicate clearly the coin-
cidence of the distribution of B. coli in single centimeters
of surface waters, with actual sanitary conditions.
They show also the slight significance of the test for
this organism in larger volumes of water. Almost
every source gave positive tests in 100 c.c., while with
i-c.c. samples only those lakes appear suspicious which
are, in fact, exposed to dangerous pollution.

Houston (1905) gives the following table, which may
be taken as another fair example of the distribution of B.
coli in small streams and lakes. Of the two lakes
studied, Loch Ericht is free from the pollution of human
or domesticated animals, while Loch Laggan receives
some drainage from farm lands; both are of large
size. The brook and river samples were collected from
adjacent streams.

DISTRIBUTION OF B. COLI IN SURFACE-WATERS

(HOUSTON, 1905)
Percentage of Samples showing B. Coli in each Dilution.

Dilution.

+ O.I C.C.

+ 1.0 C.C.
— O.I C.C.

+ 10 C.C.
— I C.C.

+ IOO C.C.
— IO C.C.

Not in

IOO C.C.

Brooks and river.
Loch Laggan.

7-7

53.8

I . 2

34-6

33 -O

3.8 •

49 -4

16 .4

Loch Ericht

I .0

19 .O

80.0

As an example of a heavily polluted stream, 011 the
other hand, the table on page 159 may be cited. It shows
in a striking way the increase of B. coli in the Thames
on its passage through London and its progressive
purification below.

SIGNIFICANCE OF COLON GROUP IN WATER 159

The river at the lower stations in this table was con-
siderably diluted with sea-water, yet it showed clearly
its large proportion of sewage. Normal sea- water,
even in the neighborhood of the shore, shows, B. coli
only in large samples. Houston (1904), in another
communication, reports the examination of 168 samples
of sea- water near the English coast. None of the samples
showed B. coli in i c.c.; 97 samples gave negative
results in 10 c.c.; 45 in 100 c.c., and 4 had no B. coli
even in 1000 c.c.

B. COLI IN THE RIVER THAMES AT VARIOUS POINTS

(HOUSTON, i904a)
Percentage of Positive Results

Place.

— 10

c.c.

+ 10
— I C.C.

+ i

— O.I

c.c.

+ .1

— O.OI

c.c.

+ .01
— 0.001

c.c.

+ 0.001
— 0.0001

c.c.

+0.0001
— O.OOOOI

c.c.

Sunbury

70 6

23 <

t; Q

Hampton.

ii. 8

64 7

177

c 0

Barking

4.2

45-8

4.2

Crossness

ii i

27 7

CQ O

II I

Purfleet

3.0

9.1

33-3

39-i

15-1

Grays

2.8

22 . 2

41 . 7

77 . 7

Mucking

30 8

C7 7

ii S

Chapman

C O

4< o

^O.O

Barrow Deep . .

12. O

36.0

40.0

12.0

Gartner (1910) has collected some interesting data
in regard to the ratio between the number of colon
bacilli and of total bacteria in waters of different
quality. The results from four different sets of experi-
ments by Konrich at Jena, Houston at London, Noble
in New York, and Hill at Giessen, may be combined
as on the opposite page:

160 ELEMENTS OF WATER BACTERIOLOGY

RATIO OF TOTAL BACTERIA TO COLON BACILLI IN
WATERS OF DIFFERENT CLASSES

Coli titer—
Smallest Por-
tion of Water
Showing
B. Coli.

Ratio of Plate Count to B. Coli.

Jena.

20°

London.

New York.

Giessen.

20°.

37°.

24°.

37°.

IOO C C

60,500
15,100
1,288

352
46

1 1, 800

1,810
255
34
2.4

1950
213
2O

280
47
7

4-9
0.4

10 C.C
I C.C
O.I C.C
O.OI C.C
O OOI C C

695
183
29.7
4.1

O.OOOI C.C.

O.OOOOI C.C.. .

The rather regular decrease in the ratio of the total
count to the B. coli count with an increase in the actual
number of colon bacilli is very interesting.

Prof. Gartner apparently holds that this fall in the
ratio of the plate count to the u coli titer " indicates a
fallacy in the method of the latter and in particular he
emphasizes the absurdity of the lowest figures in the table
which indicates that there were twice as many colon
bacilli as bacteria of all sorts. It seems to us that the
last phenomenon is quite as likely to be due to an error in
the plate count as to a failure in the enrichment pro-
cedure. Unless dilutions are very carefully made
plates inoculated with waters containing tens and
hundreds of thousands of bacteria per c.c. are pretty
likely to be so crowded that only a portion of the bac-
teria with which they are sown are able to develop.
As to the diminishing ratio with increasing coli-content,
it is exactly what might reasonably be expected. One-
tenth to one-quarter of the bacteria in sewage may be

SIGNIFICANCE OF COLON GROUP IN WATER 161

colon bacilli, and the greater the amount of sewage
in water, the nearer this ratio will be approached.

Colon Bacilli in Ground- waters. With ground-waters
the story is the same. Even in sources of excellent
quality we should expect to find, and we do sometimes
find, colon bacilli in large volumes of water. Abba,
Orlandi, and Rondelli (1899) showed by experiments
with B. prodigiosus at Turin that when bacteria are
present in great numbers on the surface of the ground,
a few may penetrate for a considerable distance and
ultimately reach the sources of ground-waters. The
chance that disease germs could survive this process in a
soil so impervious as to allow colon bacilli to appear
only in large samples of water, is infinitesimal.

An interesting contribution to the bacteriology of
ground-waters was made by the Massachusetts State
Board of Health (Massachusetts State Board of Health,
1901) in connection with the examination of the spring-
waters bottled for the sale in the State. Ninety-nine
springs were included in this study, and in almost every
instance 4 samples were examined, 2 taken directly
from the spring by the engineers of the board and 2
from the bottles as delivered for sale to the public. In
the water of one spring B. coli was found twice, once
in a sample from the spring and once in the bottled
sample. This spring was situated in woodland, but
was unprotected from surface drainage, and the method
of filling bottles subjected it to possible contamination.
In 5 other cases B. coli was found once in the sample
from the spring; all were subject to pollution from
dwellings or cultivated fields, and 4 of the 5 were shown

162 ELEMENTS OF WATER BACTERIOLOGY

to be highly contaminated, chemically. In 7 other
cases B. coli was found in the bottled samples alone;
3 of these sources were of high purity, but the bottling
process furnished opportunity for contamination.

Clark and Gage (1903), in the examination of 170
samples of water from tubular and curb wells of good
quality used as sources of water-supply, found B. coli
only 5 times, once in i c.c. and 4 times in 100 c.c. Horton
(1903), from a study of ground-waters in Ohio, concluded
that the presence of B. coli in wells and springs was
indicative of serious pollution.

Houston (i903b) makes an instructive comparison of
some more or less polluted shallow wells at Chichester
with deep ground-waters of high quality at Tunbridge
Wells. The following table shows the value of the
i cubic-centimeter sample in discriminating between
good and bad waters.

DISTRIBUTION OF B. COLI IN GOOD AND BAD WELL
WATERS

(HOUSTON, 1903^
Percentage of Positive Tests

Quantity of Water.

Chichester Shallow
Wells.

Tunbridge Wells,
Deep Wells.

IOO C.C.
IO C.C.

90
80

25
6

I C.C.

O. I C.C.

In a subsequent investigation, Houston (1905) exam-
ined still larger samples of water from the Tunbridge
Wells for B. coli: 49 samples of 100 c.c. each showed no
B. coli, and 27 liter samples showed B. coli only once.

SIGNIFICANCE OF COLON GROUP IN WATER 163

Kaiser (1905) reports an interesting correlation between
total numbers and B. coli in a series of 38 well waters.
Of 1 1 wells containing over 200 bacteria per c.c. 90 per cent
showed colon-like organisms in liter samples. Of 1 2 wells
containing from 50 to 200 bacteria per c.c. 67 per cent
gave colon-like organisms; of 26 wells with le s than 50
bacteria per c.c., only 27 per cent showed positive results.
Fromme (1910) brings out the relation between B.
coli and total numbers in 120 samples of well waters
near Hamburg in the table below.

RELATION BETWEEN TOTAL NUMBERS OF BACTERIA
AND B. COLI

(FROMME, 1910)

Colony Count.

Number of Samples.

Per cent Positive
B. coli Tests in 10 c.c.

Over 200
50-200
Under 50

35
19
66

40.0
15-8

Similar data obtained by one of us for some American
sources have been cited in Chapter I. Even Konrich
(1910), who is exceedingly rceptical as to the value of
the colon test, has shown that an increase in the colon
content of the Jena water supply (a ground-water)
always followed a heavy rain which washed through
some of the colon bacilli in the soil.

Colon Bacilli in Filtered Waters. One of the most
important applications of the colon test is in the control
of the operation of municipal water niters. It has
been used for this purpose for 10 years or more at
Lawrence, and Fuller laid stress upon its results in his
classic experiments on water purification in the Ohio
valley. At Cincinnati he records the presence of

164 ELEMENTS OF WATER BACTERIOLOGY

colon bacilli in 60 per cent of the i-c.c. samples from the
Ohio River, while the effluent from either slow sand or
mechanical niters gave positive results only half the
time in samples of 50 c.c. The results of the examina-
tions carried out at Lawrence for 6 years are brought
together in the table below from the Annual Reports
of the Massachusetts State Board of Health.

B. COLI IN MERRIMAC RIVER AND LAWRENCE FILTER
EFFLUENT

Merrimac River, Per
cent of i c.c.,
Samples containing
B. coli.

Merrimac River,
Number B. coli per
c.c.

Filtered Water, Per
cent of i c.c.,
Sample containing
B. coli.

1900

99-7

iS.I

1901

1902

99.0

4.0

1903

99-0

4.2

1904

100. 0

8.0

1905

IOO.O

118

4-7

* Not given.

At Harrisburg, Pa., mechanical nitration combined
with chlorin disinfection has yielded the results tabulated
below :

B. COLI IN RAW AND TREATED WATER AT
HARRISBURG, PA.

(HARRISBURG, 1913)

Per Cent Positive Tests in i c.c.

Raw Water.

Treated Water.

1906

71.9

2.7

1907

64.0

I .O

1908

65-7

I . I

1909

63-1

1.0

1910

55-4

O. 2

1911

77-3

0.6

1912

46.9

0.8

SIGNIFICANCE OF COLON GROUP IN WATER 165

At Washington the most complete slow sand-
filtration plant yet constructed has yielded the results
tabulated below, for which we are indebted to the
courtesy of Mr. F. F. Longley:

B. COLI IN POTOMAC RIVER AND WASHINGTON FILTER
EFFLUENT

Dalecarlia Reservoir Inlet.

Filtered- Water Reservoir
Outlet.

1906.

Samples.

Number

Tested. _^

Tested.

IO C.C.

I C.C.

10 C.C.

I. C.C.

February. . . .

iS 5

March

April

May

25 3

June

26 o

July

•y

August

September. . .

October

November. . .

December. . .

1907

January

February. . . .

March

April

May

June

2j-

July

II -

August

September. . .

It must be remembered that in the Washington plant
filtration is supplemented by thorough sedimentation,
preliminary and subsequent. The entire credit for the
good effluent obtained is not therefore due to the filters.
At Lawrence it has been shown that removal of colon

166 ELEMENTS OF WATER BACTERIOLOGY

bacilli in storage reservoirs and pipe systems may be
considerable. The figures obtained in 1900 at various
points in the distribution system may be cited as an
example.

PERCENTAGE OF SAMPLES OF WATER CONTAINING
B. COLL LAWRENCE, MASS.

(MASSACHUSETTS STATE BOARD OF HEALTH, 1901)

Effluent of
Filter.

Outlet of
Reservoir.

Tap
City Hall.

Tap Experi-
ment Station.

In i c.c

In 100 c.c. . . .

Fromme (1910) in the examination of a filtered
water averaging 35 bacteria per c.c. obtained the
following results in various quantities of water.

PERCENTAGE OF SAMPLES OF WATER CONTAINING
B. COLL

(FROMME, 1910)

Number of Samples.

Volume Examined.

Per Cent Positive.

101

2OO C.C.

30-9

412
800

10 C.C.
I C.C.

2.9
0.25

In regard to the proportion of positive colon tests
permissible in a filter effluent, Clark and Gage (Clark
and Gage, 1900) reported some specially instructive
observations made when certain of the underdrains of
the Lawrence filter were relaid in the autumn of 1898.
In doing this work the sand on some of the beds was
seriously disturbed; and in December, after the work was
completed, B. coli was found in i c.c. of the filtered

SIGNIFICANCE OF COLON GROUP IN WATER 167

effluent in 72 per cent of the samples examined. In
January and February the organisms were found in
54 per cent and 62 per cent of the samples, respectively,
while in March the number fell to a normal value of
8 per cent. Corresponding to this excess of B. coli
in the city water, there were 12 cases of typhoid fever
in December, 59 cases in January, 12 in February,
and 9 in March, all during the early part of the month.
The authors conclude that " when filtering a river-
water as polluted as that of the Merrimac, it is safe
to assume that when B. coli is found only infrequently
in i c.c. of the effluent, the typhoid germs, necessarily
fewer in number and more easily removed by the
filter, have been eliminated from the water."

The results of the daily tests carried out at municipal
filter plants are frequently expressed in monthly or
yearly averages, as in some of the cases quoted above.
It must be remembered, however, that averages of
this sort are accepted only by courtesy and with the
implied assumption that conditions are approximately
constant during the period averaged. When it is
said that an acceptable effluent may show B. coli in
3 or 4 per cent of the samples tested the statement is
true only for a series of samples collected and examined
at the same time. If in a given month 3 per cent of
the i c.c. samples tested show B. coli, the effluent may
or may not be safe. If on each of 20 days 3 B. coli
or thereabouts were present in 100 c.c. of the water
it is probably a safe one. If on 19 days no B. coli
were present, and on the twentieth day 100 c.c. showed
60 B. coli, the average result would be the same, but

168 ELEMENTS OF WATER BACTERIOLOGY

the water on one day was of a dangerous character.
With properly managed filter plants marked varia-
tions do not occur from day to day and average results
are generally reliable. It is wholly misleading, how-
ever, to compare such results with the average exami-
nations of an unfiltered surface water. With surface
waters daily variations are the rule and a low monthly
average of colon tests may include and cover up
dangerous and significant high numbers at particular
periods.

Summary of American and Foreign Opinion as to
the Value of the Colon Test. The general results of the
studies of the colon tests which have now been carried
out in great numbers all over the world may be sum-
marized by a few further citations.

In America the fact that the number of colon bacilli
in a water measures the degree of its pollution is now
universally accepted. The same conclusion has been
established in England by the elaborate investigations
of Houston and his pupils. Savage, for example,
concluded (Savage, 1902) from a study of a large num-
ber of water supplies in Wales that even in surface
waters, exposed to animal contamination from adjacent
grazing grounds, B. coli is not present in 2 c.c. unless
other pollution is present. In a more recent review
of the whole subject, the same author (Savage, 1906)
concludes that " there is no evidence or observations
which have ever shown that B. coli, reasonably defined,
is present in any numbers in sources which have not
been exposed to some form of faecal contamination."

In Germany, Petruschky and Pusch (Petruschky

SIGNIFICANCE OF COLON GROUP IN WATER 169

and Pusch, 1903) examined a considerable series of
waters from different sources by incubating measured
samples with equal amounts of nutrient broth and iso-
lating upon agar. In 45 samples of well-waters they
found B. coli 7 times in .01 c.c., 9 times in .1 c.c., and
7 times in i c.c. In the other 22 cases it could not be
found in i c.c. and in 4 cases not in 100 c.c. One sample
showed it only in 600 c.c. and i not in 750 c.c. Of
29 river-waters, only 2 failed to give positive results
in .1 c.c. and 14 showed B. coli in .001 of a c.c. or less.
In sewage the number varied from i to 1,000,000 per
c.c. The authors conclude that a quantitative estima-
tion of the B. coli content furnishes a good measure of
the faecal pollution of water. There is still a school of
bacteriologists in Germany, however, who are inclined
to place little value on the colon-test. We have pointed
out how Kruse and his pupils at Bonn led in the attack
on it in 1894. Fourteen years later Kruse (1908)
concluded after a full summary of the literature that
the colon test was on the whole less valuable than the
gelatin count, although he admitted that when the
test is made quantitative it is valuable as a supple-
ment to the plate count. Konrich (1910), working in
Gartner's laboratory at Jena, after perhaps the most
exhaustive study ever made of the whole subject,
concludes, that to include the colon test in forming
judgment on the sanitary quality of a water is to
complicate the procedure without improving it and that
one would do well to omit the test except in certain
special cases.

The German criticisms of the colon test are based

170 ELEMENTS OF WATER BACTERIOLOGY

mainly on two considerations, the inaccuracy of the
method itself and the difficulties in its interpretation.
They hold on the one hand that the errors in the enrich-
ment process and the consequent lack of correspondence
between duplicate determinations are so great that the
whole process is worthless. It is of course true that
chance errors of distribution and the overgrowth which
often occurs in the lower tubes of a dilution series do
occasionally lead to individual erroneous results. If
several dilutions are made with duplicates in each
dilution, and particularly if reliance is placed, as it
should be placed, not on single determinations, but
on the average of several tests, results are, however,
obtained which are in accord with each other and with
the results of practical epidemiological experience.

The other objection brought forward by many Ger-
man sanitarians is that the wide distribution of the colon
bacillus leads to its presence even in considerable num-
bers in waters which are really of good sanitary quality.
It is undoubtedly true that colon bacilli are often
found in surface waters which receive no sewage but
which are polluted only with the wash from roadways
or cultivated land. Even dust blowing in from a road-
way may perhaps contribute an appreciable pollution,
as we have pointed out above. Gartner (19 10) cal-
culates that in the soil of a cultivated field 100
meters square there are 15,000,000,000 colon bacilli
and points out that it is no wonder that a rain should
wash a few of them through into neighboring wells.
Konrich (1910) goes so far as to say that to reject on
principle all water containing B. coli in i c.c. portions

SIGNIFICANCE OF COLON GROUP IN WATER 171

is impossible, since many regions have no other water
available.

On the other hand, it may be urged that it is better
to be on the safe side if possible. There is no doubt
that in the United States there is no difficulty in secur-
ing water supplies, either from the ground or by storage
or filtration, which only rarely contain colon bacilli
in i c.c. samples. It may, perhaps, be that in the thickly
settled and intensively cultivated parts of Germany
this is not the case. Where it is possible to obtain such
waters we believe it to be wise to do so. Ground
waters to which the colon bacilli from cultivated soil
penetrate and surface waters which still contain many
colon bacilli from fields and roadways may be generally
inocuous but there is always the chance that infectious
material may find its way to the soil and may enter,
along with the intrinsically harmless colon bacilli of
manure. The two most recent German contributions
to the subject are on the whole favorable to the colon
test if wisely and intelligently applied. Fromme
(1910), from Dunbar's laboratory at Hamburg, concludes
from a survey of the literature and from his own studies
that " the finding of colon bacilli in water is a valuable
indicator of the quality of the water " and recommends
it particularly for ground-water spring waters and
filter effluents; and Prof. Gartner, the head of the Hygiene
laboratory at Jena, after a thorough discussion of
previous work, comes to a much more conservative
conclusion than his pupil, Konrich (Gartner, 1910).
While emphasizing the shortcomings of the process
and deploring the tendency "to set B. coli on a high

172 ELEMENTS OF WATER BACTERIOLOGY

throne and dance before it " he recognizes that the
isolation of this organism has its place. " My aim will
have been realized," he says, " if I have been able to
show that the colon test may be useful under certain
circumstances, but that it must be viewed with great
caution and that, moreover, not the mere numbers
of colon bacilli, but a careful consideration of the local
situation and all the circumstances bearing on the special
case are absolutely essential to the formation of an
opinion." With Prof. Gartner's emphasis on the
importance of a sanitary inspection all experienced
bacteriologists will certainly find themselves in agree-
ment.

Some of the best French bacteriologists are strong
supporters of the value of B. coli as an indicator of
pollution. Gautie (1905) holds that the quantitative
determination of B. coli is of the highest importance
in water analysis; and Vincent (1905), in an excellent
review of the subject, gives strong reasons for main-
taining the same position. He finds B. coli absent
from spring and well-waters of good quality and present
in polluted water in proportion to its pollution. He
concludes finally that water containing B. coli in .1 to
i.o c.c. is unfit to drink, while if the organism is found
in i.o to 10.0 c.c. it is of doubtful quality.

In Portugal too the trend of opinion is strongly in
favor of the colon test — the Anglo-American procedure,
as it is called in the publications of Dr. Bettencourt
and his associates (Ferreira, Horta and Paredes, i9o8a).

Altogether the evidence is quite conclusive that the
absence of B. coli demonstrates the harmlessness of a

SIGNIFICANCE OF COLON GROUP IN WATER 173

water as far as bacteriology can prove it. That when
present, its numbers form a reasonably close index of
the amount of pollution, the authors above quoted have
proved beyond reasonable cavil. It may safely be
said that when the colon bacillus is found in such abun-
dance as to be isolated in a large proportion of cases
from i c.c. of water, it is reasonable proof of the presence
of serious pollution.

CHAPTER VIII

VARIETIES OF COLON BACILLI AND THEIR SPECIAL
SIGNIFICANCE

Tests Used for Subdividing the Colon Group. The

group of colon bacilli as denned in Chapter VI includes
all aerobic Gram-negative bacilli which produce acid
and gas in dextrose and lactose media. By the applica-
tion of various tests, mainly bio-chemical, it may be
further split up, almost at will, into a great number of
varieties. The principal tests which have been used for
this purpose are as follows :

1. Motility.

2. Coagulation of milk.

3. Production of indol.

4. Liquefaction of gelatin.

5. Reduction of nitrates.

6. Reduction of neutral red.

7. Fermentation of various carbohydrates other than
dextrose and lactose.

8. Voges and Proskauer reaction.

9. Character of colonies on various solid media.

10. Aesculin reaction.

11. Growth and fermentation at 46°.

Motility is seldom determined in actual routine work
in this country from the general belief that its demon-

174

VAEIETIES OF COLON BACILLI 175

stration is both burdensome and needless. Motility.is a
fluctuating and uncertain property and one which
frequently requires repeated preliminary cultivations
to make manifest. Furthermore, non-motile colon bacilli
are common in the intestine and are probably as charac-
teristic of pollution as the motile forms.

McWeeney (1904) found non-motile B. coli abundant
in faeces and observed cases where the organisms were
motile at 20° and not at 37°. He quotes Stocklin as hav-
ing found 116 non-motile strains among 300 otherwise
normal B. coli from faeces. Evidence that non-motile
bacteria, otherwise resembling B. coli, occur in unpol-
luted water would furnish the only basis for requiring
this test as a routine procedure. No such evidence exists.
The great body of data which connects the presence of
B. coli with pollution includes all B. coli whether motile
or not, since scarcely any bacteriologists observe this
property in actual practice.

Howe (1912) has recently come to the conclusion that
motility has no diagnostic value. MacConkey, however
(MacConkey, 1909), after carefully reviewing the various
characters suggested for use in studying colon bacilli,
retains this one as important. He recommends that it be
made in a drop of a 6-hour broth culture on an ordinary
slide with a J-inch objective and dark ground illumina-
tion. Failure to show motility indicates in particular
B. lactis-aerogenes and B. pneumonias of his classifica-
tion (see p. 191).

Coagulation of milk is one of the most generally ac-
cepted tests for the colon group; and as a rule most
lactose fermenting forms give a positive reaction. The

176 ELEMENTS OF WATER BACTERIOLOGY

common practice in this country is to incubate litmus
milk tubes for 48 hours at 37° and then heat to boiling.
Tubes which have not coagulated spontaneously fre-
quently do so on heating. Biffi (1906) has pointed out
that milk to be used for this purpose should not be
sterilized in the autoclave. Temperatures above 100°
so alter the milk as to make its coagulation much slower.
Konrich (1910) concluded from his experiments that
the coagulation of milk by the colon bacillus is often not
due to acid production, but to the secretion of a specific
coagulating ferment, since he found that the addition of
an amount of acid similar to that produced by the
organism failed to coagulate.

The production of indol in a peptone solution is an-
other test very generally used in this country and in
England as diagnostic of " typical " B. coli. The
usual procedure has been to incubate a tube of an
aqueous solution containing i per cent peptone and
.01 per cent sodium nitrite for four days at 37° and to
test for indol by adding i c.c. of a .02 per cent solution
of sodium or potassium nitrite and i c.c. of a i to i
solution of sulphuric acid. Both the tube and the
reagents should be cooled on ice before mixing, and the
tube should be left in a cool place for an hour afterward
to allow time for the characteristic rose-red color of
nitroso-indol to develop.

Marshall (1907) and other German and English work-
ers have shown that this nitrite and sulphuric acid test
for indol often gives incorrect results and that the test
modified by Bohme from Ehrlich is both more sensitive
and more accurate. Two solutions are used; No. i is

VARIETIES OF COLON BACILLI 177

made up of 4 parts of para-dimethyl-amido-benzalde-
hyde, 380 parts of absolute alcohol, and 80 parts of con-
centrated HC1; No. 2 is a saturated aqueous solution
of potassium persulphate ; 5 c.c. of i is added to 10 c.c.
of a broth culture and then 5 c.c. of 2 is added and the
whole shaken. A red color indicates indol. Mac-
Conkey (1909) who was at first inclined to discard indol
as a routine test, believes that when made in this way
it is of much value.

Howe (1912), from a statistical study of 630 strains
of intestinal colon bacilli, concluded that indol, ammonia,
and nitrite tests were but slightly correlated with general
vigor and had but slight classificatory significance. It
does not necessarily follow, however, that this is true
of the forms which occur in stored waters. The so-
called " atypical B. coli " are of course rare in faeces but
they may occur in sufficient numbers to be important
in waters which have been remotely polluted.

The liquefaction of gelatin is another test generally
applied in any detailed study of the colon group. The
longer the tubes are kept the higher will be the propor-
tion of positive results, for there are many slowly lique-
fying forms grading by almost imperceptible degrees
into the commoner non-liquefying type. The table
cited from Gage and Phelps (1903) on p. 186 shows that
of a series of 1908 cultures from various waters, sewages,
and shellfish 8 per cent liquefied gelatin in 4 days, 10
per cent in 7 days, 13 per cent in 10 days, and 17 per
cent in 14 days.

The reduction of nitrates to nitrites has been used in
the United States as one of the five standard tests for

178 ELEMENTS OF WATER BACTERIOLOGY

B. coli, but it has never gained wide acceptance in Eng-
land or Germany. The usual practice has been to
incubate for 4 days at 37° and to test for nitrites by
adding a drop of each of the following solutions in suc-
cession :

A. Sulphanilic acid 5 gram

Acetic acid (25% sol.) 150.0 c.c.

B. Naphthylamine chloride o. i gram

Distilled water 20.0 c.c.

Acetic acid (25% sol.) 150.0 c.c.

A red or violet coloration indicates the presence of nitrites.

In making the nitrite test it is important to remember
the possibility that appreciable amounts of nitrite may
be present in the media — either derived from the air or
from the use of impure peptone (Wherry, 1905). In the
case of the nitrite reaction control tubes should always
be tested from the same batch of media and only a
distinct red color should be considered positive. The
nitrite test is particularly subject to variations of un-
explained origin. Of two duplicate tubes inoculated in
the same way, one may show a strong reaction and the
other none.

The reduction of neutral red has been extensively used
in England and less in this country. It has been re-
ferred to in Chapter VI. as one of the tests suggested for
use as a presumptive indicator of the colon group as a
whole. MacConkey (1909) concludes that both the ni-
trate test and the neutral red test should be dropped
from the procedure used in identifying colon bacilli, since
so many organisms give these reactions that they have
little significance. Houston, however, in the important
investigations which will shortly be discussed, used the

VARIETIES OF COLON BACILLI 179

production of a greenish fluorescence in neutral red
broth as one of his tests of " typical " B. coli and ap-
parently found it valuable.

Fermentation of various carbohydrate media has be-
come the most common method of subdividing the
bacilli of the colon group during the last few years, largely
as a result of the work of MacConkey. Sugar broths for
this test are generally put up in fermentation tubes of
some sort so that the gas formation may be observed
and perhaps measured, while acid production may be
indicated by the addition of litmus or accurately deter-
mined by titration. The old-fashioned fermentation
tube with a bulb and a stand has given way in most
water laboratories to a plain bent tube of even bore
and, more recently, to a still simpler device, a small vial
inverted in an ordinary test-tube of sugar broth. The
air in the top of the vial is driven out on sterilization
and the presence or absence of gas can be easily deter:
mined, although it is not possible to measure its quan-
tity with accuracy. If an ordinary bent tube is used
the amount of gas in the closed arm may be conven-
iently measured by the Frost gasometer (Frost, 1901).
If a measurement of the gas ratio is desired a few centi-
meters of strong sodium or potassium hydrate are added
and mixed with the broth by cautiously tipping the
tube; a second measurement determines the amount of
gas absorbed (assumed to be C02) .

It has been pointed out in Chapter VI that the gas
ratio appears to be a reaction of slight importance as
thus determined.

The list of fermentable substances used by various

180 ELEMENTS OF WATER BACTERIOLOGY

observers in classifying colon bacilli is a long one. It
was pointed out by Smith (1893) long ago that saccharose
divides these organisms into two groups, and Winslow
and Walker (1907) have found that strains which
attack saccharose generally ferment raffinose also.
MacConkey (1909) has made the most careful study
of the reactions of the group during recent years. He
believes that of the ordinary tests, milk and neutral
red should be discarded and fermentative reactions in
saccharose, dulcite, adonite, inulin, inosite and mannite
should be substituted in addition to motility, the indol
test, the liquefaction of gelatin, and the Voges and
Proskauer reaction.

Howe (1912), from his recent statistical study of 630
strains of intestinal bacilli, concluded that fermenta-
tion tests in mannite, dulcite, and starch media are
of little value in the classification of colon bacilli, since
they are not closely correlated with other characters.
It should be noted, however, that he worked only with
fresh intestinal strains and it is possible that types
characterized by definite reactions in these media may
be rare in faeces but may develop so as to be important
in stored waters.

The Voges and Proskauer reaction has been extensively
used by MacConkey and his followers in England and
by Bergey and Deehan (1908) in this country. After
the carbon dioxide in the fermentation tube has been
absorbed by caustic soda, if the tube be allowed to stand,
an eosin-like color gradually develops in the open
arm, due to the presence of acetyl-methyl-carbinol.
West (1909) points out that the test used by Rivas,

VARIETIES OF COLON BACILLI 181

the boiling of 1-4 c.c. of a 48-hour dextrose broth
culture with 5 c.c. of a 10 per cent caustic soda solu-
tion, is a quick method of obtaining the Voges and
Proskauer reaction. A yellow color is produced under
these conditions by the sugar alone, and a pinkish
eosin-like color when the acetyl-methyl-carbinol is
present. The reaction is hastened by shaking or
blowing into the tube to promote oxidation. West
confirms the conclusion of MacConkey and Bergey
that this reaction is characteristic of the B. lactis-
aerogenes and B. cloacae types (saccharose positive,
dulcite negative organisms).

The characters of colonies on various solid media,
such as Endo agar or Conradi-Drigalski agar appear
to be of minor importance, usually depending on one
of the simple fermentative reactions for their differential
value. The aesculin test and the Eijkman test (fer-
mentation at 46°) have been discussed in connection
with their use as enrichment procedures in Chapter VI.

Biological Significance of Variations in the Colon
Group. The general view among water bacteriologists
has been that forms differing from the " typical " B.
coli in one or more respects represented original intes-
tinal types weakened by a prolonged sojourn in an
unfavorable environment. As Whipple says (Whipple,
1903), " The type form of Bacillus coli is one which
can be defined within reasonably narrow limits, but
when the organism has been away from its natural
habitat for varying periods of time, and has existed
under abnormal conditions, its ability to react nor-
mally to the usual tests appears to be greatly im-

182 ELEMENTS OF WATER BACTERIOLOGY

paired. Its power to reduce nitrates may be lost, or
on the other hand, may be increased; its power to
produce indol may be lost, or on the other hand, it
may be increased; its power to coagulate milk, even,
is sometimes reduced, although seldom entirely lost;
its power to ferment carbohydrates may be altered
so that the amount of gas obtained in a fermentation-
tube, as well as its ratio of H to C02, is quite abnormal.
But in spite of all these facts, the bacillus tested may
have been originally a true Bacillus coli."

The results obtained by Peckham (1897) suggest
that the indol reaction in particular is highly variable.
By successive daily transfers in peptone broth she was
able to increase the amount of indol produced by nor-
mal B. coli, and by a longer continuance of the same
process to again weaken and abolish the power of form-
ing it. Gas formation too was slackened in the cul-
tures grown for too many transfers in the same medium.
Horrocks (1903) found that B. coli kept in unsterilized
well-waters and tap waters and in sterilized sewage and
Thames water for 2 to 3 months showed only a feeble
indol production and a delayed action on milk and
neutral red. These modified forms are sometimes
called " atypical B. coli," or " para-colon bacilli," and
Vincent gives them the picturesque name, " microbic
satellites of B. coli."

' Such anomalies are most frequent with cultures
freshly isolated from water, and they may often be
avoided, as Fuller and Johnson (1899) have shown,
by subjecting the organism to a process of preliminary
cultivation. For this purpose the American Public

VARIETIES OF COLON BACILLI 183

Health Association Committee recommends three suc-
cessive cultivations in broth at 20 degrees, each of 24
hours' duration, inoculation from the last broth tube
of a gelatin plate which is incubated for 48 hours at 20
degrees, inoculation of an agar streak from one colony
on the plate and incubation of this streak for 48 hours
at 20 degrees.

Often, however, the differences between types of the
colon group indicate something much more funda-
mental than temporary weakening due to unfavorable
environment. In particular sudden more or less perma-
nent mutations may suddenly appear. Twort (1907)
reports that by continued cultivation in sugar media he
was able to develop fermentative power in certain mem-
bers of the Gartner group which lacked such powers
before.

The work of Massini, Miiller, Sauerbeck, Konrich
and others (well summarized by Konrich, 1910) has
also shown that mutations capable of fermenting sugars
may suddenly arise from a parent strain lacking this
power. Burri (1910) has contributed to the same
problem and has found that the latent power to ferment
a given sugar is released by growing the organism on
that particular sugar, but that as Konrich and the others
show only a certain proportion of the cells develop
this power. The most important recent studies of
bacterial mutation have been made by Penfold. In
his latest communication (Penfold, 1912) he shows
that many bacteria of the colon-typhoid group pro-
duce a mutant capable of fermenting lactose, that
all strains of the typhoid bacillus produce dulcite and

184 ELEMENTS OF WATER BACTERIOLOGY

iso-dulcite mutants, that many paratyphoid and Gartner
group bacilli produce raffinose mutants, and that other
mutations also occur. The general phenomena are
the same in each case. A strain which normally fails
to ferment a given carbohydrate is grown upon a solid
medium containing that carbohydrate. As the colonies
develop there appear upon them raised papillae of a
different consistency from the rest of the colony and
colored red if litmus be present. Transplants from
the papillae give pure cultures of a strain fermenting
the carbohydrate in question and forming no papillae.
Transplants from the other portions of the colony
give the original strain, non-fermenting, but capable
of producing fermenting mutants as before. The
identity of derivative strain in all other respects has
been made clear by exhaustive cultural tests and serum
reactions ; and Penf old has shown that the whole proc-
ess may be repeated, starting from an isolated single
cell.

The work of MacConkey and Clemesha, which will
be discussed shortly, is based on the assumption that
a great number of minute subdivisions of the colon
group, whether they have arisen by the gradual modify-
ing effect of an unfavorable environment, or by muta-
tions, or in some other way, are for practical purposes
fairly permanent entities which they describe and name
as definite species.

Distribution of Types of the Colon Group in Waters
of Various Kinds. The sanitary importance of a study
of these minor types within the colon group depends
on the assumption that a certain set of characters is

VARIETIES OF COLON BACILLI 185

generally associated with forms fresh from the intes-
tine and may therefore be called typical. Such " typ-
ical" B. coli are understood as a rule to be motile,
to clot milk, produce indol, reduce nitrate and neutral
red and to fail to liquefy gelatin. It seems clear that
forms having these characters predominate in the
intestine itself while differing or " atypical" forms
bear to them somewhat the relation implied in Vincent's
term, satellites. Houston (1903*) examined in detail
10 1 cultures of coli-like microbes isolated from faeces and
found that 72 per cent of the cultures were typical in
all respects, while n per cent more differed only in
being non-motile. The remaining 17 per cent were
atypical, reacting abnormally to milk, indol, neutral
red, litmus whey or Capaldi and Proskauer's medium.
In a later investigation, Houston (1904) made a careful
study of the distribution of the atypical forms in faeces,
sewage, polluted water, and the filtered water-supplies
of London. According to his ingenious system of
nomenclature, " fl " indicates an organism which pro-
duces green fluorescence in neutral red broth; " ag,"
one which forms acid and gas in lactose media; " in,"
one which produces indol; and " ac " one which acidifies
and clots litmus milk. The combination of all these
properties gives " Flaginac," or typical B. coli; " aginac"
is a form which fails to reduce neutral red; " flagac,"
one which fails to form indol, etc. " Flaginac " B. coli
form the great majority of coli-like microbes in faeces,
but Houston found that in filtered water they are
outnumbered by atypical forms, of which he recognized
thirty-five distinct types.

186 ELEMENTS OF WATER BACTERIOLOGY

PERCENTAGE OF CULTURES PASSING VARIOUS TESTS IN THE ROUTINE EXAMINATION FOR
B. COLI AT THE LAWRENCE EXPERIMENT STATION OF THE MASSACHUSETTS STATE BOARD
OF HEALTH (GAGE AND PHELPS, 1903)

•ii°o 'a

Ov 00 Ov co M CO
vo vo vo rt* T^- ••^-

00
vo

*Including cultures which failed to grow on agar and streptococcus cultures, giving a very scanty, non-characteristic growth.

Per Cent of Cultures.

"rt

O O co r^ oo oo M

H <N

O O vo vo vo O vo

M (N

« "SABQ t7!

o1 - ••

O O co vo O 00 O

CO M M M <N CO

i-5 55 'sA^Q oi

O CO Ov O ON 00 O

<N M d CN

H 'SABQ Z,

O vo t>. ON ON OO vo
<N M (N

•SABQ *

<N 1-1 Cvl

~ ^ -papinuM

O hi •*$• l-< O O [~*

^3* -onsuaiowBqo

+ QS ro ^" ^1~ cs co
- O 00 00 00 t^ vo

•jopui aonpojd

1 co M vo vo Qv OO

VO OO

•sa^BJ^i^ aonpa^

'^TUK 81Bin3BO3

+ O O 00 M i>- oo
00 Ov 00 Ov vo OO

•uinoc
'H^oig 8SOJ1 9onpoj,j

O co ro O co O 1O

uot^'B5.u9uiJ91>j 'SBQ gonpojfj

+ M ^J- LO OO vO t>*
OO 00 t^» O vo vO

•Bipap^ AaoiBiiuuuoQ o^ui
pa^BJnoouj ssan^in3 jo aaquift^

• (N (N H CS VO

Per Cent of
Cultures.

•snoosjA

O M ^" H H CO O
<N

•paRauM-ntia

O to -3- 00 <N rf oo

^•H^oao oN

O O ON ^}" co O ^O

-anqmpunio

1 vo co t^- ^J" ^i" ^
~T~ oo oo oo vo t^- t^.

•WVVKH"*!"*

CO vo O vN CO Qv
• t^ 00 vO M vo t^
• tN ^ M CO

-U9uu9jj AjBUiuiijai j jo ^U9;3 J9<J

• O • cs OO co

Totals

'• t-l '.

. <L>

I S3 B •

' p4 t^ J S3 ;

VAEIETIES OF COLON BACILLI 187

On the whole, much of the English evidence tends to
the assumption that the atypical forms, or " paracolon
bacilli," generally represent weakened strains from the
intestinal B. coli stock. As Savage says, " we know
that nearly all the coli-like organisms in faeces are quite
typical B. coli, that in sewage a good many atypical
varieties are present, and that in contaminated water
and soil the proportion present is still larger."

The data tabulated on p. 186 from Gage and Phelps
(1903) lead to a similar conclusion. About 60 per cent
of the cultures isolated from polluted river water,
filtered water, and sewage proved to be typical B. coli,
while 41 and 43 per cent of those isolated from spring
water and shellfish, respectively, and 48 per cent of
those from ice belonged in this class.

Contradictory results, indicating a higher proportion
of typical forms outside the body than within it, have
been obtained by Konrich (1910) in the examination of
2387 different strains isolated in about equal propor-
tions from faeces, earth, and water. Of the 2387 coli-
like microbes studied, 308 were excluded by micro-
scopic examination (showing abnormal morphology or
positive Gram reaction) or by their liquefaction of
gelatin. The other 2079 strains were tested in sugar
media and peptone water with the results tabulated
on p. 188.

We are inclined to attribute these results of Konrich's
largely to the technique which he employed. It seems
to be clearly stated in his paper that he obtained his
faecal cultures by direct plating on solid media, while
his earth and water samples were treated to preliminary

188 ELEMENTS OF WATER BACTERIOLOGY

BIOCHEMICAL REACTIONS OF 2079 ORGANISMS OF THE
COLON GROUP

(KONRICH, 1910)

Medium.

Reaction.

Percentage of Positive
Results.

Faecal
Strains.

Earth
Strains.

Water
Strains.

Dextrose broth
Dextrose broth, 46°. . . .
Lactose broth
Lactose broth.

Gas production
Gas production
Gas production
Acid production ....
Coagulation

IOO

77
97
65

54
38

IOO

87
IOO

68
57

IOO

82
92
IOO

Milk
Neutral red dextrose
agar
Peptone solution
Endo agar.

Fluorescence

Indol production. . . .
Deep red colonies
with greenish luster

enrichment in sugar broths. Such an enrichment would
undoubtedly tend to increase the proportion of typical
B. coli. Konrich's own experiments on the storage
of pure cultures of colon bacilli in water showed a gen-
eral, though not invariable, relative increase of atypical
forms as the sojourn in water became long continued.
Houston (1911) has recently pointed out the danger
that the selective action of enrichment media may
confuse the normal relations of the bacterial flora;
and in order to obtain a more accurate idea of the
relations involved he made an elaborate study of raw
water, stored water, and stored and filtered water by
direct plating, preceded by the use of various physical
concentration methods. The results indicated in the
table on p. 189 are of interest.

VARIETIES OF COLON BACILLI

189

CHARACTERISTICS OF DEXTROSE FERMENTING BAC-
TERIA FROM RAW, STORED, AND FILTERED WATERS

(HOUSTON, 19 n)

Per Cent Positive Results.

, «

Gas Production.

en O

•g

n* "

Source.

JQ
U

1§

"3
a
i— i

o
a

|§

Raw river

water .

320

I c;

O 3

T r

Stored water. .

232

0.4

Stored and fil-

tered water .

225

The strains having the combination of positive reac-
tions in lactose and peptone solution made up 53 per cent
of those isolated from the raw waters. 46 per cent of
those from the stored waters, and 34 per cent of those
from the stored and filtered waters.

MacConkey's Classification of the Colon Group.
MacConkey is not satisfied with this general classifica-
tion of colon bacilli into typical and atypical forms,
but wishes to go much further, believing that even the
so-called " typical B. coli " chould rather be con-
sidered a complex including a considerable number of
definite individual types. After a detailed study of the
reactions of the lactose-fermenting bacteria of faeces
he outlined a new classification based on fermentative
reactions in the rarer sugars (MacConkey, 1905).
Using saccharose and dulcite he first divided the lactose-
fermenting organisms into four groups as indicated in
the table on p. 190.

190 ELEMENTS OF WATER BACTERIOLOGY
PRIMARY SUBDIVISIONS OF THE COLON GROUP

Saccharose.

Dulcite.

Type.

B. acidi-lactici
B. coli (or B. communis)
B. neapolitanus (or B. communior)
B. aerogenes

The fourth group, fermenting saccharose, but not
dulcite, was further subdivided by MacConkey into
the B. coscoroba type, which does not liquefy gela-
tin or give the Voges and Proskauer reaction, the B.
lactis-aerogenes type, which does not liquefy gelatin,
but does give the Voges and Proskauer reaction, and the
B. cloacae type, which liquefies gelatin and gives the
Voges and Proskauer reaction. Records of the preva-
lence of the four principal groups in human and animal
faeces and in milk are given in MacConkey's two papers
(1905 and 1909) as well as their relative numbers in a
suspension of faeces in water after various intervals of
time. The results do not, however, appear to us to
justify any important practical conclusions.

In his later paper MacConkey (1909) carried the
sub-division of the colon group much further. He
isolated 497 lactose-fermenting bacilli from the faeces
of man and animals, from sewage, water, grains, etc.
All were Gram-negative, fermented lactose, coagulated
milk and reduced nitrate. They were subdivided
by their action on gelatin, pepton and various fer-
mentable substances and by their motility into over
100 types of which the more important have received

VARIETIES OF COLON BACILLJ

191

names. The principal types of this classification are
indicated in the table below:

MAcCONKEY'S CLASSIFICATION OF THE COLON GROUP

(MACCONKEY,

Group.

No.

Name.

Liquefaction of
Gelatin.

Indol Produc-
tion.

Fermentation of

J|
II

Saccharose.

Dulcite.

Adonit.

Inulin.

4
5

;

B. acidi-lactici
B. levans

B. Griinthal
B. vesiculosus

II. ! 34
35

B. coli communis
B. Schafferi

;

III.

B. oxytocus pernicio-

sus

B. pneumonias

B. neapolitanus

IV.

103
104
107
108

B. lactis aerogenes. . .
B. gasoformans

;

•r-

=fc

B. coscoroba

B. cloacae

In this country the MacConkey classification was
first adopted by Bergey and Deehan (1908). These
workers used 8 diagnostic characters, motility, indol
production, liquefaction of gelatin, the Voges-Pros-
kauer reaction, and the fermentation of saccharose,
dulcite, adonite and inulin. They tabulated 256 different
combinations of these 8 characters and in the examina-

192 ELEMENTS OF WATER BACTERIOLOGY

tion of 92 colon-like bacilli from 50 samples of milk,
8 of sewage and i of kefir they found 43 of the possible
combinations.

Copeland and Hoover (1911) have recently urged
the importance of these fermentative reactions in the
rarer carbohydrates in the study of the colon group.
They confirm the positive Voges and Proskauer reaction
reported by other observers for B. lactis-aerogenes and
B. cloacae and point out that B. lactis-aerogenes is
the only form in a considerable series studied which
gives a brown coloration in aesculin media in one day.
On the other hand they record a positive dulcite reac-
tion for B. lactis-aerogenes and B. cloaca3 which is
highly confusing and makes it difficult to interpret
their results. Both these names according to the usage
of MacConkey, which has been accepted for the past
five years, are applied to dulcite-negative saccharose-
positive organisms.

Still another classification of the colon group is Jack-
son's modification of MacConkey's scheme in which
MacConkey's four primary groups are symmetrically
subdivided according to reactions in mannite and rafn-
nose with motility, indol production, nitrate reduc-
tion, liquefaction of gelatin and coagulation of milk as
secondary differential characters (Jackson, 1911). Under
each of the four groups, B. communior (MacConkey's
B. neapolitanus) , B. communis (MacConkey's B. coli),
B. aerogenes (MacConkey's Group IV), and B. acidi-
lactici, he distinguishes four types, A (fermenting
both mannite and raffinose, B (mannite +, raffinose — ),
C (mannite — , raffinose -}-), and D (fermenting neither

VARIETIES OF COLON BACILLI

193

mannite nor raffinose); and he indicates reactions in
other media by subscript letters. These types with
their subtypes are fully discussed in the last report
of the Committee on Standard Methods of Water
Analysis (1912).

Clemesha's Investigation of Stored Waters in India.
The most suggestive contribution to this subject which
has been made in recent years is a book by Major W.
W. Clemesha of the Indian Medical Service on The
Bacteriology of Surface Waters in the Tropics (Clemesha,
1912*), in which a vigorous argument is made for the
MacConkey classification in practical water work.
Major Clemesha's researches show the prevalence
of considerable numbers of all of MacConkey's primary
types in human and bovine fasces as indicated in the
table below, although the relative proportions found
in England and in India do not correspond very closely.
Clemesha's percentages are of special importance
because they are based in each case on over 1000
colonies.

RELATIVE PROPORTION OF MACCONKEY'S GROUPS IN
HUMAN F/ECES AND IN COW DUNG

Human Fasces.

Cow Dung.

Group.

MacConkey.

Clemesha.

MacConkey.

Clemesha. *

Both in human faeces and in cow dung Clemesha
finds the prevailing types to be B. coli, B. Grim thai,

194 ELEMENTS OF WATER BACTERIOLOGY

and B. coscoroba, the three together usually making
up 75 per cent all the lactose-fermenting organisms
present. A very interesting point brought out in these
investigations was the occurrence of " epidemics,"
of particular types which at certain periods become
suddenly frequent, usually prevailing in human faeces,
cow faeces and water supplies at the same time. (It
should be noted for the benefit of anyone studying
Clemesha's book that the tabular classification of the
colon group at the end contains a serious misprint.
B. lactis-aerogenes, B. gasoformans, B. coscoroba and
B. cloacae are there given as saccharose negative,
whereas they should be saccharose positive.) The
discussion in the text, however, appears to refer to the
orthodox MacConkey types. Clemesha (191 2a) made
a number of experiments on the relative resistance of
the various lactose-fermenting types by placing faecal
emulsions, with or without sand, in shallow dishes
in the sunlight and at various intervals isolating 10
colonies of the predominant types and working out
their fermentative reactions. In general the experiments
showed B. coli to be the dominant form at the beginning.
It quickly disappeared, however, and after a few hours
B. lactis aerogenes, B. acidi-lactici, B. cloacae and
others appeared. At the end of the experiments,
often on the second day, B. Grim thai or B. cloacae were
generally the only forms surviving. In a long series of
examinations of Red Hills Lake Clemesha obtained 138
colonies of lactose-fermenting organisms during rainy
periods and of these 59 belonged to MacConkey's
Group I, 10 to Group II, 14 to Group III and 55 to

VARIETIES OF COLON BACILLI

195

Group IV. Of 280 colonies isolated during dry periods,
37 belonged to Group I, 22 to Group III and 221 to
Group IV. When the forces of self-purification had
been at work, Group II (B. coli) entirely disappeared
and Group IV (B. cloacae and B. coscoroba) was pre-
dominant. B. Griinthal was the commonest of the
Group I forms. B. cloacae was especially prevalent
in bottom samples.

A study of a number of rivers in Bengal gave the
results tabulated below.

RELATIVE PREVALENCE OF CERTAIN LACTOSE-
FERMENTING TYPES IN BENGAL RIVERS

MacConkey
Group.

Types.

Dry Weather,
Dec.-June.

Wet Weather,
July-Nov.

I.
II.

IV.

Do.

B. Griinthal and B. vesiculosus
B. coii communis
B. lactis aerogenes
B. cloacae

41
3
7
ii

23
13

There are many irregularities in Dr. Clemesha's
results. For example, B. aerogenes, as well as the other
representatives of Group IV, was more abundant in Red
Hills Lake during dry periods than at times of rain.
On the whole, however, it does seem clear that his
results justify a general classification of the lactose-
fermenting organisms into three main groups accord-
ing to resistance. B. coli communis and B. oxytocus
perniciosus (representing MacConkey's Groups II and
III, both fermenting dulcite) are sensitive organisms
found in numbers only where pollution is fresh. B. lactis-
aerogenes, representing the subgroup of MacConkey's

196 ELEMENTS OF WATER BACTERIOLOGY

Group IV which ferments adonit, but does not form
indol or liquefy gelatin, occupies a somewhat inter-
mediate position, appearing in waters which have been
fairly recently polluted and later disappearing again.
Finally B. Grim thai and B. vesiculosus (MacConkey's
Group I, negative in both saccharose and dulcite)
and B. cloacae and B. coscoroba (of MacConkey's
Group IV, dulcite negative and saccharose positive),
are highly resistant organisms which occur in relatively
high proportions in stored waters. B. cloacae is most
abundant in bottom sediments and B. Griinthal and B.
vesiculosus in sunned surface-waters.

The moral drawn by Major Clemesha is that for
Indian conditions with waters stored in warm sunned
lakes and large rivers, where sensitive faecal bacteria
have ample opportunity to die out and resistant faecal
bacteria have an ample opportunity to multiply, it is
not proper to condemn water containing any members
of the colon group without distinguishing between the
more and the less resistant forms. For example, he
quotes 239 examinations of which only 74 showed no
B. coli according to the English standard, which closely
corresponds to our own, while 165 showed what we
should call positive results. Of the 165, however, 69
contained only the highly resistant B. Griinthal and
59 contained mixtures of other forms not belonging
to MacConkey's Group II (saccharose negative, dulcite
positive). Thus of the 239 samples 31 per cent would
have been passed by Houston's standard, 53 per cent
would have been condemned by Houston's standard,
although containing only resistant forms which Clemesha

VAEIETIES OF COLON BACILLI 197

believes to be unimportant, and 16 per cent would be
condemned by Clemesha as containing true B. communis.

Major Clemesha does not claim that these results
necessarily indicate any change of procedure in dealing
with the waters of temperate climates. Indeed, the
experience of English and American bacteriologists
offers pretty conclusive evidence that waters so stored
as to be safe do not contain large numbers of lactose-
fermenting organisms of any type. In other tropical
countries and perhaps in warm summer weather, the
Indian conditions may possibly be duplicated (as we
know they are in the case of the forms fermenting
dextrose but not lactose) ; and the experiments reported
in this book deserve the careful consideration of water
bacteriologists and sanitarians.

The results obtained by Houston (1911) in London
unfortunately do not correspond at all with these
Indian data. Houston studied in detail the reactions
of about 800 strains of dextrose-fermenting bacteria
from raw river-water, stored water, and stored and
filtered water. Comparison of the relative prevalence
of types from these three sources ought to furnish some
confirmation of Major Clemesha's conclusions, even
although the extreme conditions of warmth and sun-
light are lacking. We find, however, on careful study
of the figures that they do not. The Houston types
corresponding to B. communis, B. Schafferi and B.
neapolitanus (sensitive forms) are on the whole but
little more prevalent in the raw than in the stored
and filtered waters. On the other hand the types
corresponding to B. Griinthal, B. vesiculosus, B. cos-

198 ELEMENTS OF WATER BACTERIOLOGY

coroba and B. cloacae (Clemesha's resistant types)
are less abundant in the filtered and stored than in the
raw water. Houston's lactose-fermenting forms clas-
sified in MacConkey's four great groups show the rela-
tions indicated in the table below, which are almost
the reverse of what should be expected if the dulcite-
fermenting forms (Groups II and III) were indicative
of recent pollution.

DISTRIBUTION OF MACCONKEY'S GROUPS IN RAW,
STORED, AND FILTERED WATER AT LONDON

Pei

centage

Group.

Reactions.

Type.

Raw

Stored

Filtered

Water.

Saccharose — Dulcite —

B. acidi-lactici . . .

Saccharose— Dulcite +

B. coli

III

Saccharose+Dulcite+

B. neapolitanus. .

Saccharose + Dulcite —

B. lactis-aerogenes

Statistical Classification of the Colon Group. From
a biological standpoint, there is a twofold difficulty
with such a classification as that of MacConkey and
Jackson. In the first place it is enormously complex,
or soon becomes so, as new investigators add new
diagnostic tests. In the second place, it is entirely
arbitrary in its choice of the order in which particular
tests are to be used in splitting up the group. Closely
related forms may be widely separated if they chance
to differ in the one respect first chosen for dichotomic
division.

The best basis for a classification following natural bio-
logical lines seems to us to be the statistical method first

VARIETIES OF COLON BACILLI 199

suggested by Andrewes and Horder (1906) and Winslow
and Winslow (1908) in the study of the cocci. The
essential point about this method is that the characters
of the organisms studied are not considered indepen-
dently, but in relation to each other. The individual
reactions are first studied quantitatively in a considerable
series of allied strains, so that those types of reaction
which are manifested by a large number of strains
may be distinguished from the rarer intermediate
variations. In the second place, the correlations be-
tween different characters are used as a basis for group-
ing the types on the assumption that a coincidence in
several characters indicates a closer relationship than
any single character alone.

The statistical method has been applied to the colon
group in two extensive investigations, neither of which
has yet been published in full. Of the first by Howe, a
brief, abstract has appeared (Howe, 1912). The second
by L. A. Rogers, W. M. Clark, and B. J. Davis we
have had the opportunity of seeing in manuscript.
These two papers promise at last to lay a foundation
for a sound knowledge of the relationships of the colon
group.

Howe (1912) in his investigation dealt with 630
strains of fresh intestinal colon bacilli. He concluded
from his exhaustive study that in bacilli of this type
isolated directly from stools, the characters of motility,
indol formation, ammonia production, nitrate reduction,
fermentation of mannite, dulcite, and starch were not
sufficiently correlated with each other or with other
characters to be of classificatory value. Dextrose,

200 ELEMENTS OF WATER BACTERIOLOGY

lactose, saccharose and raffinose he found to constitute
a natural metabolic gradient, in the order named,
fermentation of any member of the series implying
fermentation of those preceding it. Fifty-three per cent
of his strains fermented all four sugars, 5 per cent all
but rafnnose, 41 per cent attacked dextrose and lactose
only, and i per cent dextrose alone.

CHAPTER IX
OTHER INTESTINAL BACTERIA

IT would be an obvious advantage if the evidence of
sewage contamination, furnished by the presence of
the colon group, could be reinforced and confirmed
by the discovery in water of other forms equally char-
acteristic of the intestinal canal. The attention of a
few bacteriologists in England and America has been
turned in this direction during the past few years; and
two groups of organisms, the sewage streptococci
and the anaerobic spore-bearing bacilli, have been
described as probably significant.

Significance of the Sewage Streptococci. The term
" sewage streptococci," as generally used, covers an
ill-defined group, including many cocci which do not
occur in well-marked chains. Those most commonly
found grow feebly on the surface of ordinary nutrient
agar, producing faint transparent, rounded colonies,
but under semi-anaerobic conditions flourish better,
giving a well-marked growth along the gelatin stab
and only a small circumscribed film on the surface.
They are favored by the presence of the sugars and
ferment dextrose and lactose, with the formation
of abundant acid but no gas. They are seen under
the microscope as cocci, occurring as a rule in pairs,

201

202 ELEMENTS OF WATER BACTERIOLOGY

short chains, or irregular groups. They do not show
visible growth and do not form indol and nitrite
in the standard peptone and nitrate solutions; most
of them do not liquefy gelatin, though occasionally
forms are found which possess this power. Until recently
no systematic study of the various species found in the
intestine had been made and all cocci giving the char-
acteristic growth on agar and strongly fermenting
lactose are commonly included as " sewage streptococci."
Although the significance of the streptococci as sewage
organisms is not established with the same defmiteness
which marks our knowledge of the colon group, these
forms have been isolated so frequently from polluted
sources and so rarely from normal ones that it now seems
reasonable to regard their presence as indicative of
pollution. Although originally reported by Laws and
Andrewes (Laws and Andrewes, 1894), their importance
was not emphasized until 1899 and 1900, when Hous-
ton (Houston, i899b, i9Oob) laid special stress upon
the fact that streptococci and staphylococci seem to
be characteristic of sewage and animal waste, the former
being, in his opinion, the more truly indicative of
dangerous pollution, since they are " readily demon-
strable in waters recently polluted and seemingly
altogether absent from waters above suspicion of con-
tamination." In six rivers recently extensively sewage-
polluted, he found streptococci in from one-tenth to
one ten-thousandth of a c.c. of the water examined,
although in some cases the chemical analysis would not
have indicated dangerous pollution. On the other
hand, eight rivers, not extensively polluted, showed

OTHER INTESTINAL BACTERIA 203

no streptococci in one- tenth of a c.c., although the
chemical and the ordinary bacteriological tests gave
results which would condemn the waters. Horrocks
(Horrocks, 1901) found these organisms in great abun-
dance in sewage and in waters which were known to be
sewage-polluted, but which contained no traces of
Bacillus coli. He found by experiment that B. coli
gradually disappeared from specimens of sewage kept
in the dark at the temperature of an outside veranda,
while the commonest forms which persisted were varieties
of streptococci and staphylococci.

In America attention was first called to these organisms
by Hunnewell and one of us (Winslow and Hunnewell,
1902*), and the same authors later (Winslow and Hunne-
well, i902b) recorded the isolation of streptococci from
25 out of 50 samples of polluted waters. Gage (Gage,
1902), from the Lawrence Experiment Station, has
reported the organisms present in the sewage of that
city, while Prescott (i9O2b) has shown that they are
abundant in faecal matter and often overgrow B. coli in
a few hours when inoculations are made from such
material into dextrose broth. In the monograph of
Le Gros (Le Gros, 1902) of the many streptococci
described, all without exception were isolated, either
from the body or from sewage. Baker and one of us
(Prescott and Baker, 1904), found these organisms
present in each of 50 samples of polluted waters. On
the other hand, in the study of 259 samples of presuma-
bly unpolluted waters, by the method of direct plating,
Nibecker and of the authors (Winslow and Nibecker,
1903) found streptococci in only one sample. Clemesha

204 ELEMENTS OF WATER BACTERIOLOGY

(191 2a) finds that streptococci in India are present
in .0001 or .00001 gm. of faeces, but are rare in waters
unless very grossly polluted. In a series of bottle
experiments in the laboratory and in the study of an
artificially polluted tank outdoors he showed that they
disappear very rapidly in water, within 2 or 3 days
at the most. Gordon (1904) showed that certain strep-
tococci are abundant in normal saliva and are found
in air which has been exposed to human pollution but
not in normal air. On the whole there can be no doubt
of the fact that streptococci occur on the surfaces of
the human and animal body more commonly than
anywhere else in nature.

Isolation of Sewage Streptococci. The isolation of
these organisms either from plates or liquid cultures is
easy. On the lactose-agar plate, made directly from
a polluted water, the colonies of the streptococci may
generally be distinguished from those of other acid-
formers by their small size, compact structure, and
deep-red color, which is permanent, never changing
to blue at a later period of incubation. Developing
somewhat slowly, however, they may be overlooked
if present only in small numbers. In the dextrose-
broth tube, streptococci will generally appear in abun-
dance after a suitable period of incubation. Prescott
and Baker, in the work above mentioned, found that
with mixtures of B. coli and streptococci in which the
initial ratios of the latter to the former varied from
i : 94 to 208 : i, the colon bacilli developed rapidly
during the early part of the experiment, reaching a
maximum after about 14 hours, and then diminishing

OTHER INTESTINAL BACTERIA

205

rapidly. The streptococci first became apparent after
10 to 15 hours and reached their maximum after 20 to
60 hours, according to the number originally present.
Applying the same method to polluted waters, similar
periodic changes were observed; nearly pure cultures
of B. coli were first obtained, then the gradual displace-
ment of one form by the other took place, and at length

RELATIVE GROWTH OF B. COLI AND SEWAGE STREPTO-
COCCI FROM POLLUTED WATERS IN DEXTROSE BROTH

(PRESCOTT AND BAKER, 1904)

Sample Number

Red colonies developing 1

from i c.c. of original sam- }
pie on litmus lactose agar J

460

1250

105

B. coli

200 185 400

130

332

420

410

hrs.

Strept

O O

o o o

0 0 0

B. coli

2OO

76 I3O 27O 220 2IO

140! 420

285

410

Number found,
in millions per

hrs.

Strept

25 20 10: 45 30

210 75

145

cubic centime-
ter, after,

B. coli

280

150 385 370

300 570

2OO

405

320

300

growth in dex-
trose broth for

hrs.

Strept

140

280 170

300 1700

350

370

350

various peri-
ods

B. coli

O 2IO

105

hrs.

Strept

474

420

480

300

390

170

400

105

250J

B. coli

hrs.

Strept

2 45

ISO

170

First gas noted after (hrs.).. .

8, 10

the streptococci were present either in pure culture
or in great predominance as shown by the accompany-
ing tables. The samples of water were plated directly
upon litmus lactose agar and the plates were incubated
at 37° for 24 hours, when the red colonies were counted.
At the time of plating, i c.c. from each sample was also
inoculated into dextrose broth in fermentation tubes,

206 ELEMENTS OF WATER BACTERIOLOGY

which were likewise incubated at 37°. After various
periods, as indicated by the tables below, the tubes
were shaken thoroughly and i c.c. of the contents
withdrawn. This was diluted (generally 1-1,000,000,)
with sterile water, plated on litmus lactose agar in the
usual way, and incubated for 24 hours. The colonies
of B. coli and streptococci were distinguished micro-

RELATIVE GROWTH OF B. COLI AND SEWAGE STREPTO-
COCCI FROM POLLUTED WATERS IN DEXTROSE BROTH

(PRESCOTT AND BAKER, 1904)

Sample Number

Red colonies developing from i c.c. 1
of original sample on litmus lac- j-
tose agar J

150

.04
O

170

. 12
O
380
128

200

• 55
o
330
80

IOO

30
1.6

Number found, in mil-
lions per cubic centi-
meter, after growth^
in dextrose broth for
various periods. . . .

7
hrs.

B. coli

.02

—

.OI

Strept.

i?
hrs.

B. coli

266

TOO

88 350

Sio

160

220

300

Strept.

150

• o! 40 140 240

27
hrs.

B. coli

520

610 72 700

IOOO

740
4380

7
60
35

Strept.

800

860

670

IO80
22
22

20
31

2500

66
70

7
52

3900

40
hrs.

B. coli

Strept.

252

330 260

16 38

hrs.

B. coli

Strept.

3-8

scopically, and by difference in color and general
characters.

The successive growth of these two intestinal groups
in the same dextrose-broth tube suggests the following
method for the detection of both B. coli and sewage
streptococci.

Inoculate the desired quantity of water, preferably

OTHER INTESTINAL BACTERIA 207

i c.c., into dextrose broth, in a fermentation tube,
and incubate at 37°. After a few hours' incubation
examine the cultures for gas. Within 2 or 3 hours'
after gas formation, is first evident, plate from the
broth in litmus lactose agar, incubating for 12 to 18
hours at 37°. If at the end of this time no acid-produc-
ing colonies are found, it is probably safe to assume that
there were no colon bacilli present. On the other
hand, if red colonies are developed, these must be fur-
ther examined by the regular diagnostic tests for B.
coli. After the first plating from the dextrose broth,
replace the fermentation tube in the incubator and allow
it to remain for 24 to 36 hours, then plate again on litmus
lactose agar. This plating should give a nearly pure
culture of streptococci if these organisms were originally
present in the water.

Streptococci as Indicators of Recent Pollution. The
comparative relation of the streptococci and the colon
bacilli to sewage pollution is still somewhat uncertain.
Houston (Houston, 1900) held that the former microbes
imply " animal pollution of extremely recent and there-
fore specially dangerous kind," and Clemesha's experi-
ments led to the same conclusion. Horrocks (Horrocks,
1901), on the other hand, maintains, largely on the
strength of certain experiments with stored sewage,
that the streptococci persist after colon bacilli have
disappeared and indicate contamination with old sewage
which is not necessarily dangerous. These discordant
results are probably to be explained by the different
media in which the viability of the bacteria was com-
pared. It seems likely that in sewage where there is a

208 ELEMENTS OF WATER BACTERIOLOGY

large amount of organic food material present the
streptococci may kill out the colon bacilli as they do
in the fermentation tube, and as we know they fre-
quently do in shellfish. This would explain Horrocks'
results. On the other hand, there is good evidence
that the streptococci are less resistant than B. coli to
the unfavorable conditions which exist in water of
ordinary organic purity. In waters of potable char-
acter B. coli is frequently present without the strep-
tococcus; and a negative test for streptococci has
little significance. A positive test, on the other hand,
furnishes valuable confirmatory evidence of pollution.
This evidence is of course of special importance when
through the activity of the streptococci themselves,
or from any other cause the colon isolation has yielded
an erroneous negative result.

The English Committee appointed to consider the
standardization of methods for the bacterioscopic
examination of water (1904) by a majority vote rec-
ommended the enumeration of streptococci, as a routine
procedure in sanitary water analysis, but in this
country the Committee on Standard Methods of Water
Analysis (1912) has concluded that "the information
afforded by the occurrence of these organisms seems
to be of less value than in the case of B. coli and it is
believed that for the present at least, the streptococcus
test is of subordinate importance."

Use of the Streptococci to Distinguish between Human
and Animal Pollution. There seems some reason to
hope that the streptococci may prove of assistance in
the important task of differentiating human and animal

OTHER INTESTINAL BACTERIA 209

pollution, a task in which all other tests have so far
failed. Unlike the colon bacilli, streptococci from the
intestines of cattle and men appear to belong to dis-
tinct types. The recognition of- this fact we owe
primarily to Gordon (1905), who made an elaborate
study of the fermentative power of the streptococci
in a long series of carbohydrate media. His work
and that of Houston (Houston, 1904; Houston, 1905*,
Houston, i905b) have made it clear that the streptococci
of the herbivora differ from those found in the
human body in their low fermentative power. In their
review of the genus, Andrewes and Horder (1906)
describe the type characteristic of the herbivora under
the name, Str. equinus, and define it by its failure to
ferment lactose, raffinose, inulin or mannite, or to
reduce neutral red. Five other types are described
from the human mouth and intestine; all of them
ferment lactose, and most reduce neutral red and fer-
ment raffinose. The commonest intestinal form clots
milk, reduces neutral red and ferments saccharose,
salicin, coniferin and mannite. The specific types
of the genus Streptococcus, grade into each other by
almost imperceptible degrees, and streptococci ferment-
ing lactose and raffinose and reducing neutral red are
sometimes found in bovine faeces; but the studies made
in this country by Winslow and Palmer (1910) confirm
the conclusions of the English observers that there are
specific differences between the streptococci of the
human, bovine, and equine intestines. The most im-
portant of these results are indicated in the table
below:

210 ELEMENTS OF WATER BACTERIOLOGY

COMPARATIVE FERMENTATIVE POWER OF STREPTO-
COCCI FROM THE HORSE, THE COW, AND MAN

(WINSLOW AND PALMER, 1910)

Streptococci.

Percentage of Positive Results (300 Strains).

Lactose.

Raffinose.

Mannite.

Human ....

62
8
52

4
28

28
2

Equine

Bovine

The rarity of lactose-fermenting streptococci in the
horse makes it probable that this group can be used for
distinguishing pollution by street washings from that
due to domestic sewage; and the fact that a considera-
bly larger proportion of human strains attack mannite
and a considerably larger proportion of bovine strains
ferment raffinose should make it possible to use the
ratio between results in these two media to distinguish
between the wash from pastures and cultivated land
and sewage. Clemesha (i9i2a) in India has, however,
obtained very different results. Out of 115 strains of
streptococci from human faeces 92 per cent belonged
to the "lamirasacsal" class of Houston (acid in lactose,
clot in milk, acid in raffinose, saccharose and salicin),
and none acidified mannite. Of 39 strains from cow
dung all belonged either to this same " lamirasacsal "
class or to the " larasacsal " class (differing only in
failing to clot milk). Nevertheless, in view of the
importance of distinguishing between human and animal
pollution and the hopelessness of doing so by means
of the colon group these different types of streptococci
well deserve further study.

OTHER INTESTINAL BACTERIA 211

The Anaerobic Spore-forming Bacilli. The English
bacteriologists have ascribed much importance as
indicators of sewage pollution to another group of organ-
isms, the anaerobic spore-forming bacilli, of which the
form described as B. aerogenes capsulatus (Welch
and Nuttall, 1892) and now called B. welchii, and the
form isolated by Klein (Klein, 1898; Klein, 1899) in
1895 in the course of an epidemic of diarrhoea at St.
Bartholomew's Hospital, described under the name of
B. enteritidis sporogenes (now called B. sporogenes)
are types.

The procedure originally described by Klein for
isolating B. sporogenes is as follows: a portion of the
sample to be examined is added to a tube of sterile
milk, which is then heated to 80° C. for 10 minutes
to destroy vegetative cells. The milk is next cooled
and incubated under anaerobic conditions, which may
be accomplished most conveniently by Wright's method.
A tight plug of cotton is forced a quarter way down the
test-tube, the space above is loosely filled with pyrogallic
acid, a few drops of a strong solution of caustic potash
are added, and the tube is tightly closed with a rubber
stopper. After 18 to 36 hours at 37° the appearance
of the tube will be characteristic if the B. sporogenes
is present. " 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 colorless, thin, watery whey, with a
few casein lumps adhering here and there to the sides

212 ELEMENTS OF WATER BACTERIOLOGY

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."

Since this organism is not present in very large num-
bers, even in sewage, the test of a water-supply must
be made with large samples, and the concentration
of at least 2000 c.c. of water by nitration through a
Pasteur filter is recommended by Horrocks as a necessary
prelude (Horrocks, 1901). The Committee on Standard
Methods of Water Analysis (1912) recommends the
following enrichment procedure for the isolation of B.
sporogenes which avoids physical concentration. Vari-
ous dilutions of the water to be tested are incubated
in fermentation tubes containing liver broth for 24 hours
at 37°. If B. sporogenes is present gas will be evolved
and a characteristic " vile odor " will be produced.
If this reaction is obtained the contents of each posi-
tive tube is transferred to an Erlenmeyer flask or large
test-tube and heated at 80° C. for 10 minutes to destroy
vegetative cells. One c.c. of broth containing sediment
is withdrawn from the bottom of each flask and enriched
once more in a fresh liver broth tube. B. sporogenes
will now usually be present in pure culture showing
large sluggishly motile bacilli containing spores. A
gelatin stab culture made from these 24-hour liver broth
tubes will show after 48 hours incubation at 20° a dis-
tinct liquefying anaerobic growth beginning about 2
cm. below the surface with gas bubbles at the top of the
liquefied area. In order to obtain absolutely pure
cultures it is necessary to fish from liver broth tubes

OTHER INTESTINAL BACTERIA 213

only 3-5 hours old as only young vegetative cells will
grow on plates. Transplants from the closed arm of
such tubes will grow on dextrose liver agar plates incu-
bated under anaerobic conditions.

The organisms of the B. sporogenes group are large
stout bacilli often occurring in chains. They liquefy
gelatin vigorously and on agar produce fine discrete
gray colonies. They vigorously ferment dextrose, lac-
tose and saccharose, producing acid and gas, and in sugar
agar each colony will be marked by one or more gas
bubbles surrounded by a delicate whitish fringe. The
organism is strongly pathogenic for guinea pigs, by
which character it is distinguished from the B. butyricus
of Botkin. B. welchii differs from B. sporogenes chiefly
in lacking motility and in forming spores with less read-
iness (Klotz and Holman, 1911).

The researches of Klein and Houston (Klein and
Houston, 1898, 1899) have shown that the B. sporogenes
occurs in English sewage in numbers varying from 30 to
2 200 per c.c. and that it is often absent in considerable
volumes of pure water. In Boston sewage it may
usually be isolated from .01 or .001 of a c.c. (Winslow
and Belcher, 1904). Since the spores of an anaerobic
bacillus may persist for an indefinite period in polluted
waters, their presence need not necessarily indicate
recent or dangerous pollution.

Vincent (1907) and other French observers consider the
determination of the total number of anaerobic bacteria
as significant, since the decomposition of organic matter
is accompanied by anaerobic growth. It is not claimed,
however, that bacteria of this type are characteristic of

214 ELEMENTS OF WATER BACTERIOLOGY

animal more than of vegetable decompositions, and the
total anaerobic count apparently adds nothing of impor-
tance to the information gained by the ordinary gelatin
plate method. The property of liquefaction was for-
merly believed to be of significance, inasmuch as the
liquefying bacteria were regarded as indicative of pollu-
tion. This position is, however, no longer tenable,
since many bacteria, typical of the purest waters, may
cause liquefaction.

As Savage says in summing up this question: " The
number of different species of organisms in sewage
is very great, and it is highly probable that many of
them occur in all specimens of ordinary sewage; but,
except for B. coli, streptococci, and B. enteritidis sporo-
genes, their presence has not been ascertained with
sufficient constancy, nor has their numerical occurrence
been sufficiently investigated to make them of value
as indicators of sewage pollution." (Savage, 1906.)

CHAPTER X

THE SIGNIFICANCE AND APPLICABILITY OF THE
BACTERIOLOGICAL EXAMINATION

Sanitary Inspection and Sanitary Analysis. The first
attempt of the expert called in to pronounce upon the
character of a potable water should be to make a
thorough sanitary inspection of the pond, stream, well
or spring from which it is derived. Study of the pos-
sible sources of pollution on a watershed, of the direc-
tion and velocity of currents above and below ground,
of the character of soil and the liability to contamina-
tion by surface-wash are of supreme importance in
interpreting the analyses to be made. In many cases,
however, the results of the sanitary inspection will
be found to be by no means conclusive. If house or
barnyard drainage or sewage is actually seen to enter
a water used for drinking purposes it is obviously
unnecessary to carry out delicate chemical or bacteri-
ological tests to detect pollution. On the other hand,
no reconnoissance can show certainly whether unpurified
drainage from a cesspool does or does not reach a
given well; whether sewage discharged into a lake
does or does not find its way to a neighboring intake;
whether pollution of a stream has or has not been
removed by a certain period of flow. Evidence upon

215

216 ELEMENTS OF WATER BACTERIOLOGY

these points must be obtained from a careful study of the
characteristics of the water in question, and this study
can be carried out along two lines, chemical and
bacteriological.

Sanitary Chemical Analysis. A chemical examina-
tion of water for sanitary purposes is mainly useful
in throwing light upon one point — the amount of decom-
posing organic matter present. It also gives an his-
torical picture which may be of much value. Humus-
like substances may be abundant in surface-waters
quite free from harmful pollution, but these are stable
compounds. Easily decomposable bodies, on the other
hand, must obviously have been recently introduced
into the water and mark a transitional state. " The
state of change is the state of danger," as Dr. T. M.
Drown once phrased it. Sometimes the organic mat-
ter has been washed in by rain from the surface of the
ground, sometimes it has been introduced in the more
concentrated form of sewage. In any case, it is a warn-
ing of possible pollution, and the determination of free
ammonia, nitrites, carbonaceous matter, as shown
by " oxygen consumed," and dissolved oxygen yield
important evidence as to the sanitary quality of a water.

Furthermore, nitrates, the final products of the oxida-
tion of organic matter, and the chlorine introduced as
common salt into all water which has been in contact
with the wastes of human life, furnish additional infor-
mation as to the antecendents of a sample. The results
of the chlorine determination are indeed perhaps more
clear than those of any other part of the analysis, for
chlorine and sewage pollution vary together, due allow-

BACTERIOLOGICAL EXAMINATION 217

ance being made for the proximity of the sea and other
geological and meteorological factors. Unfortunately,
it is only past history and not present conditions which
these latter tests reveal, for in a ground-water completely
purified from a sanitary standpoint such soluble con-
stituents remain, of course, unchanged. Thus, in the
last resort, it is upon the presence and amount of decom-
posing organic matter in the water that the opinion of
the chemist must be based.

Information Furnished by Bacteriological Examina-
tions. The decomposition of organic matter may be
measured either by the material decomposed or by the
number of organisms engaged in carrying out the proc-
ess of decomposition. The latter method has the advan-
tage of far greater delicacy, since the bacteria respond by
enormous multiplication to very slight increases in their
food-supply, and thus it comes about that the standard
gelatin-plate count at 20° roughly corresponds, in not
too heavily polluted waters, to the free ammonia and
" oxygen consumed," as revealed by chemical analysis.
If low numbers of bacteria are found, the evidence is
highly reassuring, for it is seldom that water could be
contaminated under natural conditions without the
direct addition of foreign bacteria or of organic matter
which would condition a rapid multiplication of those
already present. The bacteriologist in such cases
can declare the innocence of the water with justifiable
certainty. When high numbers are found the interpreta-
tion is less simple, since they may exceptionally be due
to the multiplication of certain peculiar water forms.
Large counts, however, under ordinary conditions,

218 ELEMENTS OF WATER BACTERIOLOGY

when including a normal variety of forms indicate the
presence of an excess of organic matter, derived in all
probability either from sewage or from the fresh wash-
ings of the surface of the ground. In either case danger
is indicated.

A still closer measure of polluting material may be
obtained from the numbers of colonies which develop
on litmus-lactose-agar at 37°, since organisms which
thrive at the body temperature, and particularly those
which ferment lactose, are characteristic of the intestinal
tract and occur but rarely in normal waters.

Gage (Gage, 1907) has shown that by counts at 20, 30,
40, and 50° C., information may be quickly obtained
which is of great assistance in judging the character
of the water.

" Modern methods of bacterial examination of water,
consisting usually of determinations of the numbers of
bacteria by means of plates incubated at room tempera-
ture, and of tests for the presence or absence of one or
two specific types, occasionally lead to an erroneous
interpretation of the quality of a water, owing to the
fact that they do not yield adequate data by which
abnormal and inaccurate results may be separated from
those which are truly indicative of purity or pollution.
Furthermore, as several days must elapse before the
bacterial tests can be completed, the results when
obtained may have passed their usefulness. If, however,
we can so modify our procedure that the varied char-
acter of the bacteria in waters of different classes may
be quickly and accurately recognized, the value of
bacterial water analysis will be enormously increased.

BACTERIOLOGICAL EXAMINATION 219

Much of this information may be obtained by the use
of selective media, selective temperatures, or by a
proper combination of the two.

" By the use of litmus-lactose-agar in place of agar
or gelatin we obtain similar counts of total bacteria,
and in addition are able to separate those bacteria into
two groups, which do and do not produce acid fermenta-
tion of lactose, and the numbers of the two classes of
bacteria so obtained indicate more completely the
character of the water than would the numbers of either
class alone. By incubating our plates at temperatures
of 30 or 40° C. we are able to obtain counts in 12 to 18
hours, which counts, while smaller than those on plates
incubated for a longer period at a lower temperature,
appear to be fully as significant. If we increase our
number of determinations by incubating duplicate
plates at two or more temperatures, the various results
and the ratios between them furnish a check upon one
another in addition to increasing the available data
upon which to base an interpretation." (Gage, 1907.)

Finally, the search for the Bacillus coli furnishes the
most satisfactory of all single tests for f a3cal contamina-
tion. This organism is preeminently a denizen of the
alimentary canal and may be isolated with ease from
waters to which even a small proportion of sewage has
been added. On the other hand, it is never found
in abundance in waters of good sanitary quality, and
its numbers form an excellent index of the value of
waters of an intermediate grade. The streptococci
appear to be forms of a similar significance useful as
yielding a certain amount of confirmatory evidence.

220 ELEMENTS OF WATER BACTERIOLOGY

The full bacteriological analysis should then consist
of three parts, the gelatin-plate count, as an estimate
of the amount of organic decomposition in process;
the total count, and the count of red colonies, on
litmus-lactose-agar, as a measure of the organisms
which form acids and thrive at the body temperature;
and the study of a series of lactose bile tubes for the
isolation of colon bacilli.

Special Advantage of the Bacteriological Examination.
The results of the bacteriological examination have, in
several respects, a peculiar and unique significance.
First, this examination is the most direct method of
sanitary water analysis. The occurrence of nitrites
or free ammonia in a small fraction of one part per
million, or of chlorine in several parts per million, do
not in themselves render a water objectionable or
dangerous. They merely serve as indicators to show
that germ-containing and germ-sustaining organic mat-
ter is present. By a determination of the chlorine
and study of the relations of carbon and nitrogen,
it is possible to determine with some degree of accuracy
whether this organic matter is of plant or animal origin,
and hence to rate its objectionable or dangerous char-
acter. By the bacteriological examination, on the
other hand, we are able to determine directly whether
particular kinds of organisms characteristic of sewage
are, or are not, actually present in the water. What
we dread in drinking-water is the presence of pathogenic
bacteria, mainly from the intestinal tract of man,
and it is quite certain that the related non-pathogenic
bacteria from the same source will behave more nearly

BACTERIOLOGICAL EXAMINATION 221

as these disease germs do than will any chemical com-
pounds. In the second place, the bacteriological
methods are superior in delicacy to any others. Klein
and Houston (1898) showed by experiment with dilu-
tions of sewage that the colon test was from ten to one
hundred times as sensitive as the methods of chemical
analysis; and studies of the self-purification of streams
have confirmed their results on a practical scale. Thus
in the Sudbury River it was found that while chem-
ical evidences of pollution persisted for 6 miles beyond
the point of entrance, the bacteria introduced could
be detected for 4 miles further (Woodman, Winslow,
and Hansen, 1902).

The statement is sometimes made that while bac-
teriological methods may be more delicate for the
detection of pollution in surface-waters, contamination
in ground-waters may best be discovered by the chemical
analysis. That such is not the case has been well
shown by Whipple (Whipple, 1903) who cites the fol-
lowing two instances in which the presumptive test
revealed contamination not shown by the chemical
analysis :

" A certain driven- well station was located in swampy
land along the shores of a stream, and the tops of the
wells were so placed that they were occasionally flooded
at times of high water. The water in the stream was
objectionable from the sanitary standpoint. The wells,
themselves were more than 100 feet deep; they pene-
trated a clay bed and yielded what may be termed arte-
sian water. Tests for the presence of Bacillus coli had
invariably given negative results, as might be naturally

222 ELEMENTS OF WATER BACTERIOLOGY

expected. Suddenly, however, the tests became positive
and so continued for several days. On investigation it
was found that some of the wells had been taken up to
be cleaned, and that the workmen in resinking them had
used the water of the brook for washing them down.
This allowed some of the brook- water to enter the system.
It was also found that at the same time the water in the
brook had been high, and because of the lack of packing
in certain joints at the top of the wells the brook-water
leaked into the suction main. The remedy was obvious
and was immediately applied, after which the tests for
Bacillus coli once more became negative. During all
this time the chemical analysis of the water was not
sufficiently abnormal to attract attention. On another
occasion a water-supply taken from a small pond fed
by springs, and which was practically a large open well,
began to give positive tests for Bacillus coli, and on
examination it was found that a gate which kept out
the water of a brook which had been formerly connected
with the pond was open at the bottom, although it was
supposed to have been shut, thus admitting a contam-
inated surface-water to the supply." Whipple also
calls attention to the report on the Chemical and
Bacteriological Examination of Chichester Well-waters
by Houston (Houston, 1901), in which the results of
chemical and bacteriological examinations of thirty
wells were compared. It was found that the bacteri-
ological results were in general concordant and satis-
factory. The wells which were highest in the number
of bacteria showed also the greatest amount of pollu-
tion, as indicated by the numbers of B. coli, B. sporo-

BACTERIOLOGICAL EXAMINATION

223

genes, and streptococci. On the other hand, the chlorine
and the albuminoid ammonia showed no correspondence
with the bacteriological results.

Vincent (Vincent, 1905) cites an interesting case of
the detection of progressive pollution of a ground-
water by bacteriological methods. The well of a
military camp in Algeria showed 200 bacteria per c.c.
before the arrival of a regiment of troops. Its sub-
sequent history is indicated in the table below:

PROGRESSIVE POLLUTION OF A WELL

(VINCENT, 1905)

Bacteria per c.c.

Bacillus coli per c.c.

Before arrival of troops

2OO

6 days after arrival

77O

14 days after arrival
41 days after arrival

4,240

6 060

I
2

60 days after arrival

14,900

Thirdly, negative tests for Bacillus coli and low bac-
terial counts may be interpreted as proofs of the good
quality of water, with a certainty not attainable by any
other method of analysis. Many a surface-water with
reasonably low chlorine and ammonias has caused epi-
demics of typhoid fever; but it is impossible, under any
natural conditions (except perhaps in a well polluted
with urine) that a water could contain the typhoid
bacillus without giving clear evidence of pollution in the
bile tube or on the lactose-agar plate.

In the examination of springs, especially those used
for domestic supplies at country houses, the authors have
found that the bacteriological examination offers a

224 ELEMENTS OF WATER BACTERIOLOGY

more delicate and more certain index of the quality
than may be obtained by chemical analysis. In a
number of instances, springs located in pastures have
become slightly polluted by animals, but to so small
an extent that the chemical examination gave no indi-
cation of trouble. The bacteria, however, increased
greatly in number, and colon bacilli could be readily
isolated from 75 per cent of the i-c.c. samples of a
long series used in making the presumptive test. A
single case may suffice as an illustration. This was a
spring located on a hill in Hopkinton, Mass.
The chemical analysis was as follows:

Color None

Turbidity None

Sediment None

Odor (hot) None

Odor (cold) None

Parts per Million.

Total solids 33 . oooo

Loss on ignition 7 . oooo

Fixed residue 26 . oooo

Hardness 1 1 . oooo

Chlorine 10 . oooo

Nitrogen as —

Albuminoid ammonia o . oooo

Free ammonia o . oooo

Nitrites o . oooo

Nitrates o . oooo

The bacteriological examination showed a total count
of 375 bacteria per c.c. and a 37° count of 350 per c.c.
The presumptive tests for Bacillus coli showed that
gas-producing organisms were present in a majority
of i -c.c. samples, and typical colon bacilli were isolated.
In this case the contamination was brought about by
cattle gaining access to the area immediately surround-

BACTERIOLOGICAL EXAMINATION 225

ing the spring; but the same conditions might easily
have led to infection from human beings.

Fromme (1910) cites several interesting examples
of temporary pollution detectable only by bacteri-
ological tests. The most striking case was that of an
artesian well. Its average bacterial content had been
38 per c.c. and colon bacilli were absent from 200 c.c.
In May, 1908, this well became polluted from a broken
stable drain 10 meters away. The number of bacteria
rose to 4370 and colon bacilli were found in 10 c.c. sam-
ples. The source of pollution was removed, but the
well water in July still contained 7100 bacteria and B.
coli in i c.c. In September the number had fallen to
105 and colon bacilli were present in 200 c.c. In Novem-
ber the bacteria numbered 120 and colon bacilli were
absent from 200 c.c. At no time did chemical tests
give any indication of danger, while the bacteriological
data obviously measured very delicately a comparatively
slight but real pollution and its gradual disappearance.

Similar results have been reported by Savage and
Bulstrode (Savage, 1906) in the examination of the
water-supply of Bridgend.

It seems to the writers that the real application of
chemistry begins where that of bacteriology ends. When
pollution is so gross that its existence is obvious and
only its amount needs to be determined, the bacteri-
ological tests will not serve, on account of their exces-
sive delicacy. In studying the heavy pollution of small
streams, the treatment of trades wastes, and the
purification of sewage, the relations of nitrogenous
compounds and of oxygen compounds are of prime

226 ELEMENTS OF WATER BACTERIOLOGY

importance. In other words, when pollution is to be
avoided, because the decomposition of chemical sub-
stances causes a nuisance, it must be studied by chem-
ical methods. When the danger is sanitary and comes
only from the presence of bacteria, bacteriological
methods furnish the best index of pollution.

In the study of certain special problems the para-
mount importance of bacteriology is generally recognized.
The distribution of sewage in large bodies of water
into which it has been discharged may thus best be
traced on account of the ready response of the bacterial
counts to slight proportions of sewage, particularly
since the ease and rapidity with which the technique
of plating can be carried out make it possible to examine
a large series of samples with a minimum of time and
trouble. The course of the sewage carried out by the
tide from the outlet of the South Metropolitan Dis-
trict of Boston was studied in this way by E. P. Osgood
in 1897, and mapped out by its high bacterial content
with greater accuracy than could be attained by any
other method. Some very remarkable facts have
been developed by similar studies as to the persistence
of separate streams of water in immediate contact
with each other. Heider showed that the sewage of
Vienna, after its discharge into the Danube River,
flowed along the right bank of the stream, preserving
its own bacterial characteristics and not mixing per-
fectly with the water of the river for a distance of
more than 24 miles (Heider, 1893). Jordan (Jordan,
1900), in studying the self -purification of the sewage
discharged from the great Chicago drainage canal,

BACTERIOLOGICAL EXAMINATION 227

found by bacteriological analyses that the Des Plaines
and the Kankakee Rivers could both be distinguished
flowing along in the bed of the Illinois, the two streams
being in contact, yet each maintaining its own indi-
viduality. Finally, the quickness with which slight
changes in the character of a water are marked by
fluctuations in bacterial numbers renders the bacteri-
ological methods invaluable for the daily supervision
of surface supplies or of the effluents from municipal
nitration plants.

In the commoner case, when normal values obtained
by such routine analyses are not at hand, the problem
of the interpretation of any sanitary analysis is a more
difficult one. The conditions which surround a source
of water supply may be constantly changing. No en-
gineer can measure the flow of a stream in July and
deduce the amount of water which will pass in February;
yet the July gauging has its own value and significance,
so a single analysis of any sort is not sufficient for all
past and future time. If it gives a correct picture of
the hygienic condition of the water at the moment
of examination it has fulfilled its task, and this the
bacteriological analysis can do. The evidence fur-
nished by inspection and by chemical analysis should
be sought for and welcomed whenever it can be obtained,
yet we are of the opinion that, on account of their
directness, their delicacy, and their certainty, the
bacteriological methods should least of all be omitted.

CHAPTER XI

BACTERIOLOGY OF SEWAGE AND SEWAGE EFFLUENTS

Bacteriological and Chemical Examination of Sewage.

The first object of modern sewage disposal is the oxida-
tion of putrescible organic matter. Chemical, rather
than bacterial, purification is usually the prime requisite;
and chemical tests therefore serve best as criteria of
the results obtained. Bacteria are the agents in the
process of sewage purification; but the most generally
useful measure of the work accomplished is the chemical
oxidation attained. " To employ a simile, it is a case
of the saw and the 2 -foot rule — the saw will do the
cutting, but the rule will measure the work cut."
(W. J. Didbin.)

In certain cases, however, bacterial as well as chemical
purity must be effected, in view of special local require-
ments. The sewage from a contagious disease hospital,
for example, should be freed from infectious material as
a factor of safety. Sewage discharged into a body of
water adapted for bathing may well be so treated as to
protect those using the water. In the case of seaboard
cities where sewage effluents are likely to contaminate
oyster beds and other layings of edible shellfish the
problem assumes great importance. Where bacterially
impure effluents are discharged into streams used for

228

BACTERIOLOGY OF SEWAGE 229

sources of water-supply the town taking water may
protect itself by nitration. It should so protect itself,
at any rate, from the pollution necessarily incident to
surface waters; and, unless the bacterial condition of a
stream or lake is made very materially worse by the
discharge of sewage effluents, it is fair that the respon-
sibility of purification should rest on the water works,
rather than on the sewage purification plant. Shell-
fish, on the other hand, cannot be purified. Either
pollution must be prevented, or the industry abandoned.
Under such circumstances sanitary authorities may
rightly demand, as they have demanded at Baltimore,
that bacteria, as well as putrescible organic matter,
shall be removed in sewage treatment. Under such
circumstances the bacterial control of purification
plants is as essential as in the case of water filters.
Methods of Bacteriological Examination of Sewage and
Effluents. In England, considerable attention has
been devoted to this subject, and numerous methods
have been recommended as furnishing valuable criteria
of the bacterial quality of sewage effluents. Houston
(i902b), for example, suggests various tests involving
the use of litmus milk, peptone solution, gelatin tubes,
and neutral-red broth, as well as the inoculation of
animals. He considers the determination of the num-
bers of B. coli and B. sporogenes as of greatest moment,
while the identification of streptococci is of value in
certain cases and the enumeration of liquefying bacteria,
spore-forming aerobes, thermophilic bacteria, and hydro-
gen sulphide producing bacteria is of subsidiary impor-
tance. Rideal (1906) has recently recommended a some-

230 ELEMENTS OF WATER BACTERIOLOGY

what less extensive series of tests, including aerobic
and anaerobic counts, both at 20 and 37°, with the
determination of the number of liquefiers and the num-
ber of spore-formers. The results attained do not
seem to warrant any such elaborate procedure. As
far as the authors are aware, the determination of
liquefying bacteria, anaerobic bacteria and thermophilic
bacteria does not add any information of material
importance to that obtained from the total count.
Some test for specific sewage organisms is of course
desirable. Here again, however, the determination
of B. sporogenes and sewage streptococci tells the
observer little more than can be learned from the routine
use of the colon test. In the United States the practise
of sewage bacteriologists is crystallizing around the
total count and the estimation of B. coli. In the absence
of evidence as to the specific value of other data, the
routine control of filter plants may well be limited
to these two determinations.

The total count of bacteria should be made, as in
the case of waters, at 20°. Determinations carried
out in duplicate at 37° give additional information of
considerable value. The ratio of the 37° count to the
20° count varies with different sewages. At Boston
the body temperature count is 70 to 80 per cent of
the total count; at Lawrence it appears to be propor-
tionately much lower (Gage, 1906). In using either
medium, it is well to add lactose and litmus and note
the number of red colonies, as a check on the enumera-
tion of B. coli.

It should be borne in mind, as Lederer and Bach-

BACTERIOLOGY OF SEWAGE . 231

mann (1911) have recently pointed out, that the sampling
error is a very serious one with sewage. Duplicate
tests made at i-minute intervals for a period of 10
minutes in their experiments gave extreme values of
190,000 and 550,000 per c.c.

The determination of the number of colon bacilli
in sewage and effluents should furnish an integral part
of bacteriological sewage analysis, since it is important
to know whether the decrease of intestinal bacteria in
the process of purification is proportional to the reduc-
tion of total bacteria. The State Sewerage Commis-
sion of New Jersey has adopted this procedure in its
supervision of the disposal plants in that State; and
the results seem amply commensurate with the labor
involved. As in the case of polluted waters the enumera-
tion of B. coli may be carried out, either by the study
of the red colonies which appear on litmus-lactose-agar
plates inoculated with the sample directly, or by the
use of a preliminary enrichment process. The com-
plete identification of B. coli seems unnecessarily
tedious, however, where the organisms are present
in such abundance. Some approximate presumptive
method is indicated here, if anywhere; and the
experience with polluted water, reviewed in Chapter
VI, points to the Jackson bile medium as the most
promising one. Experience at the Sewage Experiment
Station of the Massachusetts Institute of Technology
has shown that this presumptive test in general yields
good results. As pointed out above, a 48-hour incu-
bation period at 37° is required. All tubes showing
20 per cent gas at the end of this time may be con-

232 ELEMENTS OF WATER BACTERIOLOGY

sidered positive tests for the colon group, without
serious error.

Numbers of Bacteria in Sewage. The total number
of bacteria and the number of colon bacilli naturally
vary widely in the sewages of different cities and towns.
European sewages, being more concentrated, show
as a rule higher numbers than are found in America.
Results compiled from various sources show from
1,000,000 to 5,000,000 bacteria in the sewages of Essen,
Berlin, Charlottenburg, Leeds, Exeter, Chorley, and
Oxford, 2,000,000 to 10,000,000 in the sewages of Lon-
don, Walton, and W. Derby and over 10,000,000 in the
sewages of Paris, Ballater and Belfast (Winslow, 1905).
The number of colon bacilli in English sewages varies
from 50,000 to 750,000. In American sewages, on the
other hand, bacteria are somewhat less numerous.
At Lawrence the determinations made from 1894 to
1901 showed on the average 2,800,000 bacteria per c.c.
At Worcester, Eddy reported 3,712,000 in 1901 (Eddy,
1902); at Ames, Iowa, Walker (1901) found 1,248,256
in the same year. At Columbus, Johnson (1905)
reports an average of 3,600,000 bacteria per c.c.; the
individual numbers varied from 320,000 to 27,000,000.
The number of colon bacilli varied from 50,000 to
1,000,000 and averaged 500,000. Day samples of
Boston sewage collected three times a week, from
October, 1906, to April, 1907, showed an average of
1,200,000 bacteria per c.c. In the summer months
numbers are notably higher than at other seasons
in many sewages. Thus in 1903, Boston sewage con-
tained 2,995,000 bacteria in July, 4,263,600 in August,

BACTERIOLOGY OF SEWAGE 233

11,487,500 in September, 3,693,000 in October, 587,100
in November, and 712,000 in December (Winslow,
1905). There is also a marked diurnal variation
in the bacterial content of sewage, since the flow con-
tains a smaller proportion of intestinal matter at night
than at other times. For example, a series of hourly
samples at the Sewage Experiment Station of the
Massachusetts Institute of Technology showed the
following results:

BACTERIA IN BOSTON SEWAGE— AVERAGES FOR EACH
FOUR-HOUR PERIOD. AUGUST 13-14, 1903

(WINSLOW AND PHELPS, 1905)

Period.

Bacteria per c.c.

7:30-11:30 A.M

1 1 :3O A.M.-3 :3o P.M

I,8oo,OOO
3,2OO,OOO

3:30-7:30 P.M.

4 600 ooo

7 :3o-i i :3O P.M
ii '30 P M -3 '30 A.M

3,500,000

I OOO OOO

3 :3o— 7 :3O A.M

400,000

It is evident that many published results of bacterial
examinations of sewage are in excess of the average
values, since they refer in most cases to day samples
only.

Bacterial Content of Sewage Effluents. The bacterial
content of sewage effluents varies widely according to
the process of purification adopted and the efficiency
of the particular plant. The only process which
yields a notably purified effluent from the bacteri-
ological standpoint is that of filtration through sand.
Processes of this type when operated with care may
give a bacterial purification well over 99 per cent as

234 ELEMENTS OF WATER BACTERIOLOGY

shown by bacteriological examinations at the Brockton
(Mass.) filters, reported by Kinnicutt, Winslow and
Pratt (1910) as follows:

BACTERIA IN SEWAGE AND EFFLUENTS AT BROCKTON,
AVERAGE OF FOUR EXAMINATIONS, AUTUMN OF
1908

Bacteria per c.c.
Gelatin 20°.

Colon Bacilli per
c.c. Lactose Bile.

Sewage

3,1 ^O.OOO

1 50 ooo

Effluent A

1,900

4.OO

6 300

j r

I2Z

1,400

•? OOO

Such high efficiencies as this table indicates are
often not realized under the actual working condi-
tions of a municipal plant. At Vineland, N. J., for
example, the intermittent niters show a reduction
of 90 to 95 per cent in total bacteria and a somewhat
higher reduction of B. coli. The results of three
examinations made in 1906 are given below.

BACTERIA IN SEWAGE AND SAND FILTER EFFLUENT
AT VINELAND, N. J.

(N. J. STATE SEWERAGE COMMISSION, 1907)

Bacteria per c.c.

B. Coli in

Sewage.

Effluent.

Sewage.

Effluent.

March 2

480,000

20,000

O.OOOI C.C.

O.OI C.C.

July 26

496,000

. 6l,000

O.OOOI C.C.

O.OOI C.C.

July 26

511,000

38,000

O.OOOOI C.C.

O.OOI C.C.

BACTERIOLOGY OF SEWAGE

235

The newer bacterial processes, contact beds, and
trickling niters naturally show a much less satisfactory
bacterial removal than sand nitration beds. In the
Columbus experiments, Johnson (1905) found from
1,000,000 to 2,000,000 bacteria in the effluents of con-
tact beds and from 750,000 to 1,900,000 in the effluent
from trickling niters.

At the experiment station of La Madeleine, in Lille,
Calmette (1907), reports 5,000,000 bacteria per c.c. in
the crude sewage, 2,900,000 in the second contact
effluent and 800,000 in the effluent from the trickling
bed. Of 20,000 B. coli per c.c. applied to the filters,
the contact system delivered 4000 and the trickling
bed 2000 per c.c. The average results of examinations
made three times a week at the Sewage Experiment
Station of the Massachusetts Institute of Technology,
during two different periods, were as follows:

BACTERIA IN SEWAGE, SEPTIC EFFLUENT AND
TRICKLING EFFLUENT AT BOSTON

(WINSLOW AND PHELPS, 1907)

Bacteria per c.c.

B. Coli.
Positive Tests
in o.oooooi

C.C.*

July-Sept., 1906.

Oct., I9o6-April,
1907.

July-Sept.,
1906.

No.

Per Cent
Reduc-
tion.

No.

Per Cent
Reduc-
tion.

Per Cent.

Sewage
Septic effluent. . .
Effluent from

1,300,000
1,650,000

1,200,000
750,000

65
66

Inc.

trickling bed . .
Septic tank and
trickling bed . .

750,000
750,000

42
42

200,000
180,000

35
35

Jackson bile test.

236 ELEMENTS OF WATER BACTERIOLOGY

The following average data for two of the largest
trickling filter plants in the United States are cited
by Kinnicutt, Winslow and Pratt (1910).

BACTERIAL CONTENT OF SEWAGE AND EFFLUENTS
FROM TRICKLING FILTERS

Place.

Period.

Bacteria per c.c.

Screened
Sewage.

Septic
Effluent.

Filter
Effluent.

Reading, Pa
Columbus, Ohio. . .

1908-1909
1909

3,100,000

2,370,000

I,8oo,OOO
1 ,050,000

600.0OO
560,000

It is obvious that effluents of this character cannot be
considered satisfactory from the standpoint of bacterial
purification. As Houston concluded, after a careful
review of the subject, " The different kinds of bacteria
and their relative abundance appear to be very much
the same in the effluents as in the crude sewage. Thus,
as regards undesirable bacteria, the effluents frequently
contain nearly as many B. coli, proteus-like germs,
spores of B. enteritidis sporogenes and streptococci,
as crude sewage. In no case, seemingly, has the reduc-
tion of these objectionable bacteria been so marked
as to be very material from the point of view of the
epidemiologist" (Houston, 1902^.

Experimental studies with specific bacteria have
confirmed these conclusions. Houston (igo4b) found
that B. pyocyaneus appeared in the effluent of a trickling
bed 10 minutes after application to the top and con-
tinued to be discharged for 10 days. In septic tanks
and contact beds, the same germ persisted for 10 days.

BACTERIOLOGY OF SEWAGE 237

Rideal (1906) quotes experiments by Pickard at Exeter,
which show that typhoid bacilli may persist for 2 weeks
in a septic tank and that contact bed treatment only
effects a 90 per cent removal of these organisms.

Disinfection of Sewage Effluents. Where bacterial
purity is required, some special process of disinfection
must be combined with the contact bed or the trickling
filter. For this purpose treatment with chloride of
lime or other chemicals is rapidly gaining ground as an
important adjunct to bacterial disposal plants; and in
connection with this process bacteriological control is
an essential.

Rideal (1906) first showed at Guildford that 30 parts
of available chlorine per million would reduce the
number of bacteria in crude sewage from several mil-
lions to 50,000, while 50 parts would reduce their
number to 20 per c.c. Colon bacilli were reduced from
one million per c.c. to less than one per c.c. by 30 parts
of chlorine. In septic effluent 25 to 44 parts of chlorine
per million reduced B. coli from two and a half to four
and a half million per c.c. to less than one per c.c.
With contact effluents smaller amounts of chlorine
proved efficient, The primary effluent required 20
parts per million, the secondary effluent 10.6 parts
per million and the tertiary effluent 2.5 parts per mil-
lion to reduce the number of B. coli so that this organism
could not be isolated in 5 c.c.

In this country Phelps and Carpenter (1906) demon-
strated the practical usefulness of bleaching powder
disinfection, at the Sewage Experiment Station of the
Massachusetts Institute of Technology. As indicated

238 ELEMENTS OF WATER BACTERIOLOGY

in the table below smaller amounts of chlorine than
were used by Rideal will give good results with more
dilute American sewages.

BACTERIA IN TRICKLING FILTER EFFLUENT BEFORE
AND AFTER TREATMENT WITH CHLORIDE OF LIME
(5 PARTS PER MILLION AVAILABLE CHLORINE)

(PHELPS AND CARPENTER, 1906)

T-Jai_

Bacteria per c.c.

B. Coli, Jackson Bile Test.

Before.

After.

Before

O.OOOOOI C.C.

After

I.O C.C.

IQ06

August 1 1 ....

270,000

+ o

13....

630,000

O 0

+ o

' 14

135,000

406

+ +

+ o

' IS--.-

230,000

o o

0 O

16....

250,000

+ o

0 0

HO,OOO

0 O

+ o

20....

90,000

+ o

0 0

21....

220,000

0 O

0 0

23. .

+ o

0 0

Average

24O,OOO

33% +

22% +

Average

removal ....

99.96%

99-993%

The success of chemical disinfection varies with the
character of the sewage or effluent treated, since the
organic matter present consumes a certain amount
of the disinfectant and renders it inoperative. Dis-
cordant results are therefore reported from different
sources.

An important series of experiments tarried out in
Ohio by Kellerman, Pratt, and Kimberly (1907)
showed good results with sand filter effluents and

BACTERIOLOGY OF SEWAGE 239

contact effluents. Septic sewage, on the other hand,
required large amounts of chlorine to produce a rea-
sonable bacterial reduction. The table on page 240
shows the results obtained at Marion, Ohio.

In Germany, on the other hand, Schumacher (1905),
Kranepuhl (1907), and Kurpjuweit (1907) found larger
amounts of chlorine necessary, in the neighborhood of
60 parts per million parts of sewage. Their tests
were somewhat severe, however, the criterion of success
being the absence of B. coli in a large proportion of
liter samples.

Standards for Sewage Effluents. The science of sew-
age bacteriology is in its infancy; and it is difficult to
give any general rules for the interpretation of bac-
teriological examinations designed to indicate whether
disposal plants are successful or not. Houston stated
provisionally that the 20° count should be under 100,000
and the 37° count under 10,000, while B. coli should
be absent from .001 c.c. and B. sporogenes from .1 c.c.
(Houston, i902b). This standard seems to us far too
lenient. Either organic purity alone is necessary, as
at many sewage disposal plants, or a higher grade of
purity than this should be attained. It seems wisest
at the present time to avoid fixing any general standards
of purity for sewage effluents. Each case should
be judged intelligently on its own merits. In general,
however, where bacterial purification is indicated at all,
it seems fair to demand that the effluent should be
of such a quality as not to increase materially the
bacterial content of the body of water into which it
is discharged.

240 ELEMENTS OF WATER BACTERIOLOGY

BACTERIA IN SEPTIC EFFLUENT, CONTACT EFFLUENT,
AND SAND EFFLUENT AT MARION , O., BEFORE AND
AFTER TREATMENT WITH CALCIUM HYPOCHLORITE

(KELLERMAN, PRATT, AND KIMBERLY, 1907)

Date.

Effluent.

Average
Available
Chlorine.
Parts per
Million.

Bacteria per c.c.

20° C.

37° C. Total Count.

Untreated.

Treated.

Untreated.

Treated.

1907
Apr. ii
Apr. 12
Apr. 15
Apr. 28
Apr. 29
Apr. 30
Mar. 21
Mar. 22
Mar. 26

Septic
Septic
Septic
Contact
Contact
Contact
Sand
Sand
Sand

4-3

6.2

7-6
2.9
5-0
4.4
3-8
3-o
i-S

850,000
4,400,000
6oo,OOO
IIO,OOO
65,000
500,000
49,OOO
56,000
70,000

I,IOO,OOO
550,000
400,000
2,500
1, 600
800

570
140
4,OOO

1,200,000
850,000
450,000

240,000
260,000
190,000

73,000
160,000
9,800
7,000
20,000

370
400

ISO
60
1 60

Date.

Effluent.

Average
Available
Chlorine.
Parts per
Million.

Bacteria per c.c.

37° C. Red Colonies.

B. Coli.

Untreated.

Treated.

Untreated.

Treated.

1907
Apr. ii
Apr. 12
Apr. 15
Apr. 28
Apr. 29
Apr. 30
Mar. 21
Mar. 22
Mar. 26

Septic
Septic
Septic
Contact
Contact
Contact
Sand
Sand
Sand

4-3
6.2

7.6
2.9

5-o
4-4
3-8
3-0
i-5

55,000
6o,OOO
IOO,OOO

7,400
15,000
51,000

20,000
I5,OOO
2O,OOO

I,OOO

2,000

Not in o . 5
" 0.5

" I.O

" I.O
" I.O

In i .0

IO,OOO

21,000
1,300
800

4,000

0
O

BACTEEIOLOGY OF SEWAGE 241

Bacteriology of the Sewage Filters Themselves. Before
leaving the subject of sewage bacteriology, brief
reference must be made to the importance of bacteri-
ological studies in relation to the processes of sewage
purification which bring about the removal of the
organic matter itself. Nothing is more necessary to the
development of the present art of sewage disposal
than knowledge of the micro-organisms concerned and
of the conditions which favor their activity; but such
knowledge is woefully deficient. Something is known
of the nitrifying organisms long ago discovered by
Winogradsky. More recent work, like that of Schultz-
Schultzenstein (1903), Boullanger and Massol (1903)
and Calmette (1905), has cleared up many points
concerning these forms; but much remains to be done.
In regard to the reducing action of bacteria in the
septic tank and contact bed we are almost wholly in
the dark. Septic tanks work well with some sewages
and badly with others; and the presence or absence
of the right bacteria is probably largely responsible
for the different results. In some cases, as at
Plainfield, N. J., the seeding of a tank with cesspool
contents has produced a material improvement in
septic action.

Knowledge of the kinds of bacteria involved would
make it possible to substitute scientific control for
such empiricism and might well lead to improved
methods of a more intensive character than are yet
available. The work already done upon a laboratory
scale furnishes promise of such results. The student

242 ELEMENTS OF WATER BACTEEIOLOGY

who wishes to follow out this line of investigation
will find a good summary of what is already known
of the hydrolysis and denitrification of nitrogenous
bodies and the decomposition of cellulose and other
carbohydrates in Rideal's " Sewage and the Bacterial
Purification of Sewage " (1906).

Gage (1905) has made a suggestive study of the
bacteria which carry on the reducing changes in sewage
which deserves the study of all who are interested in
the more theoretical aspects of sewage treatment.
His method consisted in plating sewages and effluents,
isolating typical cultures and determining their power
to decompose peptone and nitrates with the produc-
tion of ammonia and free nitrogen. The rate of gelatin
liquefaction, the amount of nitrate reduced, the amount
of free ammonia formed, and the amount of nitrogen
liberated were quantitatively determined for each culture
thus isolated. The numerical values obtained, multiplied
by the number of bacteria, apparently of the same type,
observed in the plates, gave coefficients of the liquefying,
denitrifying, ammonifying, and nitrogen-liberating power
of the effluent; and these coefficients may be considered
as measures for a given sample of the tendency of the
bacterial flora to set up certain changes. The results
of further studies made by Clark and Gage (1905),
on sewages and on sand, contact, and trickling effluents,
show that there may be important differences between
various sewages in this respect which must render
their purification more or less easy. They indicate
that the effluents obtained from intermittent sand

BACTERIOLOGY OF SEWAGE 243

filters in cold weather contain larger numbers of ammo-
nifying and denitrifying bacteria than appear at other
seasons, which may help to explain the poorly nitrified
effluents obtained in the winter season. Along these
lines research work in sewage bacteriology promises
to be fruitful of results.

CHAPTER XII

BACTERIOLOGICAL EXAMINATION OF SHELLFISH

Shellfish and Disease. The pollution of areas
devoted to the growing of shellfish and the consequent
pollution of the shellfish themselves is a matter of
much sanitary importance. Oysters, clams and mussels
are the shellfish commonly used as food, and since
they are likely to be eaten in an uncooked or partially
cooked condition, it is important to be assured as to
their character from the bacteriological standpoint.
In their normal habitats, in clean sea-water, or in river
estuaries free from pollution, shellfish are unquestionably
free from dangerous bacteria, although their feeding
habits make it probable that the types of bacteria
indigenous to the waters in which they are found might
be present in considerable numbers. With the pollu-
tion of streams by unpurified sewage the areas in \vhich
oysters and clams develop may easily become infected
by organisms of intestinal types, and there is, therefore,
offered an easy means for the typhoid bacillus and other
pathogenes to pass from the sewage directly into the
intestinal tract of the consumer of the raw oysters or
clams.

The history of this subject is well summarized by
Newlands and Ham (1910), from whose excellent report
the following paragraphs are adapted:

244

EXAMINATION OF SHELLFISH 245

Attention was first drawn to the danger from shell-
fish by the remarkable outbreak of typhoid fever which
occurred in Middletown, Conn., in 1894, as a result
of the serving of raw oysters at college fraternity
banquets. The oysters used in this case were all
derived from a certain portion of Long Island Sound,
where they had been put down, or planted, in order
to fatten. Investigation showed that the stream
entering the Sound at this point was highly polluted,
and furthermore, that at a nearby house there were
two severe cases of typhoid fever from which the intes-
tinal discharges were turned into the drain and thence
into the stream without disinfection. The course
of the passage of the bacteria from the patient suffer-
ing with the disease to the oyster and so on to the
young men at the banquets was, therefore, traced
out in a most complete and thorough \vay. This
investigation, which was conducted by Prof. H. W.
Conn, of Wesleyan University, caused immediate invest-
igations to be set on foot in England and in this coun-
try. Two years later there followed a report by the
Local Government Board of Great Britain dealing
with pollution of shellfish along the English coast, and
the matter has also received much attention in this
country.

A study of the literature reveals only a few references
to oysters as carriers of disease germs previous to
1880. In that year Cameron, in a paper entitled
" Oysters and Typhoid Fever," read before the British
Medical Association, suggested that outbreaks of typhoid
fever and cholera might be caused by eating oysters.

246 ELEMENTS OF WATER BACTERIOLOGY

In 1893 Thorne-Thorne, in a report to the Local Gov-
ernment Board, wrote that, in his opinion, certain
cases of cholera which had occurred that year at various
inland towns in England were due to eating contam-
inated oysters from beds at Grimsby, where there
had been a small cholera epidemic. Following ' the
suggestions embodied in this report the English Govern-
ment began a series of investigations which have made
many important additions to our present knowledge
of the subject.

In 1902 the famous oyster epidemics at Winchester
and Southampton, England, were proven beyond
reasonable doubt to have been caused by contami-
nated oysters taken from grounds at Emsworth. Here
again we have to deal with banquets given in different
cities where the only common source of infection
appears to have been contaminated oysters. Of the
267 guests at these banquets 118 were attacked with
intestinal disorders and 21 cases of typhoid fever
developed, 5 of which were fatal.

Although a great many sensational attacks have
been made against oysters as carriers of disease germs
which have been based on little or no evidence, the
above-mentioned investigations and others, among
which might be mentioned those of Thresh, Marvel,
and Soper, have brought out sufficient trustworthy
evidence to show that contaminated oysters must be
considered as a real factor in the dissemination of
typhoid fever and other water-borne diseases. An esti-
mate of the extent to which such illness is due to
oysters would be impossible at the present time. The

EXAMINATION OF SHELLFISH 247

Royal Sewage Commission after an extensive investi-
gation on this subject came to the following conclusion :
" After carefully considering the whole of the evidence
on this point, we are satisfied that a considerable num-
ber of cases of enteric fever and other illness are
caused by the consumption of shellfish which have been
exposed to sewage contamination; but in the present
state of knowledge, we do not think it possible to make
an accurate numerical statement.

" Moreover an examination of the figures which
have been placed before us as regards those towns in
which the subject has been most carefully studied
shows that there may be occasional errors. Indeed
the witnesses themselves recognized that absolutely
accurate figures were not obtainable.

" We are far from denying that isolated cases may
have been due to contaminated shellfish, but we must
remember that the possibility of some of them being
due to other causes cannot be altogether excluded."

In the above-mentioned cases, where oysters have
been proven or reasonably suspected of being the cause
of disease, it was found that the oysters in ques-
tion had been floated or grown in heavily polluted
water where direct contact with specific infection
could be proven or readily assumed. The Wesleyan
epidemic is a case in point. Oysters had undoubtedly
been floated in the contaminated waters at Fair Haven
for a number of years previous to 1894 without any
noticeable effect on the health of persons eating them,
but specific infection of the water from two patients
in a house near by was followed by a serious epidemic.

248 ELEMENTS OF WATER BACTERIOLOGY

Valuable studies of the relation between shellfish
and disease have recently been published by Bulstrode
(1911) and Wilhelmi (1911) and Stiles (1912).

Effect of Cookery upon Polluted Shellfish. It should
be noted that it is unfortunately not only raw shellfish
which are responsible for the spread of disease. Most
of the processes of cookery to which these foods are
subjected are insufficient to destroy pathogenic germs.
Clark (1906) found that clams and oysters in stews
and fried and scalloped in the usual manner were
generally free from colon bacilli and streptococci.
With steamed clams, however, the bacteria present
could not be destroyed except by a temperature high
enough and prolonged enough to ruin the clams for
eating. Rickards (1907) confirmed these results as
to the danger from steamed clams, while he found fried
clams and clams in chowder and scalloped oysters to
be practically sterilized. Oyster stew, however, is
not exposed to long continued heat as is clam chowder,
and fried oysters are less thoroughly heated than
fried clams in the ordinary processes in use. Oysters
in both of these forms and fancy roast oysters still
contained colon bacilli and streptococci. Buchan (1910)
finds that the ordinary methods of cooking mussels
do not remove the risk of typhoid infection.

Bacteriological Examination of Shellfish. Without
further discussing the general sanitary aspects of the
subject it is important to consider just how one may
determine whether the oysters from a given region are
polluted or not. The methods which have been
developed for this work are essentially modifications

EXAMINATION OF SHELLFISH 249

of the methods used in water examination, involving
sometimes total counts of bacteria at different tem-
peratures, but especially the application of the various
tests for the determination of the colon bacillus, since
here, as in water examination, this organism may be
taken as an index of pollution and its occurrence in
considerable numbers must be looked upon not merely
with suspicion, but as a practical proof that the
supernatant waters are polluted and that the shell-
fish themselves may contain organisms of pathogenic
importance, such as B. typhi, B. dysenteries, B. sporo-
genes and others. Determinations of the pollution
of the water above the beds are sometimes made
as bearing indirectly and inferentially on the possi-
bility of the pollution of the shellfish contained therein.
Results of the two determinations are not always
in close agreement, however, owing to the rapidly
changing local conditions due to tide, etc. The gen-
eral relations and the individual variations between
water and shellfish determinations are well illus-
trated in the table on page 250 from the report
by Newlands and Ham (1910) on conditions in New
Haven Harbor.

Study of the methods of examination of shellfish
has been conducted with great care at the Lawrence
Experiment Station by Gage, at the Sanitary Research
Laboratory at the Institute of Technology by Phelps,
at Brown University by Gorham, and in New York
by Pease. Other officials of the Shellfish Commissions
of different States have also carried out investigations
upon this subject. The Lawrence Experiment Station

250 ELEMENTS OF WATER BACTERIOLOGY

BACTERIA IN WATER AND SHELLFISH, NEW HAVEN

HARBOR

Water.

Oysters.

Station.

Samples
Taken.

Av. Number
Bacteria per c.c.

Average
Number
B. Coli *
per c.c.

Number
Oyster
Samples

Character,
of Bottom.

37° C.

20° C.

Ferry St.

Bridge

2 IO

I26O

Soft

Tomlin-

son Br

9IO

2650

« *

No. i ...

510

1680

No. 2

375

9IO

1 <

No. 3. . .

255

835

Buoy 10

155

450

' *

No. 4. . !

160

1720

••

No. 5. . .

615

1340

308

Buoys. •

3iS

715

* *

Buoy 8. .

205

4IO

• '

No. 6. . .

145

485

"o"

Seaweed

No. 7. . .

215

74°

425

Hard

No. 96. .

220

260

* '

No. QA. .

100

185

No. 9. . .

195

200

1 *

No. 7 A .

I2O

240

255

* '

No. 8. . .

180

270

370

No. 10. .

300

615

100

Soft

No. ii ..

405

5io

* *

Buoy 6. .

815

1690

. „ . .

Buoy 3 . .

175

590

291

Hard

No. 12. .

620

1190

* *

No. 13. .

240

I2O

No. 14. .

285

I 100

i —

' '

Buoy 4. .

375

I4OO

No. 15- .

455

1680

Hard

No. 16. .

280

1025

—

i —

Soft

No. 17. .

300

I26O

Hard

No. 19. .

800

300

No. 20. .

135

860

Hard

Buoy 2. .

375

90S

No. 22. .

305

560

—

7-3

Hard

Buoy i . .

995

—

No. 18. .

255

675

—

No. 24. .

710

1340

Hard

No. 23 . .

450

240

—

' *

No. 25 . .

130

IOOO

—

No. 26. .

130

465

—

No. 27. .

630

695

Soft

No. 28. .

415

1400

—

Mud and sand

No. 29. .

370

1700

—

Soft

No. 30. .

185

440

—

Hard

No. 31. -

320

130

—

* *

No. 32. .

1050

—

No. 33-.

485

405

—

* *

No. 34- -

535

495

—

' '

No. 35- •

120

270

Sticky

* Jackson's lactose bile presumptive test used.

Minus sign after figure i indicates that the average was less than i.

EXAMINATION OF SHELLFISH 251

method was published in the Massachusetts State
Board of Health Report for 1905 (Clark, 1906). This
method consisted in the total counts of bacteria de-
veloping at 20° and 37° and the fermentation reaction
in dextrose broth. Experience indicated that it was
not merely necessary to examine the stomach contents
of the oysters but the " shell water " as well was sub-
jected to examination. With the advent of lactose bile
as a better medium for the development of B. coli
without interference with other types of bacteria, the
substitution of this medium for dextrose broth was
commonly made, and this is now one of the standard
media employed for the determination.

It has been noted that the superiority of lactose
bile to dextrose broth is greatest in water examinations
when the water is most polluted. In the study of
shellfish the danger of overgrowths is even greater
than in polluted waters, since the organic matter in
the oyster and its surrounding shell water furnishes a
culture medium for many bacteria. Streptococci are
particularly abundant. As pointed out in Chapter
IX, streptococci die out more rapidly than colon bacilli
in potable waters, but where organic matter is present
in abundance the former may survive the latter.

We have compiled the table from results given
by Clark (1906). It will be noted that in all cases
except in that of the shell water there is a consider-
able difference between the dextrose fermentation tests
and the colon isolations, indicating an overgrowth by
streptococci and other forms, of colon bacilli originally
present. The B. sporogenes is also very frequently

252 ELEMENTS OF WATER BACTERIOLOGY

responsible for such anomalous results in shellfish
examinations.

COLON BACILLI AND STREPTOCOCCI IN DIFFERENT
PORTIONS OF THIRTY CLAMS

Per Cent of Samples Showing

Fermenta-
tion in
Dextrose

B. Coli.

Strepto-
cocci.

B. Coli and
Strepto-

Broth.

Shell water

Gills

I ;;

Stomach (intestine).. . .

Rectum (intestine)

Liver

Visceral tissue

It will be noted from Clark's table above that the
shell liquor is not only freer from overgrowths than the
portions of the body of the clam, but that the propor-
tion of positive reactions is in each case higher. Since
the shell water is of course easier to examine than the
macerated animal, this is now generally adopted as the
standard material for examination.

Self -purification of Shellfish. In connection with
the bacteriological examination of shellfish for colon
bacilli certain investigations have been carried out
which are of great importance from the commercial
as well as from the sanitary standpoint. Phelps (1911)
has shown that oysters which develop in waters sub-
ject to sewage pollution may be purified or entirely
freed from colon bacilli by the removal of the oysters
themselves to waters of purer character, when, after

EXAMINATION OF SHELLFISH 253

sufficient time has elapsed, the oysters will have cleansed
themselves through their metabolic processes and
become entirely safe even for consumption in the raw
state. It is of considerable importance to determine
the length of time necessary for this self-purification
to take place. Obviously, from the commercial stand-
point it is desirable to make it as short as possible,
while from the sanitary standpoint it must be long
enough to insure a thorough and satisfactory removal
of all traces of polluted matter. Oyster beds which
are free from pollution or which are sufficiently good
for the re-laying for polluted oysters are difficult to
find and limited in areas because of their nearness to
sources of pollution. The investigations in question
were conducted by Phelps in the Providence River
and the upper part of Narragansett Bay. The oysters
were removed from heavily polluted regions and car-
ried to waters which were practically free from pollu-
tion, where they were planted. Examinations were
made from day to day in order to determine the length
of time that these particular oysters showed pollution
and it was found that within 4 days the organisms
of the colon type were practically all eliminated.

It must be borne in mind that, if shellfish are care-
lessly opened and handled, they may suffer a considerable
additional pollution in the process, and may therefore
be much worse instead of better than when they were
taken. This is well brought out by the table on page
254, taken from a report by Stiles (1911) in which
shucked market oysters show much worse pollution
than market oysters in the shell.

254 ELEMENTS OF WATER BACTERIOLOGY

COMPARATIVE BACTERIOLOGICAL CONDITION OF MAR-
KET OYSTERS, SHUCKED AND IN THE SHELL

(STILES, 191 1).

Number
of
Samples.

Average per c.c. Liquor.

Bacteria.

B. Coli.

Strepto-
cocci.

Plain Agar.

Bile

Salt
Agar.

25°

37°

Shell oysters
Shucked oysters . .

36
33

6,000
867,000

I,OOO

268,000

2OO

45,000

7
74,000

8000

Seasonal Variation of Bacteria in Oysters. It has
been observed by Gorham (1912) and others that the
examination of oysters from certain regions made in
the' summer failed to agree with the similar analyses
from the same beds made in the winter. With the
advent of cold weather there seems to be a great
improvement in the sanitary quality, so that oysters
taken from beds in close proximity to the outfalls of
large sewers show in the colder months entire absence
of any evidence of contamination, judged solely by the
bacteriological data. Thus Gorham found in the sum-
mer of 1910 that all oysters on the beds in the Prov-
idence and Warren Rivers and the upper part of Narra-
gansett Bay were so badly polluted by sewage as to
be unfit for jood. Colon bacilli were found in the
" shell water " of every oyster in amounts as small as
.01 of a cubic centimeter or less. Chemical and
bacteriological examination of the waters over these

EXAMINATION OF SHELLFISH 255

beds showed them to be heavily sewage polluted. In
December of the same year the analyses of the oysters
were strikingly different, although the condition of
the water was apparently unchanged. In the examina-
tion five oysters were selected in each case and the
average total number of bacteria per cubic centimeter
was determined and the presence of colon bacilli was
tested by the bile tube and subsequent isolation and
identification of the organisms. The table on page 256
shows the numbers of bacteria found, and the propor-
tion of the five oyster samples in which colon bacilli
were present in cubic centimeter amounts and also
in o.i and o.oi of a cubic centimeter.

The conclusions arrived at by Gorham are that
during the cold weather the oysters assume a condition
of rest or hibernation, during which time ciliary move-
ment ceases and the process of feeding is suspended.
No organisms are therefore taken in from the outside
water and those inside the oyster are gradually elim-
inated, so that the total number of organisms is
reduced very considerably and the oyster becomes
practically free from colon bacilli.

Standard Methods for the Examination of Shellfish.
The examination of shellfish for pollution is regarded
as of such importance by the American Public Health
Association (1912) that a committee was established
to report upon methods of examination and estima-
tion of the numbers of colon bacilli found. The
following abstract of the second report of this
committee gives the recommendations for standard
methods for bacteriological examination of shellfish

256 ELEMENTS OF WATER BACTERIOLOGY

SEASONAL VARIATION IN THE BACTERIAL CONTENT OF
OYSTERS

(GORHAM, 1912)

_"£ u, "

Proportion of

Five

Date.

till

Oysters Showing
B. Coli in

Score.

B. Coli
Present in
Water in

Tempera-
ture of
Water

> rt •*-! •+.

<; w o o

I C.C.

O.I C.C.

o.oic c.

BED No.

PROVIDENCE RIVER

Dec. 20, 1910

IOOO

0 01 C.C.

-1°

Jan. 14, IQII

750

Jan. 2<. .

O.OI C.C.

1°

Tan 27.

Feb. 10

130

O *

I .O C.C.

0.1°

Feb. 28

140

O.OOOI C C.

1°

Mar. ii

2OO

O OI C.C.

1.75°

April 14

275

O.OI C.C.

-*• i O

8-5°

April 28

700

410

O.OOOI C.C.

12-5°

May 12

I7OO

=;oo

0 0001 C.C.

15°

BED No. 44. PROVIDENCE RIVER

Jan. 7, 191 1..

425

140

0.25°

Feb 10

2<co

0°

Feb. 28

240

o q°

March n. . . .

**f-w

IOO

o
c

• J

2°

April 14

2IO

8 S0

April 28

IOOO

4IO

tj J

ii 7<5°

May 1 2 .....

I IOO

AIO

•*••*••/ 0

14 7^°

BED No. 204. WARREN RIVER

**T " / 0

Jan. 25, 1911

600

5°

0°

Feb. 10

140

o
o

I .O C.C.

0°

Feb. 28

400

O.OI C.C.

0-75°

March 4

7^0

o 7=5°

March n.. . .

/ o

O
0

O.OI C.C.

/ 0

3°

March 14. ...

3400

O.OI C.C.

8-75°

April 28

1050

410

O.OI C.C.

13°

BED No. 205. WARREN RIVER

Dec. 22, 1910

250

-i°

Feb. 10, 1911

325

I .O C.C.

0°

Feb. 28

450

O.OI C.C.

1°

March 4

600

o.75°

March 1 1 . . . .

O.OI C.C.

2°

April 14

325

O.OI C.C,

8.25°

April 28

4000

500

O.OI C.C.

ii-5°

* Only three oysters used.

EXAMINATION OF SHELLFISH 257

which were adopted by the Association at its meet-
ing in 1912:

RECOMMENDATIONS FOR STANDARD METHODS FOR THE
BACTERIOLOGICAL EXAMINATION OF SHELLFISH

Oysters in the Shell

Selection of Sample. Twelve (12) oysters of the average
size of the lot under examination, with deep bowls, short
lips, and shells tightly closed, shall be picked out by hand
and prepared for transportation to the laboratory.

As complete a record of such data as is possible to obtain
shall be made covering the following points :

The exact location of the bed from which the sample
has been selected.

The depth of the water over the bed at time of collection.

The state of the tide.

The direction and velocity of the wind.

Other weather conditions.

The day and hour of the removal of the stock from the
water.

The conditions under which the stock has been kept
since removal from the water and prior to the taking of
the sample.

The day and hour of the taking of the sample.

Transportation of the Sample. The oysters so selected
shall be packed in suitable metal or pasteboard containers
of such size and shape that a number of them can be
enclosed in a shipping case capable of satisfactory refriger-
ation by means of ice. The important points in this con-
nection are:

A. The prevention of the mixing of the oyster liquor
of different samples, and of the mixing of the ice-water
with the oysters.

258 ELEMENTS OF WATER BACTERIOLOGY

B. The icing of the samples, if they are not to arrive
at the point of laboratory examination inside of 36 hours,
or if the outside temperature is above 50° F.

It is not necessary to enclose the oysters in an absolutely
tight container, providing the above conditions are main-
tained.

Condition of Samples. Record shall be made of the
general condition of the oysters when received, especially
whether the shells are open or closed; of the presence of
abnormal odors; and of the temperature of the stock.

Technical Procedure. The bacteriological examination
shall be started as soon as possible after the receipt of the
sample.

The oysters shall be thoroughly cleaned with a stiff
brush and clean running water and then dried. The edges
of the shell shall be passed through the flame or burned
with alcohol.

The opening of the shell shall be accomplished by either
of the following methods:

A . By the use of a sterile oyster knife in the usual manner.

B. By drilling through a flamed portion of the shell
near the hinge with a sterile drill. The drill shall be
sterilized, and the site of the operation on the shell shall
be flamed at least once during the drilling process.

Bacterial Counts. Bacterial counts shall be made of
a composite sample of each lot obtained by mixing the
shell liquor of five oysters. Agar shall be used for the
culture medium and in general the procedure shall be in
accordance with the method recommended for the exam-
ination of water by the Committee on Standard Methods
of Water Analysis of the American Public Health Association.

The water used for dilution purposes shall contain i per
cent of sodium chloride, in order to approximate the natural
salinity of oyster liquor.

The agar plates shall be incubated at 20° C. for three
days and the colonies then counted.

EXAMINATION OF SHELLFISH 259

Determination of Bacteria of the Bacillus Coli Group.
The quantitative determination of the presence of B. coli
shall be in accordance with the following procedure:

Measured quantities (i.o, o.i, o.oi c.c., etc., or their
equivalents in dilutions) of the shell water of each of 5
oysters selected from the dozen, shall be placed in fer-
mentation tubes containing lactose peptone bile, prepared
according to the method recommended by the Committee
on Standard Methods of Water Analysis. These shall be
incubated for three days at 37° C., and the presence or
absence of gas noted daily. For all ordinary purposes of
routine work a development of from 10 to 85 per cent of
gas during this time period shall constitute a positive test
indicating a presumption of the presence of at least one
bacterium of the Bacillus coli group in the quantity of shell
water tested. But no final B. coli rating based on these
results shall be used for official approval or condemnation
unless positive confirmatory tests for the presence of organ-
isms of the B. coli group shall have been obtained from
the tube of highest or next highest dilution from each
oyster, showing the presence of gas. These confirmatory
tests shall be begun immediately upon noting the formation
of gas, and carried out in accordance with the procedure
recommended by the Committee on Standard Methods of
Water Analysis.

Statement of Results. The results of the bacterial
counts shall be expressed as Number of Bacteria per c.c.
The results of the tests for B. coli shall be expressed either
in the form of the following arbitrary numerical system
to be known as " The American Public Health Association
Method of Rating Oysters for B. coli; or in Estimated
Number of Bacteria of the B. coli Group per c.c. of the
Sample."

260 ELEMENTS OF WATER BACTERIOLOGY

The American Public Health Association Method of Rating
Oysters for B. Coli.*

The following values shall be assigned to the presence
of bacteria of the B. coli group in each of the 5 oysters
examined, these figures being the reciprocals of the greatest
dilutions in which the test for B. coli was positive:

If present in i.o c.c. but not in o.i c.c., a value of i.
If present in o.i c.c. but not in o.oi c.c., a value of 10.
If present in o.oi c.c, but not in o.ooi c.c., a value

of 100, etc.

The sum of these values for the 5 oysters gives the total
value for the sample and this figure shall be taken as the
" rating for B. coli."

The results shall be expressed in the following tabular
form:

RESULTS OF TESTS FOR B. COLI IN DILUTIONS
INDICATED

Oysters.

I.O C.C.

O.I C.C.

O.OI C.C.

Numerical
Value.

Total, or rat.

ing for B. coli

= 23

+ = presence of bacteria of the B. coli group in fermentation tube
test with lactose bile,
o = failure to demonstrate presence of bacteria of the B. coli group.

Estimated Number of B. Coli per c.c. If the standard
B. coli rating above described is divided by 5, or, in general,
if the rating is divided by the number of oysters tested,
the result will be approximately the number of B. coli

* Where the term B. coli is used, it refers in all cases to bacteria of
the B. coli group and not to the specific prototype.

EXAMINATION OF SHELLFISH

261

per c,c. of shell water. Partly because it does not do this
exactly, but also for simplicity and the avoidance of fractions,
the method of stating results as an arbitrary rating is
preferred by the committee. Practical experience with the
method also has appeared to justify this preference.

Illustrations of the Application of the Method of Rating
Oysters for B. Coli. Sometimes results similar to the
following are obtained, that is, one or more oysters may
show positive results in small quantities of shell water,
while an equal number may show negative results in larger
quantities. In this case the next lower numerical value
shall be given to the positive results in the high dilutions,
and such positive results shall be considered as being trans-
ferred to a lower dilution giving negative results in another
oyster. This is done on the theory that inconsistent results,
mathematically considered, may follow naturally from an
unequal distribution of the bacteria in the shell water.
This recession of the assigned values, however, shall not
be carried beyond the point where the number of such
recessions is greater than the number of instances where
other oysters in the series failed to give positive B. coli
results.

As examples of the method of obtaining the rating for
B. coli, the following illustrations are given. They repre-
sent results that may be met with in practice :

CASE A.— RESULTS OF B. COLI TESTS IN
INDICATED

DILUTIONS

Oysters.

I.O C.C.

O.I C.C.

0.01 C.C.

Numerical
Value.

IO
10

10 (not i)

10 (not 100)

50= rating

262 ELEMENTS OF WATER BACTERIOLOGY

CASE B.— RESULTS OF B. COLI TESTS IN DILUTIONS
INDICATED

Oysters.

I.O C.C.

O.I C.C.

0.01 C.C.

Numerical
Value.

10 (not 100)

i (not o)

23 = rating

CASE C— RESULTS OF B. COLI TESTS IN DILUTIONS
INDICATED

Oysters.

I.O C.C.

O.I C.C.

0.01 C.C.

Numerical
Value.

I
2

10
IO

IOO

-f

10 (not 100)

10 (not i)

1 40= rating

Oysters removed from the Shell (Opened or Shucked Stock).

Except as hereinafter stated, all the procedures and
requirements for the examination of opened oysters, i.e.,
shucked stock, shall be those specified for the examination
of oysters in the shell.

Selection and Preparation of Sample. The stock in the
container from which the sample is to be taken shall be
thoroughly mixed, and one or more wide-mouthed sterile
jars of a total capacity of one quart shall be each half filled
with the sample by means of a clean ladle or other instru-
ment sterilized by flaming alcohol. The jar or jars shall
be so sealed as to exclude all possibility of contamination
from without.

EXAMINATION OF SHELLFISH 263

Transportation of Samples. When the time between
the collection of the sample and its examination exceeds
3 hours, or if the outside temperature is above 50° F., the
sample shall be thoroughly refrigerated by means of ice
placed around, but not in, the sample jars.

Technical Procedure. The bacteriological examination
shall be begun as soon as possible after taking the sample.
The sample shall be thoroughly shaken at least 25 times
immediately before opening.

Bacterial Counts. The procedure specified for oysters
in the shell shall be followed:

Determination of Bacteria of the Bacillus Coll Group.
The procedure specified for oysters on the shell shall be
followed, but attention is called to the fact that higher
dilutions than -\\-§ c.c. are usually required. Triplicate
fermentation tubes shall be inoculated from each dilution
of the sample.

Statement of Results. The results of the bacteriological
examination of the opened oysters, or shucked stock, shall
be expressed in the same way as that specified for oysters
in the shell, except that in the calculation of B. coli rating
the values for the results of the positive fermentation tests
after confirmation shall be recorded for each of the inocula-
tions of each dilution. In order that the rating from these
triplicate tests may be compared with that obtained from
testing 5 oysters in the shell, the sum of the values for
the triplicate tests shall be multiplied by f . If, instead, the
sum is divided by 3, the result will give approximately the
number of B. coli per c.c.

Clams and Other Shellfish

The methods for examining clams and shellfish other
than oysters shall be those given above. Certain modi-
fications are necessary in the method of handling the samples
and the opening of the shells, etc.

264 ELEMENTS OF WATER BACTERIOLOGY

Clams are more likely to lose water during transportation
than oysters. It is therefore necessary to take greater
precautions to separate different samples of clams from
each other than in the case of oysters.

In opening soft clams it has been found that if two
incisions are made through the mantle the shell water
may be poured out without opening the shell.

Hard clams are more difficult to open, but if the shell
be struck over the dorsal muscle with a small hammer
an opening will be formed permitting the insertion of the
knife to cut the muscle.

Sometimes clams and other shellfish contain too little
liquor to make all of the tests above described. This is
always the case when the shells are very small. Under
these conditions the water from two or more shellfish shall
be taken together and tested and considered as one.

Standards of Interpretation. As in the case of water
it is neither practicable nor desirable to attempt to formulate
any hard and fast standard for passing or condemning
shellfish. It is very clear from the work carried out
by the English Commission, and at Lawrence, Boston,
Providence and New Haven in this country that shell-
fish from entirely unpolluted regions are free from colon
bacilli and that the proportion of positive tests for
these organisms increases with the increase in pollution.
Just where to draw the line, however, it is not easy to
say. According to Newlands and Ham (1910), the stand-
ards of permissible pollution adopted by various English
and American workers vary from a positive test in i
c.c. to a positive test in o.i c.c. of shell liquor. The
Bureau of Chemistry of the United States Department
of Agriculture condemns oysters sold in interstate com-
merce which show three positive tests out of five in o.i c.c.
portions, and the same standard has been adopted by the
Rhode Island Shell Fish Commission.

APPENDIX

PREPARATION OF CULTURE MEDIA

IN view of the marked influence upon bacteriological
reactions of variations in culture media caused by differ-
ences both in ingredients and in technique of preparation,
it is necessary that uniform methods be used in order to
obtain comparable data. In specifying the various ingre-
dients used in culture media it is the intention that they
shall be uniform in quality, but it is not the intention
that the recommendations as to ingredients and technical
manipulations shall stand in the way of true progress as
to improvements. When, however, improved or modified
methods are used, the variations from the standard methods
shall be plainly set forth together with the reasons for the
modifications.

INGREDIENTS

Distilled water shall be used in the preparation of stand-
ard culture, media.

Infusions of fresh lean meat, and not meat extract,
shall be used as the basis of various media.

Peptone shall be that of Witte (dry from meat).

Gelatin shall be the best French brand, so called. It
shall be as free as possible from acids and other impurities,
and shall be of such a character that a 10 per cent solution

1 From the Report of the Committee on Standard Methods of the
American Public Health Association.

265

266 APPENDIX

prepared in the usual way shall not soften when kept at
a temperature of 25° C.

Commercial agar in threads shall be of as high a grade
as can be obtained. Agar may be purified by washing.

The various sugars, such as dextr se, lactose, and sac-
charose, shall be as nearly as possible the chemically pure
compounds designated. Unusual effort to obtain such
sugars is considered to be necessary.

Glycerine shall be double distilled.

In place of litmus, a i per cent aqueous solution of
Kahlbaum's azolitmin may be used.

Of the various other ingredients used, nearly all of which
are of a mineral nature, special effort shall be made to
see that they are chemically pure products within the
full meaning of this expression.

STERILIZATION

Sterilization in the autoclave seems to be preferable to
that in flowing steam. Both in the lowering of the melting
point of gelatin and in the breaking down of sugar media
the time of sterilization has a greater effect than does the
temperature within the standard limits. It is, therefore,
suggested that small containers be used and that media
be sterilized in the autoclave at 120° C. (15 pounds pressure)
for 15 minutes. A shorter period than this, in practice,
sometimes results in incomplete sterilization, and a longer
time results in the inversion of sugars or the lowering of
the melting point of gelatin. Agar media should be melted
before placing in the autoclave.

An important point in the sterilization of gelatin and
sugar media is to have the sterilizer hot when the media
are introduced, so that heating to the point of sterilization
will be accomplished as quickly as possible. Also when
sterilization is complete the media should be cooled rapidly.
This not only reduces the time of heating, thus preserving

APPENDIX 267

the gelatin and sugar, but also assists in the actual sterili-
zation. It is also advisable in the use of the autoclave
to keep the small pet cock at the bottom partially open
so that steam is escaping during sterilization. This insures
the removal of all air. If practicable, store media at room
temperature for two days to see that sterilization is complete.
In intermittent sterilization, media shall be placed on
each of three successive days in streaming steam for 30
minutes after the steam fills the sterilizer.

REACTION

Phenolphthalein shall be the standard indicator used
in obtaining the reaction of all media. Turmeric paper
possesses similar properties, and its use is advised where
phenolphthalein is not available.

Titrations and adjustment of reactions shall be made
as follows:

Put 5 c.c. of the medium to be tested into 45 c.c. dis-
tilled water. Boil briskly one minute. Add i c.c. of
phenolphthalein solution (5 gm. of commercial salt in i
liter of 50 per cent alcohol). Titrate while hot (preferably
while boiling) with N/20 caustic soda. A faint but distinct
pink color marks the true end-point. This distinct pink
color may be more precisely described as a combination
of 25 per cent of red (wave length approximately 658)
with 75 per cent of white as shown by the disks of the
color top (described under Record of Tints and Shades
of Apparent Color, p. 10 of Standard Methods Report).
In practice, titration is continued until the pink color of
alkaline phenolphthalein matches that of the fused disks.

All reactions shall be expressed with reference to the
phenolphthalein neutral point and shall be stated in per-
centages of normal acid or alkaline solutions required to
neutralize them. Alkaline media shall be recorded with
the minus ( — ) sign before the percentage of normal acid

268 APPENDIX

needed for their neutralization, and acid media with the
plus (+) sign before the percentage of normal alkaline
solution necessary for their neutralization.

The standard reaction of culture media shall be +1.0
per cent. If it differs from i per cent by more than 0.2
per cent it should be readjusted.

Wherever reactions other than the standard above given
are used it shall be clearly stated in all results of bacterial
work, and the reasons therefor also stated.

STORAGE OF MEDIA

It is recognized by the committee that it is desirable
to prepare media in large quantities in order to guard
against discrepancies in composition; but, all things con-
sidered, the complications resulting from the varying
amounts of heating incident to withdrawing portions from
time to time and tubing it, are believed to more than
offset this advantage. Consequently, when possible, media
shall be put at once into tubes and placed in cold storage.

To guard against changes due to evaporation all media
not used promptly shall be stored in a moist atmosphere,
preferably in an ice-box, or else the flask shall be sealed
by dipping the cotton plug in paraffin.

NUTRIENT BROTH

Nutrient broth shall be prepared as follows: Infuse
500 gm. chopped lean meat 24 hours with 1000 c.c. dis-
tilled water in refrigerator. Restore loss by evaporation.
Strain infusion through cotton flannel. Add i per cent
peptone. Warm on water bath, stirring until the pep-
tone is dissolved. Heat over boiling water (or steam)
bath 30 minutes. Restore loss by evaporation. Titrate.
Adjust reaction to -fi per cent by adding normal hydro-
chloric acid or normal sodium hydrate, as required. Boil

APPENDIX 269

2 minutes over free flame, constantly stirring. Restore loss
by evaporation. Filter through absorbent cotton and cotton
flannel, passing the liquid through until clear. Titrate and
record final reaction. Tube, using 10 c.c. in each tube.
Sterilize.

SUGAR BROTHS

Sugar broths shall be prepared in the same general manner
as the standard nutrient broth, with the addition of i
per cent of dextrose, lactose, saccharose or other sugar,
just before sterilizing. The removal of muscle-sugar by
inoculation with B. coli is unnecessary if small amounts
of gas are disregarded.

If, however, test-tubes of small dimensions are used and
the presence of any gas, however small, is taken to indicate
gas formation the removal of muscle sugar is necessary.

The reaction of sugar broths shall be neutral to phenol-
phthalein.

Sterilization may be done in streaming steam, as usual,
or in the autoclave at 120° C. (15 pounds pressure) for
15 minutes. Sterilization in the autoclave seems to be
preferable to that in flowing steam.

For the routine work of testing samples in water for
B. coli, especially large volumes of water are to be mixed
with broths of such strength that the resulting mixture
will be one of normal strength. Dextrose broth made
with Liebig's Beef Extract is not equal in effectiveness
to that made of fresh beef extract and should not be sub-
stituted for the latter.

LACTOSE BILE

The lactose bile medium consists of sterilized undiluted
fresh ox gall (or a 10 per cent solution of dry fresh ox gall)
to which has been added i per cent of peptone and i per
cent of lactose. The addition of peptone is important.

270 APPENDIX

GELATIN

No gelatin media should be employed having a melting
point below 25° C. The percentage of gelatin added may
be increased to bring the melting point up to the desired
figure. With most gelatin on the market n per cent
seems to be preferable to the standard of 10 per cent,
provided the gelatin is weighed out without correcting for
contained moisture, as appears to have been the custom.
Ten per cent, or even 20 per cent of moisture commonly
occurs in laboratory gelatin, and unless this is taken into
account in weighing, the stiffness of the media is materially
affected and the bacterial results obtained considerably
modified. All gelatin should be tested for moisture before
using by drying a sample for half an hour at 105° C. The
stock should be kept under uniform conditions in tight
containers, so that the percentage of water present may
then be properly accounted for and the weight on a dry
basis be used in making up the medium.

AGAR

For bacterial counts 10 gm. of agar per liter should be
used. The smaller amount seems to be sufficient to carry
the added water and the medium is less stiff. This appears
to give higher and more consistent counts. Fifteen grams
may be employed for keeping cultures.

North medium is especially valuable for keeping cultures,
particularly cocci. This is composed as follows:

500 c.c. Extract of 500 gms. of chopped beef.
500 c.c. Distilled Water.
10 gm. Agar.
20 gm. Gelatin.
20 gm. Peptone.
5 gm. Sodium Chloride.
Reaction neutral.

It is well to determine the reaction of the media after
sterilization, as during this process the reaction often changes

APPENDIX 271

and the final results should correspond to the acidity
recommended by the standard methods. What has been
said regarding the necessity of correcting for moisture in
gelatin applies with equal force to agar and for the same
reasons.

NUTRIENT GELATIN AND AGAR

Nutrient gelatin and agar shall be prepared as follows:

Gelatin. Agar.

1. Boil 15 gm. thread agar in 500 c.c.

water for half an hour and make
up weight to 500 gm. or digest for
15 minutes in the autoclave. Let
this cool to about 60° C.

2. Infuse 500 gm. lean meat 24 Infuse 500 gm. lean meat 24 hours
hours with 1000 c.c. of dis- with 500 c.c. of distilled water in
tilled water in refrigerator. refrigerator.

3. Make up any loss by evaporation.

4. Strain infusion through cotton flannel.

5. Weigh filtered infusion.

6. Add i per cent Witte's pep- Add 2 per cent of Witte's pep-
tone and 10 per cent Gold tone.

Label sheet gelatin on dry
basis.

7. Warm on water bath, stirring till peptone and
gelatin are dissolved and not allowing the
temperature to rise above 60° C.

8. To 500 gm. of the meat infusion

add 500 c.c. of the 3 per cent
agar, keeping the temperature
below 60° C.

9. Titrate, after boiling one minute to expel
carbonic acid.

10. Adjust reaction to +1.0 per cent by adding
normal hydrochloric acid or sodium hydrate
as required.

11. Heat over boiling water (or steam) bath for
40 minutes.

12. Restore loss by evaporation.

13. Readjust to +1.0 per cent if necessary and
boil 5 minutes over free flame, constantly
stirring.

14. Make up loss by evaporation.

272 APPENDIX

15. Filter through absorbent cotton and cotton
flannel, passing the nitrate through the filter
until clear.

16. Titrate and record the final reaction.

17. Tube, using 10 c.c. of medium in each tube.

1 8. Sterilize 15 minutes in the autoclave at 120
degrees, or for 30 minutes in streaming steam
on three successive days. Put the gelatin at
once into ice-water till solidified.

19. Store in the ice-chest in a moist atmosphere,
to prevent evaporation.

LACTOSE LITMUS AGAR

Lactose litmus agar shall be prepared in the same manner
as nutrient agar, with the addition of i per cent of lactose
to the medium just before sterilization. The reaction shall
be made neutral to phenolphthalein (see p. 267).

If the medium is to be used in tubes the sterilized azo-
litmin solution shall not be added until just before the
final sterilization.

If the medium is to be used in Petri dishes the steril-
ized azolitmin solution shall not be added to the medium
until it is ready to be poured into the dishes.

More colonies and better general results are obtained
on the lactose litmus agar plates, when the litmus and
lactose are each sterilized separately and added to the
plate with the neutral agar at the time of planting. Good
results can, however, be obtained, if the agar and lactose
are mixed and sterilized in an autoclave at 120° C. for
15 minutes only.

The azolitmin on the market varies considerably, much
of that sold being entirely unreliable for the purpose. A i
per cent solution of Kahlbaum's azolitmin, if boiled for
5 minutes, readily dissolves and needs no correction for
acidity if added to standard agar. Many bacteriologists
prefer pure litmus to azolitmin, and it is therefore sug-
gested that its use be made optional. Both total and
red colonies may be counted after from 18 to 24 hours

APPENDIX 273

when incubated at 37° C. Such tests are sometimes used
in the control of nitration plants.

LIVER BROTH

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 (Witte's) 10 gm.

Dextrose 10 gm.

Di-Potassium Phosphate (K^HPO-i) i gm.

Water 1000 gm.

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

2. Boil slowly for 2 hours in a double boiler, starting
cold and stirring occasionally.

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

4. To the filtrate add 10 gm. of peptone, 10 gm. of
dextrose and i gm. of potassium phosphate.

5. After warming this mixture in a double boiler and
stirring it for a few minutes to dissolve the ingredients,
titrate with N/2O sodium hydrate, using phenolphthalein as
an indicator, and neutralize with normal sodium hydrate.

6. 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.

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

274 APPENDIX

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

Other valuable liver media (for use in the identification
of B. sporogenes and other species) are prepared as given
below:

LIVER GELATIN

1. Proceed as in steps i, 2, 3, in preparing liver broth.

2. Cool the filtrate to 50° C. Add 10 per cent of sheet
gelatin and stir a few minutes until dissolved.

3. Add i per cent of peptone, i per cent of dextrose
and o.i per cent of potassium phosphate.

4. Stir until the ingredients are dissolved, keeping the
temperature below 50° C., and then proceed as in steps

5, 6, 7, 8.

LIVER AGAR

1. Chop 500 gm. of beef liver into small pieces, add 500
c.c. of distilled water, and boil slowly for 2 hours, stirring
occasionally.

2. Add 5 gm. of agar (dried at 105° C. for 30 minutes)
to 500 c.c. of distilled water and digest for 30 minutes in
an autoclave at 120° C. (15 pounds pressure).

3. After making up the loss by evaporation, pass the
liver infusion through a wire strainer, add 500 c.c. of the
filtrate to the agar solution and proceed as in steps 4, 5,

6, 7, 8, in preparing liver broth.

It is very important to note that liver broth should not
be exposed to the high temperature attained in the auto-
clave any longer than 15 minutes, as prolonged heating
above the boiling-point causes caramelization of the carbo-
hydrates, rendering the medium less delicate for bacterial
development. For the rejuvenation of attenuated cultures,
especially B. sporogenes, the addition of very small pieces
of liver tissue, which have been sterilized in Petri dishes

APPENDIX 275

in the autoclave for 15 minutes improves the rejuvenating
properties of the medium. They should be added to the
tubes after sterilization.

Bacterial growth being very rapid in this medium, pre-
liminary rejuvenation at 37° C. should be concluded between
6 and 12 hours.

HESSE AGAR

Agar (absolutely dry) 4.5 gm.

Peptone, Witte 10 gm.

Liebig's Extract of Beef 5 gm.

Salt 8.5 gm.

Distilled Water 1000 c.c.

Dissolve 4.5 gm. of dry agar in 500 c.c. distilled water
by heating over a free flame, making up loss in weight by
evaporation. Into another vessel 500 c.c. of distilled water
is poured and to this is added 10 gm. of peptone, 5 gm. of
Liebig's Beef Extract, and 8.5 gm. of salt. This is heated
until all is dissolved and the loss in weight by evaporation
is made up by adding distilled water.

Add the two solutions together; boil 30 minutes; make
up loss in weight with distilled water, filter through absorb-
ent cotton held in the funnel by cotton flannel, passing
the filtrate through several times until perfectly clear.
Test the reaction; adjust, if necessary, to +1.0 per cent
normal acid, and tube, using 10 c.c. of medium in each
tube. Sterilize for 15 minutes at 120° C. (15 pounds
pressure) in an autoclave. Cool with running water and
store in an ice-chest the air of which is saturated with
moisture.

CONRADI-DRIGALSKI MEDIUM

These authors have modified lactose litmus agar by
adding to it nutrose and crystal violet and by using 3 per
cent of agar instead of i per cent. The crystal violet
strongly inhibits the growth of many other bacteria, especially
cocci, which would also color the medium red; the 3 per

276 APPENDIX

cent agar makes the diffusion of the acid which is formed
more difficult.

Three pounds of chopped beef are allowed to stand 24
hours with 2 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 3
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 per cent soda solution. To this feebly alkaline
mixture 4 c.c. of hot sterile 10 per cent soda solution are
added and 20 c.c. of a sterile solution (i to 1000) of crystal
violet (Hochst B.).

ENDO MEDIUM

Make 3 per cent agar as follows:

1. To i liter of cold water add 30 gm. of powdered agar
by sifting successive small portions upon the surface and
allowing them to settle.

2. Add 10 gm. of Witte's peptone and 5 gm. of Liebig's
meat extract.

3. Heat in double boiler until the ingredients are thoroughly
dissolved.

4. Neutralize the sodium carbonate, using litmus as an
indicator, and then add 10 c.c. of a 10 per cent solution
of sodium carbonate.

5. Store the medium in flasks in lots of 100 c.c. or in larger
known quantities, the flasks being large enough for the
other ingredients to be added later.

6. Sterilize 2 hours in streaming steam.

It is essential to use powdered agar and cold water,
since the agar settles more readily in cold water.

APPENDIX 277

To use Endo medium: (a) make a 10 per cent solution
of sodium sulphite in water; (b) make a 10 per cent solution
of basic fuchsin in 96 per cent alcohol; (c) add 2 c.c. of
the fuchsin solution to 10 c.c. of the sulphite solution, pre-
pared as above, and steam a few minutes in an Arnold
sterilizer.

7. Add i gm. of chemically pure lactose to each 100 c.c.
of Endo medium.

8. Melt the Endo medium in streaming steam and add
| c.c. of the fuchsin-sulphite solution described above.

9. Pour plates, and allow them to harden thoroughly
in the incubator.

The lactose used must be chemically pure, and the sul-
phite solution must be made up fresh every day.

HISS' MEDIA

Two media are used: one for the isolation of the typhoid
bacillus by plate culture, and one for the differentiation
of the typhoid bacillus from all other forms in pure culture
in tubes.

Plate Medium. The plate medium is composed of 10
gm. of agar, 25 gm. of gelatin, 5 gm. of sodium chloride,
5 gm. of Liebig's extract, 10 gm. of dextrose, and 1000 c.c.
of water. When the agar is thoroughly melted the gelatin
is added and completely dissolved by a few minutes' boiling.
The medium is then titrated, to determine its reaction,
phenolphthalein being used as an indicator. The requisite
amount of normal hydrochloric acid or of sodium hydrate
solution is added to bring it to the desired reaction; i.e.,
H-2. To clear the medium add the whites of one or two
eggs, well beaten in 25 c.c. of water, boil for 45 minutes,
and filter through a thin filter of absorbent cotton. Add
the dextrose after clearing. The reaction of the medium
is most important; it should contain never less than 2 per
cent of normal acid.

278 APPENDIX

Tube Medium. The tube medium contains agar 5 gm.,
gelatin 80 gm., sodium chloride 5 gm., meat extract 5 gm.,
and dextrose 10 gm. to the liter of water; reaction +1.5.
The mode of preparation is the same as for the plate medium,
care being taken always to add the gelatin alter the agar
is thoroughly melted, so as not to alter this ingredient by
prolonged exposure to high temperature. The dextrose is
added after clearing.

MILK

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 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 separa-
tory funnel. Should the milk be too acid the reaction shall
be corrected to +i 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 i
per cent azolitmin. As it is impossible to make each lot
of litmus milk with the same shade of color, it is recom-
mended that a control tube be always exposed with the
inoculated tubes for purposes of comparison.

NITRATE BROTH

Dissolve i gm. peptone in i liter of tap water, and add
0.2 gm. of nitrite-free potassium nitrate.

PEPTONE SOLUTION FOR INDOL TEST

Broth which has been inoculated with B. coli to remove
the muscle sugar contains toxins which interfere with the
proper growth of many species, hence peptone solution,

APPENDIX 279

(i per cent peptone in water) is recommended for use in
the indol test.

^ESCULIN MEDIA

Broth. Dissolve 10 gm. peptone, 5 gm. commercial
sodium taurocholate (or glycocholate), o.i gm. aesculin and
0.5 gm. soluble iron citrate in i liter distilled water.

Tube in 5 c.c. quantities in test-tubes and sterilize.

It is preferable to keep the iron citrate solution separate
and to add it to the rest of the media as used, since the
iron causes a precipitate. Sometimes the iron citrate will
not readily dissolve. In such case a few drops of ammonia
will cause immediate solution, especially on warming. The
excess of ammonia should then be boiled off.

A gar. This medium is made in the same manner as the
broth with the addition of the usual amount of agar.

APPARATUS

Few definite requirements need be made respecting
apparatus. The quality of the glass shall be such as not
to be easily acted upon by the reagents used, and all glass-
ware shall be scrupulously clean when used. When nec-
essary it shall be sterilized by dry heat for one hour at
about 150° C. A slight browning of the cotton stoppers
is a good index of proper exposure.

In some operations, as for example, the determination
of the thermal death point, it is necessary to use test-tubes
of a definite size and thickness. For this purpose the
standard size culture tube shall be 15 cm. long, 1.6 cm.
in diameter, and of medium weight. Tubes to be filled
with gelatin for quantitative work may be those described
as 6 in. Xf in. " heavy."

The standard loop for making transfers shall be pre-
pared as follows:

280 APPENDIX

Bend the end of a piece of No. 27 platinum wire about
10 cm. long over a bit of No. 10 wire, and fasten the loop
thus formed into a glass rod to serve as a handle. A
loopful of culture shall be interpreted as meaning all the
fluid that the loop can hold. That is, the fluid shall form
a bi-convex body and shall not be simply a film covering
the space in the loop.

The standard fermentation tube shall be a tube 1.5 cm.
in diameter, bent at an acute angle, closed at one end and
provided with a bulb at the other which is large enough
to receive all the liquid contained in the closed portion.
The length of the closed end of the tube shall be about
14 cm. The bulb shall have a diameter of about 3.8 cm.

INCUBATION

There shall be two standard temperatures of incubation—
20° C. and 37° C., the first corresponding to ordinary room
temperature, the second to blood heat. The temperature
of the incubators shall not be allowed to vary from these
two standards by more than i° C. in either direction.

The atmosphere of the incubator shall be kept moist,
preferably near the point of saturation. The incubator shall
be ventilated so as to insure a reasonably good circulation
of air in order to prevent the accumulation in the incubator
of gases which might be prejudicial to the development of
the bacteria.

No definite period of incubation can be prescribed which
will be suitable for all the work of species determination,
but in reporting results the period used shall always be
stated and form a part of the report. General statements
as to the necessary periods will be found in connection
with the principal tests.

REFERENCES

ABBA, F. 1895. Sulla presenza del bacillus coli nelle acque pota-

bili, e sopra un metodo per metterlo in evidenza. La Riforma

medica, XI, 302.
ABBA, F., Orlandi, E. and Rondelli, A. 1899. Ueber die Filtra-

tionskraft des Bodens und die Fortschwemmung von Bakterien

durch das Grundwasser. Zeitschrift fur Hygiene, XXXI, 66.
ABBOTT, A. C. 1899. The Principles of Bacteriology. Philadelphia

and New York, 1899.
ADAMI, J. G., and CHOPIN, J. A. 1904. On a Simple Method of

Isolating from Water Forms which Agglutinate with Typhoid

Serum. Journal of Medical Research, XI, 469.
ALTSCHULER, E. 1903. Eine typhusanraicherungsmethode. Cen-

tralblatt fiir Bakteriotogie, Abth. i, XXXIII, 741.
AMYOT, J. A. 1902. Is the Colon Bacillus a Normal Inhabitant

of the Intestines of Fishes? Transactions of the American Public

Health Association. 1901 Meeting. XXVII, 400.
ANDREWES, F. W. and HORDER, T. J. 1906. A study of the strepto-
cocci pathogenic for man. Lancet, CLXXI, 708.
AYERS, S. H. 1911. A Counting Apparatus for Litmus Plates with a

cold counting Plate for Use with Gelatin Plates. Journal of the

American Public Health Association, I, 892.
BARTHEL, C. Bidrag till kannedonnen om mjolksyre-bakteriernas

forekomst och utbredunig utom mjolken. Rev. Centralblatt fiir

Bakteriologie, Abth. II, XVI, 550.
BECKMANN, W. 1894. Ueber die typhusahnlichen Bakterien des

Strassburger Wasserleitungswassers. Archiv fiir experimentelle

Pathologic und Pharmakologie, XXXIII, 466.
BELITZER. 1899. Zur Lehre iiber das Bacterium coli commune.

Rev. Jahresbericht iiber die Fortschritte in der Lehre von den

pathogenen Mikroorganismen, XV, 326.
BERGEY, D. H., and DEEHAN, S. J. 1908. The colon-aerogenes group

of bacteria. Journal of Medical Research, XIX, 175.

281

282 REFERENCES

BETTENCOURT, A., and BORGES, I. 19080. Recherches sur le B. coli,
des Vertebres Inferieurs et des Cereales. Archives do Real Insti-
tute Bacteriologico Camara Pestana, Tome II, Ease. II, 221.

BETTENCOURT, A., and BORGES, I. 19086. Peut-on distinguer le
Coli bacille de I'Homme de celui des animaux au moyen de la
fixation du complement. Archives do Real Institute Bacteri-
logico Camara Pestana, Tome II, Fasc. II, 245.

BIFFI, U. 1906. La coagulabilita al calore delle culture in latte
come elemento di diagnosi bacteriologica. Soc. Medico-Chirurgica
di Bologna, 16, Diz. 1906. Rev. Centralblatt fur Bakteriologie,
Abth. I, Ref., XL, 312.

BIFFI, N. 1907. Ancora sulla prova della coagulabilita delle culture
in latte nella diagnostica bacteriologica del gruppo tifocoli. Societa
Medico-Chirurgica di Bologna, Mar. 14, 1907. Rev. Bulletin de
1'institut Pasteur, V, 380.

BLACHSTEIN. 1893. Contribution a 1'etude microbique de 1'eau.
Annales de ITnstitut Pasteur, VII, 689.

BOLTON, M. 1886. Ueber das Verhalten verschiedener Bakterien-
arten im Trinkwasser. Zeitschrift fur Hygiene, I, 76.

BOULLANGER, E. and MASSOL, L. 1903. Etudes sur les microbes
nitrificateurs. Annales de 1'institut Pasteur, XVII, 492.

BOYCE, MACCONKEY, GRUNBAUM and HILL. 1902. The Reduc-
tion in the Number of Bacteria in a Biological Effluent (Double
Contact), by means of Filtration. Second Report of the Royal
Commission on Sewage Disposal, 87.

BRAUN, A. 1906. Le rouge neutre et le diagnostic rapide de la souil-
lure des eaux de boisson par le colibacille. Bulletin de 1'institut
Pasteur, IV, 561.

BREZINA, E. 1906. Die Donau vom Leopoldsberge bis Pressburg,
die Abwasser der Stadt Wien und deren Schicksal nach ihrer
Einmundung in den Strom. Zeitschrift fur Hygiene, LIII, 369.

BROTZU, L. 1895. Sulla desinfezione del canale intestinale. Annali
del 1'instituto d'igiene sperimentale della R. Universita di Roma,
IV, 427. Rev. Centralblatt fur Bakteriologie, XVII, 726.

BROWN, C. C. 1893. Fourth Report on Mohawk and Hudson Rivers.
Thirteenth Annual Report of the State Board of Health of New
York for 1892, 680.

BUCHAN, G. F. 1910. Typhoid Fever and Mussel Pollution. Journal
of Hygiene, X, 569.

BUCHNER. 1893. Ueber den Einfluss des Lichtes auf Bakterien
und liber die Selbstreinigung der Fliisse. Achiv fur Hygiene,
XVII, 179.

BULSTRODE, H. T. 191 1. Report on Shellfish other than Oysters in

REFERENCES 283

Relation to Disease. Supplement to the Thirty-ninth Annual
Report of the Local Government Board in continuation of the
Report of the Medical Officer for 1909-10.

BURRI, R. 1895. Nachweis von Fakalbakterien im Trinkwasser.
Hygienische Rundschau, V, 49.

BURRI, R. 1910. Ueber scheinbar pldtzliche Neuerwerbung eines
bestimmten Garungsvermogens durch Bakterien der Coligruppe.
Centralblatt fur Bakteriologie, Abth. II? XXVIII, 321.

CALMETTE, A. 1905. Recherches sur 1'epuration biologique et
chimique des eaux d'egout effectuees a 1'institut Pasteur de Lille
et a la station experimentale de la Madeleine, I.

CALMETTE, A. 1907. Recherches sur 1'epuration biologique et
chimique des eaux d'egout effectuees a 1'institut Pasteur de Lille
et a la station experimentale de la Madeleine. II.

CHICK, H. 1901. The Distribution of Bacterium coli commune.
The Thompson- Yates Laboratories Report, III, i and 117.

CHRISTIAN. 1905. Zum Nachweis fakaler Verunreinigung von Trink-
wasser. Archiv fur Hygiene, LIV, 386.

CLARK, H. W. 19060. Experiments upon the Purification of Sewage
and Water at the Lawrence Experiment Station during the Year
1905. Thirty-seventh Annual Report of the State Board of Health
of Massachusetts for 1905.

CLARK, H. W. 19066. Studies at the Lawrence Experiment Station
on the Pollution of Shellfish. Thirty-seventh Annual Report of
the State Board of Health of Massachusetts for 1905, 429.

CLARK, H. W., and GAGE, S. DEM. 1900. The Significance of the
Appearance of B. Coli communis in Filtered Water. Journal
of the Boston Society of Medical Sciences, IV, 172.

CLARK, H. W., and GAGE, S. D. 1905. The Furtbtions of Various
Types of Bacteria in the Purification of Sewage, with some Methods
for their Quantitative Determination. Engineering News, LIII, 27.

CLARK, H. W., and GAGE, S. D. 1909. Disinfection as an Adjunct to
Water Purification. Journal of the New England Water Works
Association, XXIII, 302.

CLEMESHA, W. W. 19120. The Bacteriology of Surface Waters in the
Tropics. Calcutta and London.

CLEMESHA, W. W. 19126. A Criticism of Dr. A. C. Houston's Report
on the Biological Characters of B. coli isolated from (i) Raw,
(2) Stored River Water, and (3) Stored and Filtered Water. Jour-
nal of Hygiene, XII, 463.

COMMITTEE OF BACTERIOLOGISTS. 1898. Procedures recommended
for the Study of Bacteria, with especial Reference to Greater
Uniformity in the Description and Differentiation of Species,

284 REFERENCES

Being the Report of a Committee of Bacteriologists to the Com-
mittee on the Pollution of Water-supplies of the American Public
Health Association. Transactions of the American Public Health
Association. 1897 Meeting. XXIII, 60.

COMMITTEE APPOINTED TO CONSIDER THE STANDARDIZATION OF METHODS
FOR THE BACTERIOSCOPIC EXAMINATION OF WATER. 1904. Report.
Journal of State Medicine, XII, 471.

COMMITTEE ON STANDARD METHODS FOR THE BACTERIOLOGICAL EXAM-
INATION OF SHELLFISH. 1912. American Journal of Public
Health, II, 34-

COMMITTEE ON STANDARD METHODS OF WATER ANALYSIS. 1905.
Report of Committee on Standard Methods of Water Analysis to
the Laboratory Section of the American Public Health Association.
Journal of Infectious Diseases, Supplement No. i, i.

COMMITTEE ON STANDARD METHODS OF WATER ANALYSIS. 1912.
Report of Committee on Standard Methods for the Examination
of Water and Sewage to the Laboratory Section of the American
Public Health /Association. Second edition, New York.

COPELAND, W. R. 1901. The Use of Carbolic Acid in Isolating
the Bacillus coli communis from River Water. Journal of the
Boston Society of Medical Sciences, V, 381.

COPELAND, W. R., and HOOVER, C. P. 1911. The Interpretation of
Tests for B. coli communis. Journal of Infectious Diseases, IX,

343-
CRAMER, C. 1885. Die Wasserversorgung von Zurich und ihr

Zusammenhang niit den Typhus-epidemic des Jahres 1884.

Zurich.
DITTHORN, F., and GILDEMEISTER, E. 1906. Eine Anreicherungs-

methode fur den Nachweis von Typhusbacillen in Trinkwasser

bei der chemischen Fallung mit Eisenoxychlorid. Hygienische

Rundschau, XVI, 1376.
DOEBERT, A. 1906. Wachstum von Typhus- und Koli-Reinkul-

turen auf verschiedenen Malachitgriin-Nahrboden. Archiv fur

Hygiene, LIX, 370.
DOLT, M. L. 1908. Simple Synthetic Media for the Growth of B.

coli and for its Isolation from Water. Journal of Infectious Diseases,

V, 616.
DOMBROWSKY. 1903. Zur Biologic der Ruhrbacillen. Archiv fur

Hygiene, XL VII, 243.
DOWNES, A., and BLUNT, T. P. 1877. Researches on the Effect

of Light upon Bacteria and other Organisms. Proceedings of

the Royal Society, XXVI, 488.
DREW, G. H. 1912. Report of Investigations on Marine Bacteria

EEFERENCES 285

carried on at Andros Island, Bahamas, British West Indies, in

May, 1912. Year Book of the Carnegie Institution of Washington,

No. n, 136.
DRIGALSKI, V. 1906. Ueber ein Verfahren zur Zuchtung von Typhus-

ba/dllen aus Wasser und ihren Nachweis in Brunnenwasser.

Arbeiten aus dem Kaiserlichen Gesundheitsamte, XXIV, 68.
DRIGALSKI, V., and CONRADI, H. 1902. Ueber ein Verfahren zum

Nachweis der Typhusbacillen. Zeitschrift fur Hygiene, XXXIX,

283.

DUCLAUX, E. 1898. Traite de Microbiologie, Vol. I. Paris, 1898.
DUGGELI, M. 1904. Die Bakterienflora gesunder Samen und daraus

gezogener Keimpflanzchen. Centralblatt fur Bakteriologie, Abth.

II, XII, 602.
DUNBAR. 1896. Zur Differentialdiagnose zwischen den Cholera-

vibrionen und anderen denselben nahestehenden Vibrionen. Zeit-
schrift fur Hygiene, XXI, 295.
DUNHAM, E. K. 1887. Zur chemischen Reaction der Cholera-

bakterien. Zeitschrift fur Hygiene, II, 337.

DUNHAM, E. K. 1889. On the Bacteriological Test of Drinking-
water. Medical Record, XXXVI, 367.
DURHAM, H. E. 1898. On the Serum Diagnosis of Typhoid Fever.

with special reference to the Bacillus of Gartner and its Allies.

Lancet, 1898, I, 154.
DURHAM, H. E. 1901. Some theoretical Considerations upon the

Nature of Agglutinins, together with further Observations upon

Bacillus typhi abdominalis, Bacillus enteritidis, Bacillus coli

communis, Bacillus lactis aerogenes, and some other Bacilli of

allied Character. Journal of Experimental Medicine, V, 353.
DYAR, H. G., and KEITH, S. C. 1893. Notes on Normal Intestinal

Bacilli of the Horse and of certain other Domesticated Animals.

Technology Quarterly, VI, 256.
EDDY, H. P. 1902. Annual Report of the Superintendent of Sewers

of the City of Worcester for the Year ending November 30, 1901.

Worcester. 1902.
EIJKMAN, C. 1904. Die Garungsprobe bei 46° als Hilfsmittel bei

der Trinkwasseruntersuchung. Centralblatt fur Bakteriologie,

Abth. I, Orig., XXXVII, 742.
ELSNER. 1896. Untersuchungen liber electives Wachsthum der

Bacterium coli-arten und des Typhusbacillus. Zeitschrift fur

Hygiene, XXI, 25.
ENDO, S. 1904. Ueber ein Verfahren zum Nachweis der Typhus

bacillen. Centralblatt fur Bakteriologie, Orig., XXXV, 109.
ENGINEERING NEWS. 1912. Bacterial Contents of Allegheny, Monon-

286 REFERENCES

gahela and Ohio Rivers at and below Pittsburgh. Engineering

News, LXVII, 483.
ESCHERICH, T. 1885. Die Darmbakterien des Neugeborenen und

Sauglings. Fortschritte der Medicin, III, 515 and 547.
EYRE, J. W. H. 1904. On the Distribution of B. coli in Nature.

Lancet, 1904, I, 648.
FEHRS. 1906. Die. Beeinflussung der Lebensdauer von Krank-

hcitskeimeri im Wasser durch Protozoen. Hygienische Rund-
schau, XVI, 113.
FERREIRA, A., HORTA, A., and PAREDES, C. 19080. Recherches sur

le B. coli communis de I'intestin de 1'homme. Archives do Real

Institute Bacteriologico Camara Pestana, Tome II, Fasc. II, 153.
FERREIRA, A., HORTA, A., and PAREDES, C. 19086. Recherches sur le

B. coli de I'intestin des Mammiferes et des Oiseaux. Archives do

Real Institute Bacteriologico Camara Pestana, Tome II, Fasc. II,

203.
FICKER, M. 1904. Ueber den Nachweis von Typhusbacillen im

Wasser durch Fallung mit Eisen sulfat. Hygienische Rundschau.

XIV, 7.
FISCHER, A. 1900. The Structure and Functions of Bacteria. Trans.

by Jones, A. C. Oxford, 1900.
FISCHER, B., and FLATAU, G. 1901. Typhusbacillen in einer ein-

gesandten typhusverdachtigen Wasserprobe. Centralblatt fill

Bakteriologie, XXIX, 329.
FORD, W. W. 1903. The Classification and Distribution of the

Intestinal Bacteria in Man. Studies from the Royal Victoria

Hospital, I, also Studies from the Rockefeller Institute for Medical

Research, II.

FRANKLAND, G. and P. 1894. Micro-organisms in Water. Lon-
don, 1894.
FREMLIN. 1893. Vergleichende Studien an Bact. coli commune

verschiedener Provenienz. Archiv fiir Hygiene, XIX, 295.
FREUDENREICH, E. VON. 1888. De 1'antagonisme des bacteries et

de 1'immunite qu'il confere aux milieux de culture. Annales de

1'Institut Pasteur, II, 200-.
FREUDENREICH, E. VON. 1895. Ueber den Nachweis des Bacillus

coli communis im Wasser und dessen Bedeutung. Centralblatt

fur Bakteriologie, XVIII, 102.
FROMME, W. 1910. Ueber die Beurteilung des Colibakterienbefundes

im Trinkwasser nebst Bemerkungen iiber den Nachweis und das

Vorkommen der Colibazillen. Zeitschrift fiir Hygiene, LXV, 251.
FROST, W. D. 1901. A Laboratory Guide in Elementary Bacteriology.

Madison, 1901.

REFERENCES 287

FROST, W. D. 1904. The Antagonism exhibited by certain Sapro-
phytic Bacteria against the Bacillus typhosus, Gaffky. Journal
of Infectious Diseases, I, 599.

FROST, W. H. 1910. Note on an organism isolated from Washington
Tap Water agglutinated readily by the serum of Typhoid Patients.
American Journal of Public Hygiene, XX, 670.

FULLER, G. W. 1895. On the Proper Reaction of Nutrient Media
for Bacterial Cultivation. Transactions of the American Public
Health Association. 1894 Meeting. XX, 381.

FULLER, G. W. 1899. Report on the Investigations into the Puri-
fication of the Ohio River Water for the Improved. Water-supply
of the City of Cincinnati, Ohio. Cincinnati, 1899.

FULLER, G. W. 1902. Report of Committee on Standard Methods
of Water Analysis. Transactions of the American Public Health
Association. 1901 Meeting. XXVII, 377.

FULLER, G. W., and FERGUSON, C. E. 1905. Concerning Tests for
B. coli communis in Water. Journal of Infectious Diseases,
Supplement No. 2, 142.

FULLER, G. W., and JOHNSON, G. A. 1899. On the Differentiation
and Classification of Water Bacteria. Journal of Experimental
Medicine, IV, 609.

GAEHTGENS, W. 1905. Ueber die Erhohung der Leisterungs-
fahigkeit des Endoschen Fuchsinagars durch den Zusatz von
Kaffein. Centralblatt fur Bakteriologie, Orig. XXXIX, 634.

GAGE, S. D. 1901. Notes on Testing for B. coli in Water. Journal
of Applied Microscopy, IV, 1403.

GAGE, S. D. 1902. Bacteriological Studies at the Lawrence Experi-
ment Station, with Special Reference to the Determination of
B. coli. Thirty-third Annual Report of the State Board of Health
of Massachusetts for 1901, 397.

GAGE, S. D. 1905. Contributions to the Biochemistry of Sewage
Purification; the Bacteriolysis of Peptones and Nitrates. Tech-
nology Quarterly, March, 1905. Also Journal of the American
Chemical Society, XXVII, 327.

GAGE, S. D. 1906. A Study of the Numbers of Bacteria Develop-
ing at different Temperatures and of the Ratios between such
Numbers with Reference to their Significance in the Interpre-
tation of Water Analysis. Biological Studies by the Pupils of
William Thompson Sedgwick, Boston, 1906.

GAGE, S. D. 1910. Methods for Testing Shellfish for Pollution,
Journal of Intectious Diseases, VII, 78.

GAGE, S. D., and ADAMS, G. O. 1904. Studies of Media for the

288 REFERENCES

Quantitative Estimation of Bacteria in Water and Sewage. Jour-
nal of Infectious Diseases, I, 358.
GAGE, S. D., and PHELPS, E. B. 1902. Studies of Media for the

Quantitative Estimation of Bacteria in Water and Sewage.

Transactions of the American Public Health Association. 1901

Meeting. XXVII, 392.
GAGE, S. D., and PHELPS, E. B. 1903. Notes on B. coli and Allied

Forms, with Special Reference to the Neutral Red Reaction.

Transactions of the American Public Health Association. 1902

Meeting. XXVIII, 402.
GARRE, C. 1887. Ueber Antagonisten unter den Bakterien. Cor-

respondenzblatt fur Schweizer Aertze. XVII, 385.
GARTNER, A. 1910. Das Bacterium coli als Indikator fiir fakale Ver-

unreinigung eines Wassers. Zeitschrift fiir Hygiene. LXVII, 55.
GAUTIE, A. 1905. Sur la determination quantitative du colibacille dans

les eaux d'alimentation. Annales de ITnstitut Pasteur, XIX, 124.
GORDAN. 1904. Bakteriologische Untersuchungen zur Beurteilung

von Kleien nach ihrer Neigung zur Schimmelbildung. Rev.

Centralblatt fiir Bakteriologie, Abth. II, XIII, 561.
GORDON, M. H. 1904. Report on a bacterial test for estimating

pollution of air. Supplement to the Thirty-second Annual Report

of the Local Government Board containing the Report of the

Medical Officer for 1902-3.
GORDON, M. H. 1905. Report of some characters by which various

streptococci and staphylococci may be differentiated and identified.

Supplement to the Thirty-third Annual Report of the Local Gov-
ernment Board containing the Report of the Medical Officer for

1903-4, 388.
GORHAM, F. P. 1912. Seasonal Variations in the Bacterial Content

of Oysters. American Journal of Public Health, II, 24.
HALE, F. E., and MELIA, T. W. 1910. Studies on Inhibition, Attenua-
tion and Rejuvenation of B. coli. American Journal of Public

Hygiene, XX, 622.
HALE, F. E., and MELIA, T. W. 1911. The Value of the Esculin Test .

Journal of the American Public Health Association, I, 833.
HAMMERL, H. 1897. Ueber das Vorkommen des Bacterium coli

im Flusswasser. Hygienische Rundschau, VII, 529.
HANKIN, E. H. 1899. On the Detection of the B. typhi abdominalis

in Water and other Substances. Centralblatt fiir Bakteriologie,

XXVI, 554-
HARRISBURG. 1913. Twenty-fifth Annual Report of the Board of

Commissioners of the Water and Lighting Department of the City

of Harrisburg, Pa., for the year 1912.

REFERENCES 289

HARRISON, F. C., and VAN DER LECK, J. 1909. Aesculin Bile Salt

Media for Water Analysis. Centralblatt fiir Bakteriologie, Abth.

I, XXII, 547.
HEIDER, A. 1893. Untersuchungen iiber die Verunreinigungen der

Donau durch die Abwasser der StadtWien. Das 6'sterreichische,

Sanitatswesen. V, 53.
HEKAEUS, W. 1896. Ueber das Verhalten der Bakterien im Brun-

nenwasser, sowie iiber reducirende und oxydirende Eigenschaften

der Bakterien. Zeitschrift fiir Hygiene, I, 193.
HESSE, W., and NIEDNER. 1898. Die Methodik der bakteriolo-

gischen Wasseruntersuchung. Zeitschrift fiir Hygiene, XXIX,

454-

HESSE, G. 1904. Beitrage zur Herstellung von Nahrboden und
zur Bakterienziichtung. Zeitschrift fiir Hygiene, XL VI, i.

HESSE, W., and NIEDNER. 1906. Die quantitative Bestimmung von
Bakterien in Fliissigkeiten. Zeitschrift fiir Hygiene, LIII, 259.

HILGERMANN, R. 1906. Der Nachweis der Typhusbazillen in Wasser
mittels der Eisenfallungsmethoden. Archiv fiir Hygiene, LIX, 355.

HILGERMANN. 1909. Der Wert des Bacillus-coli Befundes zur Beur-
teilung der Reinheit eines Wassers. Der Wert der Eijkmanschen
Garungsprobe. Klinischen Jahrbuch, XXII, 315.

HILL, H. W. 1899. Report of Director of Bacteriological Labo-
ratory. Twenty-seventh Annual Report of the Health Depart-
ment of the City of Boston for the year 1898, 63.

HILL, H. W. 1904. Porous Tops for Petri Dishes. Journal of Medical
Medical Research, XIII, 93.

HILL, H. W., and ELLMS, J. W. 1898. An Apparatus for the Collec-
tion of Samples of Water for Chemical, Microscopical, and Bac-
teriological Analysis. Transactions of the American Public Health
Association. 1897 Meeting. XXIII, 193.

Hiss, P. H. 1902. New and Simple Media for the Differentiation
of the Colonies of Typhoid, Colon and Allied Bacilli. Journal of
Medical Research, VIII, 148.

Hiss, P. H. 1904. On the Fermentative and Agglutinative Char-
acters of Bacilli of the " Dysentery Group." Journal of Medical
Research, XITI, i.

HOFFMANN, W. 1905. Untersuchungen iiber die Lebensdauer von
Typhusbazillen in Aquariumwasser. Archiv fiir Hygiene, LII, 208.

HOFFMANN, WT., und FICKER, M. 1904. Ueber neuer Methoden des
Nachweises von Typhusbacillen. Hygienische Rundschau, XIV, i.

HORHAMMER, C. iQii. Untersuchungen iiber das Verhalten niederer
Krustacean gegeniiber Bakterien im Wasser. Archiv fiir Hygiene,
LXXIII, 183.

290 REFERENCES

HORROCKS, W. H. 1899. On the Behavior of the Bacillus Typhosus
in Sewage. Journal of the Sanitary Institute, XX, 664.

HORROCKS, W. H. 1901. An Introduction to the Bacteriological
Examination of Water. London, 1901.

HORROCKS, W. H. 1903. The Bacillus coli communis; considered
as an Indicator of Sewage Contamination of Water Supplies.
Journal of the Royal Army Medical Corps, I, 362.

HORTON, E. G. 1903. The Colon Bacillus in Ground Waters. Jour-
nal of Hygiene, III, 155.

HOUSTON, A. C. 1898. Chemical and Bacteriological Examination
of Soils with reference to the Amount and Nature of the Organic
Matter and the Number and Character of the Bacteria contained
in them. Supplement to the Twenty-seventh Annual Report of the
Local Government Board, containing the Report of the Medical
Officer for 1897-98, 251.

HOUSTON, A. C. 1899. a. Chemical and Bacteriological Examina-
tion of the " Washings " of Soils, with Reference to the Amount
and Nature of the Organic Matter and the Number and Character
of the Bacteria contained in them. Supplement to the Twenty-
eighth Annual Report of the Local Government Board containing
the Report of the Medical Officer for 1898-99, 439.

HOUSTON, A. C. 1899. b. Bacterioscopic Examination of Drinking-
water with Particular Reference to the Relations of Streptococci
and Staphylococci with Water of this Class. Supplement to the
Twenty-eighth Annual Report of the Local Government Board
containing the Report of the Medical officer for 1898-99, 467.

HOUSTON, A. C. 1900. a. Chemical and Bacteriological Examina-
tion of the " Washings " of Soils, with Reference to the Amount
and Nature of the Organic Matter, and the Number and Char-
acter of the Bacteria contained in them. Supplement to the
Twenty-ninth Annual Report of the Local Government Board con-
taining the Report of the Medical Officer for 1899-1900, 489.

HOUSTON, A. C. 1900. b. On the Value of Examination of Water
for Streptococci and Staphylococci with a view to detection of
its Recent Contamination with Animal Organic Matter. Sup-
plement to the Twenty-ninth Annual Report of the Local Govern-
ment Board containing the Report of the Medical Officer for
1899-1900, 458.

HOUSTON, A. C. 1901. Report on the Chemical and Bacteriological
Examination of Chichester Well-water. Supplement to the
Thirtieth Annual Report of the Local Government Board contain-
ing the Report of the Medical Officer for 1900-01, 511.

HOUSTON, A. C. 1902. a. Bacteriological Standards in relation to

REFERENCES 291

potable and non-potable streams. Second Report of the Royal
Commission on Sewage Disposal, 25.

HOUSTON, A. C. 1902. b. Some of the chief Methods used in the
Bacteriological Examination of Sewage and Effluents. Second
Report of the Royal Commission on Sewage Disposal, 135.

HOUSTON, A. C. 1903. a. Report on the Bacteriological Examina-
tion of the Normal Stools of Healthy Persons. Supplement to
the Thirty-second Annual Report of the Local Government Board
containing the Report of the Medical Officers for 1902-03, 511.

HOUSTON, A. C. 1903. b. Report on the Chemical and Bacterio-
logical Examination of Tunbridge Wells Deep Well Water. Supple-
ment to the Thirty-second Annual Report of the Local Govern-
ment Board containing the Report of the Medical Officer for
1902-03, 581.

HOUSTON, A. C. 1904. a. The Bacteriological Examination of the
River Thames and the Thames Estuary. Fourth Report of the
Royal Commission on Sewage Disposal. Vol. Ill, i.

HOUSTON, A. C. 1904. b. The Inoculation of Sewage with B. pyo-
cyaneus and the subsequent Isolation of this microbe from (lA)
the effluent from a continuous Filter; (2 A) Septic Tank Liquor;
and (26) the Effluent from Contact Beds. Fourth Report of the
Royal Commission on Sewage Disposal, Vol. Ill, 77.

HOUSTON, A. C. 1904. c. 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. Journal of Pathology and Bacteriology,
X, 456.

HOUSTON, A. C. 1904. d. Report on the Bacteriological Examina-
tion of (i) the Normal Stools of Healthy Persons; (2) the Intestinal
Contents of Sea-Fowl and Fish; and (3) Certain of our Public
Water Supplies. Supplement to the Thirty-third Annual Report
of the Local Government Board containing the Report of the
Medical Officer for 1903-04, 472.

HOUSTON, A. C. 1905. a. The bacteriological examination of milk.
Report to the London County Council, July n, 1905.

HOUSTON, A. C. 1905. b. Report on the Bacteriological Examination
of (i) Deep Well Waters; (2) Upland Waters; and (3) Cow-Dung.
Supplement to the Thirty-fourth Annual Report of the Local
Government Board containing the Report of the Medical Officer
for 1904-05, 326.

HOUSTON, A. C. 1906. a. Metropolitan Water Board. Reports for
November and December, 1905.

HOUSTON, A, C, 1906. b. Metropolitan Water-Supply. Monthly

292 REFERENCES

Reports for 1906 of the Water Examiners and of the Director of
Water Examinations.

HOUSTON, A. C. 1908. The Vitality of the Typhoid Bacillus in arti-
ficially infected samples of raw Thames, Lee, and New River
Waters, with special reference to the question of storage. First
Report on Research Work, Metropolitan Water Board, London.
HOUSTON, A. C. 1909. Third Report on Research Work on the Storage
of Raw River Water Antecedent to Filtration, Metropolitan Water
Board, London.

HOUSTON, A. C. 1911. Seventh Report on Research Work, Metro-
politan Water Board, London.

HOWE, E. C. 1912. A Biometric Investigation of Certain Non-spore-
forming Intestinal Bacilli. Science, N. S., XXXV, 225.
HUNTEMULLER, O. 1905. Vernichtung von Bakterien im Wasser

durch Protozoen. Archiv fiir Hygiene, LIV, 89.

IRONS, E. E. 1901. Some Observations on Methods for the Detec-
tion of B. coli communis in Water. Transactions of the American
Public Health Association, 1900 Meeting, XXVI, 310.
IRONS, E. E. 1902. Neutral Red in the Routine Examination of

Water. Journal of Hygiene, II, 314.

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.

JACKSON, D. D. 1907. The Use of Lactose-Bile Medium in Water

Analysis. Journal of Infectious Diseases, Supplement No. 3, 30.

Also Transactions of the American Public Health Association, 1906

Meeting, XXXII, Part II, 30.

JACKSON, D. D. 1911. Classification of the B. coli Group. Journal

of the American Public Health Association, I, 930.
JACKSON, D. D., and MELIA, T. W. 1909. Differential Methods for
Detecting the Typhoid Bacillus in Infected Water and Milk. Jour-
nal of Infectious Diseases, VI, 194.
JACKSON, D. D., and MUER, T. C. 1911. Liver Broth. Journal of

the American Public Health Association, I, 927.
JANOWSKI, T. 1888. Ueber den Bakteriengehalt des Schnees. Cen-

tralblatt fiir Bakteriologie, IV, 547.

JOHNSON, G. A. 1904. Isolation of Bacillus coli communis from the
Alimentary Tract of Fish and the Significance thereof. Journal
of Infectious Diseases, I, 348.
JOHNSON, G. A. 1905. Report on Sewage Purification at Columbus,

Ohio. Columbus, 1905.

JOHNSON, G. A. 1907. Some Notes on Oriental Water-works. Engi-
neering News, LVIII, 481.

REFERENCES 293

JORDAN, E. O. 1890. A Report on Certain Species of Bacteria observed
in Sewage. Special Report of the State Board of Health of Massa-
chusetts on the Purification of Sewage and Water, II, 821.

JORDAN, E. O. 1900. Some Observations upon the Bacterial Self-
purification of Streams. Journal of Experimental Medicine,
V, 271.

JORDAN, E. O. 1901. The Relative Abundance of Bacillus coli com-
munis as an Index of the Self-purification of Streams. Journal
of Hygiene, I, 295.

JORDAN, E. O. 1905. The Persistence of Agglutinability in Typhoid
Bacilli in Water. Journal of Infectious Diseases, Supplement
No. i, 172.

JORDAN, E. O., RUSSELL, H. L., and ZEIT, F. R. 1904. The Lon-
gevity of the Typhoid Bacillus in Water. Journal of Infectious
Diseases, I, 641.

KABRHEL, A. 1906. Studien iiber den Filtrationseffekt der Grund-
wasser. Archiv fur Hygiene, LVIII, 345.

KAISER, M. 1905. Ueber die Bedeutung des Bacterium coli im
Brunnenwasser. Archiv fur Hygiene, LII, 121.

KELLERMAN, K. F., PRATT, R. W., and KIMBERLY, A. E. 1907. The
Disinfection of Sewage Effluents for the Protection of Public Water
Supplies. Bulletin No. 115, Bureau of Plant Industry, U. S.
Department of Agriculture.

KELLERMAN, K. F., and WHITTAKER, H. A. 1909. Farm Water
Supplies of Minnesota. Bulletin 154, Bureau of Plant Industry,
U. S. Department of Agriculture.

KEYES, F. G. 1909. The Gas Production of Bacillus coli. Journal
of Medical Research, XXI, 69.

KINNICUTT, L. P., WINSLOW, C.-E. A., and PRATT, R. W., 1910. Sewage
Disposal. New York.

KISSKALT, X. 1906. Die Verunreinigung der Lahn und der Wieseck
durch die Abwasser der Stadt Giessen, mit besonderen Beriick-
sichtigung der Brauchbarkeit der iiblichen Methoden zur Unter-
suchung von Flussverunreinigungen. Zeitschrift fur Hygiene
LIII, 305.

KLAUMANN, F. 1904. Beitrag zur Frage der Wirkung des Kaffeins
auf Typhus und Colibakterien. Centralblatt fur Bakteriologie,
Orig. XXXVI, 312.

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 rela-
tions with Milk, with Sewage, and with Manure. Supplement to
the Twenty-seventh Annual Report of the Local Government

294 REFERENCES

Board containing the Report of the Medical Officer for
1897-98, 210.

KLEIN, E. 1899. Further Report on the Bacillus enteritidis sporo-
genes. Supplement to the Twenty-eighth Annual Report of the
Local Government Board containing the Report of the Medical
Officer for 1898-99, 312.

KLEIN, E., and HOUSTON, A. C. 1898. Report on Bacteriological
Evidence of Presumably Recent and therefore Dangerous Sewage
Pollution of elsewise Potable Waters. Supplement to the Twenty-
seventh Annual Report of the Local Government Board containing
the Report of the Medical Officer for 1897-98, 318.

KLEIN, E., and HOUSTON, A. C. 1899. Further Report on Bacterio-
logical Evidence of Recent and therefore Dangerous Sewage Pollu-
tion of elsewise Potable Waters. Supplement to the Twenty-
eighth Annual Report of the Local Government Board containing
the Report of the Medical Officer for 1898-99, 498.

KLEIN, E., and HOUSTON, A. C. 1900. Preliminary Account of
the Results of a Bacterioscopic Analysis of Different Cereals and
Foodstuffs. Supplement to the Twenty-ninth Annual Report of
the Local Government Board containing the Report of the Medical
Officer for 1899-1900, 593.

KLOTZ, O. 1904. On a Bacillus Isolated from Water and Aggluti-
nated by High Dilutions of Typhoid Serum. Journal of Medical
Research, XI, 475.

KLOTZ, O, and HOLMAN, W. L., 1911. Infection by the Gas-Bacillus
in Coal Mines. Journal of Infectious Diseases, IX, 251.

KOCH, R. 1893. Ueber den augenblicklichen Stand der bakterio-
logischen Choleradiagnose. Zeitschrift fur Hygiene, XIV, 319.

KOHN, E. 1906. Zur Biologic der Wasserbakterien. Centralblatt
fur Bakteriologie, Abth. II, XV, 690.

KOLLE, W., and GOTSCHLICH, E. 1903. Untersuchungen iiber die
Bakteriologische Choleradiagnostik und Specificitat der Koch-
schen Cholera vibrio. Zeitschrift fur Hygiene, XLIV, i.

KONRADI, D. 1904. Ueber die Lebensdauer pathogenen Bakterien
im Wasser. Centralblatt fur Bakteriologie, Abth. I, Orig.
XXXVI, 203.

KONRICH. 1910. Zur Bewertung des Bacterium coli im Wasser.
Klinisches Jahrbuch, XXIII, i.

KORSCHUN, S. W. 1907. Zur Frage der Verbreitung des Abdominal
typhus durch Trinkwasser. Archiv fiir Hygiene, LXI, 336.

KRANEPUHL. 1907. Beitrage zur Frage der Abwasserdesinfektion
mittels Chlorkalkes. Mitteilungen aus der Koniglichen Prii

REFERENCES 295

fungsanstalt fur Wasserversorgung und Abwasserbeseitigung zu

Berlin, IX, 149.
KRUSE, W. 1894. Kritische und experimentelle Beitrage zur hygieni-

schen Beurtheilung des Wassers. Zeitschrift f iir Hygiene, XVII, i .
KRUSE. 1908. Beitrage zur Hygiene des Wassers. Zeitschrift fur

Hygiene, LIX, 6.
KUBLER and NEUFELD, F. 1899. Ueber einen Befund von Typhus-

bacillen im Brunnenwasser. Zeitschrift fur Hygiene, XXXI, 133.
KURPJUWEIT, O. 1907. Zur Frage der Desinfektion ungereinigter

und gereinigter stadtischer Abwasser mit Chlorkalk. Mitteilun-

gen aus der Koniglichen Priifungsanstalt fiir Wasserversorgung

und abwasserbeseitigung zu Berlin, IX, 162.
LAWS, J. P., and ANDREWES, F. W. 1894. Report on the Result

of Investigations of the Micro-organisms of Sewage. Report to

London County Council, December 13, 1894.
LEDERER, A., and BACHMANN, F. 1911. Fallacies in the Bacterial

Control of a Sewage Purification Plant. Proceedings of the Third

Meeting of the Illinois Water Supply Association, 144.
LE GROS, F. L. 1902. Monographic des Streptocoques et des agents

des septicemies metadiphteriques particulierement des diplocoques.

Paris, 1902.
LEMKE. 1911. Ueber Anreicherung von Typhusbacillen in Wasser.

Deutsche medicinische Wochenschrift, XXXVII, 1698.
LENTZ, O. and TIETZ, J. 1903. Eine Anreichungsmethode fiir Typhus-

und Paratyphusbazillen. Muenchener Medizinische Wochen-
schrift, L., 2139.
LENTZ, O., and TIETZ, J. 1905. Weitere Mittheilungen iiber die

Anreichungsmethode fiir Typhus- und Paratyphusbazillen, mittelst

einer Vorkultur auf Malachitgriinagar. Klinisches Jahrbuch,

XIV, 495-
LEVY, E., and BRUNS, H. 1899. Zur Hygiene des Wassers. Archiv

fiir Hygiene, XXXVI, 178.
LIGNIERES. 1894. Nouveau moyen d'isolement du colibacille. Comptes

rendus de la societe de biologie, 1894, 200.
LOEFFLER. 1893. Zum Nachweis der Cholerabakterien im Wasser.

Centralblatt fiir Bakteriologie, XIII, 380.
LOEFFLER, F. 1903. Neues Verfahren zum Nachweise der Typhus-

bazillen in Faeces, Wasser, Erde. Deutsche Medizinische Wochen-
schrift, XXIX, 36.
LOEFFLER, F. 1906. Der Kulturelle Nachweis der Typhusbazillen

in Faeces, Erde und Wasser mit Hilfe des Malachitgriins. Deutsche

Medizinische Wochenschrift, XXXII, 289.
LONGLEY, F. F., and BATON, W. U. C. 1907. Notes on the Deter-

296 REFERENCES

mination of B. coli in Water. Journal of Infectious Diseases,

IV, 397-
LUBENAU, C. 1907. Das Kaffeinanreichungverfahren zum Typhus-

nachweis im Stuhle. Archiv fur Hygiene, LXI, 232. «

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 Thomp-
son-Yates Laboratories Report, III, 41.
MACCONKEY, A. 1901. Further Note on Bile Salt Lactose Agar.

The Thompson- Yates Laboratories Report, III, 151.
MACCONKEY, A. 1905. Lactose-fermenting Bacteria in Faeces,

Journal of Hygiene, V, 333.
MACCONKEY, A. 1906. A Contribution to the Bacteriology of Milk.

Journal of Hygiene, VI, 3§5-
MACCONKEY, A., and HILL, C. A. 1901. A simple Test for Fascal

Contamination. The Thompson- Yates Laboratories Report, IV,

151-
MAKGILL, R. H. 1901. The Neutral-red Reaction as a means of

detecting Bacillus coli in Water-supplies. Journal of Hygiene,

I, 430.
MARSHALL, W. E. 1907. The Para-Dime thyl-amido-benzaldehyde

Test for Indole. Journal of Hygiene, VII, 581.
MASCHEK, J. 1887. Bakteriologische Untersuchungen des Leit-

meritzer Trinkwasser. Jahresbericht der Oberrealschule zu Leit-

meritz. 1887.
MASSACHUSETTS STATE BOARD OF HEALTH. 1899. Thirtieth Annual

Report, for 1898.
MASSACHUSETTS STATE BOARD OF HEALTH. 1900. Thirty-first Annual

Report, for 1899.
MASSACHUSETTS STATE BOARD OF HEALTH. 1901. Thirty-second

Annual Report, for 1900.
MASSACHUSETTS STATE BOARD OF HEALTH. 1902. Thirty-third

Annual Report, for 1901.
MATHEWS, A. P. 1893. On Wurtz's Method for the Differentiation

of Bacillus typhi abdominalis from Bacillus coli communis and

its Application to the Examination of Contaminated Drinking-
water. Technology Quarterly. VI, 241.
MAYER, G. 1902. Untersuchungen von Wasserlaufen in China.

Centralblatt fur Bakteriologie, Abth. I, Orig., XXXIII, 412.
McGowAN, G., HOUSTON, A. C., and KERSHAW, G. B. 1904. Reports

on Land Treatment of Sewage. Fourth Report of the Royal Com-
mission on Sewage Disposal, Vol. IV.
McWEENEY, E. J. 1904. Report on the Shellfish Layings on the

REFERENCES 297

Irish Coast as Respects their Liability to Sewage Contamination.

British Medical Journal, 1904, II, 497.
MIQUEL, P. 1886. De la richesse en bacteries des eaux d'essangeage.

Revue d'Hygiene, VIII, 388.
MIQUEL, P. 1891. Manuel pratique d'analyse bacteriologique des

eaux. Paris, 1891.
MIQUEL, P., and CAMBIER, R. 1902. Traite de Bacteriologie pure

et appliquee. Paris, 1902.
MOORE, V. A., and WRIGHT, F. R. 1900. A Comparison of B. coli

communis from Different Species of Animals. Journal of the

Boston Society of Medical Sciences, IV, 175.
MOORE, V. A., and WRIGHT, F. R. 1902. Preliminary Observations

on B. coli communis from Certain Species of Animal. Science,

N. S., XV, 372.
MORONI, A. 1898. Bacterium coli commune. Archives per le

Scienze mediche, XXII, 261.
MORONI, A. 1899. La presenza del bacillus coli communis nelle

acque. La Riforma medica, XV, in.
MULLER, O. 1905. Ueber den Nachweis von Typhusbazillen im

Trinkwasser mittels Chemischer Fallungsmethoden, insbesondere

durch Fallung mit Eisenoxychlorid. Zeitschrift fiir Hygiene, LI, i.
MULLER, P. 1900. Ueber die Verwendung des von Hesse und Niedner

empfohlenen Nahrbodens bei der bacteriologischen Wasserunter-

suchung. Archiv fiir Hygiene, XXXVIII, 350.
NEUMANN, G. 1906. Der Nachweis des Bacterium coli in der

Ausserwelt unter Zuhilfnahme der Eijkmannschen Methode.

Archiv fiir Hygiene, LIX, 174.
NEUMANN, G. 1910. Der Nachweis des Bacterium coli in der Aussen-

welt, besonders auf Nahrungsmitteln. Deutsche medicinische

Wochenschrift, XXXVI, 2046.
NEW JERSEY STATE COMMISSION, 1907. Report of the State Sewage

Commission to the Legislature of 1907.
NEWLANDS, J. A., and HAM, G. C. 1910. Report to the State Board

of Health of Connecticut on the Sanitary Investigation of Oyster

Grounds in the New Haven Harbor.
NEW ORLEANS, 1903. Report on Water Purification Investigations

and on Plans Proposed for Sewerage and Water-works Systems.
NIETER, A. Ueber den Nachweis von Typhusbazillen in Trinkwasser

durch Fallung mit Eisenoxychlorid. Hygienische Rundschau,

XVI, 57.
NOWACK, K. 1905. Ueber die Grenzen der Verwendbarkeit des Mala-

chitgriinagars zum Nachweise der Typhusbazillen in Stuhle.

Archiv fiir Hygiene, LIII, 374.

298 REFERENCES

NOWACK, K. 1907. Untersuchungen iiber die Zuverlassigkeit der
Eijkmannschen Probe. Mitteilungen aus der Koniglichen Prii-
fungsanstalt fiir Wasserversorgung und Abwasserbeseitigung zu
Berlin, IX, 197.

OTTO, M., and NEUMANN, R. O. 1904. Ueber einige bakteriologische
Wasseruntersuchungen in Atlantischen Ozean. Centralblatt. fiir
Bakteriologie, Abth. II, XIII, 481.

PAKES, E. C. C. 1900. The Application of Bacteriology to Public
Health. Public Health, XII, 385.

PAPASOTIRIU, J. 1901. Untersuchungen iiber das Vorkommen des
Bact. coli in Teig, Mehl und Getreide, nebst einigen Bemerkungen
iiber die Bedeutung des Bact. coli als Indikator fur Verunreinigung
von Wasser mit Fakalein. Archiv fiir Hygiene, XLI, 204

PARIETTI, E. 1890. Metodo di ricerca del Bacillo del tifo nelle acque
potabili. Rivista d'Igiene e Sanita publica. I, 409.

PARK, W. ?L, COLLINS, K. R., and GOODWIN, M. E. 1904. The
Dysentery Bacillus Group and the Varieties which should be
included in it. Journal of Medical Research, XI, 553.

PECKHAM, A. W. 1897. The influence of Environment upon the
Biological Processes of the Various Members of the Colon Group
of Bacilli. Journal of Experimental Medicine, II, 549.

PENFOLD, W. J. 1912. On the Specificity of Bacterial Mutation,
with a Resume of the Results of an Examination of Bacteria found
in Faeces and Urine, which undergo Mutation when grown on
Lactose Media. Journal of Hygiene, XII, 195.

PENNINGTON, M. E., and KUSEL, G. C. 1900. An Experimental
Study of the Gas-producing Power of Bacillus coli communis
under Different Conditions of Environment. Journal of the
American Chemical Society, XXII, 556.

PENNSYLVANIA, 1908. The Dectection of Bacillus typhosus in the Water
Supplied to the City of Scran ton. Second Annual Report of the
Commissioner of Health of the Commonwealth of Pennsylvania
for 1907, 410.

PETRUSCHKY, J., and PUSCH, H. 1903. Bacterium coli als Indi-
cator fiir Fakalverunreinigung von Wassern. Zeitschrift fiir
Hygiene, XLIII, 304.

PHELPS, E. B. 1907. A Method for Calculating the Numbers of B.
Coli from the Results of Dilution Tests. American Journal of
Public Hygiene, XVIII, 141.

PHELPS, E. B. 1911. Some Experiments upon the Removal of Oysters
from Polluted to Unpolluted Waters. Journal of the American
Public Health Association, I, 305.

REFERENCES 299

PHELPS, E. B., and CARPENTER, W. T. 1906. The Sterilization of

Sewage Filter Effluents. Technology Quarterly, XIX, 382.
PHELPS, E. B., and HAMMOND, F. S. 1909. A Study of Certain Para-
colon Forms found in Polluted Deep Wells. American Journal of

Public Hygiene, XIX, 545.
PHILBRICK, B. G. 1905. Changes in the Bacterial Content of Water

in passing through a Distributing Reservoir. Journal of Medical

Research, XIII, 419-
POUJOL, M. G. 1897. Sur la presence tres frequent du Bacterium

coli dans les eaux naturelles. Comptes rendus de la Societe de

Biologic, IV, 982.
PRAUSNITZ, W. 1890. Der Einfluss der Miinchener Kanalisation

auf die Isar mit besonderer Berticksichtigung der Frage der Selb-

streinigung der Fliisse. Munich, 1890.
PRESCOTT, S. C. 1902. a. On the Apparent Identity of the Cultural

Reactions of B. coli communis and certain Lactic Bacteria. Science,

N. S., XV, 363-
PRESCOTT, S. C. 1902. b. A Note on Methods of Isolating Colon

Bacilli. Science, N. S., XVI, 671.
PRESCOTT, S. C. 1903. On Certain Precautions Required in Making

and Interpreting the so-called " Colon Test " for Potable Waters.

Medicine, XI, 20.
PRESCOTT, S. C. 1906. The Occurrence of Organisms of Sanitary

Significance on Grains. Biological Studies by the Pupils of William

Thompson Sedgwick, Boston, 1906.
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. Journal

of Infectious Diseases, I, 193. Also, Transactions of the American

Public Health Association. 1903 Meeting, XXIX, 369.
PROCACCINI, R. 1893. Influenza della luce solare sulle acque di

rifiuto. Annali dell' Istituto d'Igiene Sperimentale di Roma,

HI, 437-
RAPP, R. 1903. Ueber den Einfluss des Lichtes auf organische Sub-

stanzen mit besonderer Beriicksichtigung der Selbstreinigung der

Fliisse. Archiv fur Hgyiene, XLVIII, 179.
REFIK. 1896. Sur les divers types de coli-bacille des eaux. Annales

de 1'Institut Pasteur, X, 242.
REINSCH, A. 1891. Zur bakteriologischen Untersuchung des Trink-

wassers. Centralblatt fur Bakteriologie^ X, 415.
REMLINGER, P., and SCHNEIDER, G. 1897. Contribution a 1'etude

du bacille typhique. Annales de ITnstitut Pasteur, XI, 55.
REYNOLDS, A. R. 1902. Report of Streams Examination, Chemic

300 REFERENCES

and Bacteriologic, of the Waters between Lake Michigan at Chicago
and the Mississippi River at St. Louis, for the purpose of determin-
ing their condition and quality before and after the Opening of
the Drainage Channel. Chicago, 1902.

RICKARDS, B. R. 1907. Pollution of Shellfish by Sewage. Thirty-
fifth Annual Report of the Health Department of the City of Boston,
for the year 1906, 80.

RIDEAL, S. 1902. Water and its Purification. London, 1902.

RIDEAL, S. 1906. Sewage and the Bacterial Purification of Sewage.
Third Edition, London and New York, 1906.

ROTH, E. 1903. Versuche iiber die Einwirkung des Kaffeins auf
das Bakterium Typhi und Coli. Hygienische Rundschau, XIII,
489.

ROTHBERGER, C. J. 1898. Differential diagnostische Untersuchungen
mit gefarbten Nahrboden. Centralblatt fur Bakteriologie, Abth.
I, XXIV, 513-

RUEDIGER, G. F. 1911. Studies on the Self-purification of Streams.
Journal of the American Public Health Association, I, 411.

RUSSELL, H. L. 1891. Untersuchungen iiber im Golf von Neapel
lebende Bakterien. Zeitschrift fiir Hygiene, XI, 166.

RUSSELL, H. L. 1892. Bacterial Investigation of the Sea and its
Floor. Botanical Gazette, XVII, 312.

RUSSELL, H. L., and BASSETT, V. H. 1899. The Significance of
Certain Gas-forming Bacteria of Non-colon Type in Sanitary
Water Analysis. Proceedings of American Public Health Asso-
ciation, XXV, 1899 Meeting, 570.

RUSSELL, H. L., and FULLER, C. A. 1906. The Longevity of Bacillus
typhosus in Natural Waters and in Sewage. Journal of Infectious
Diseases, Supplement No. 2, 40. Also Transactions of the Ameri-
can Public Health Association, 1905 Meeting, XXXI, Part II.

SAVAGE, W. G. 1901. Neutral Red in the Routine Bacteriological
Examination of Water. Journal of Hygiene, I, 437.

SAVAGE, W. G. 1902. The Significance of Bacillus coli in Drinking
Water. Journal of Hygiene, II, 320.

SAVAGE, W. G. 1903. The Pathogenicity of B. coli in Relation to
the Bacteriological Examination of Water. Journal of Hygiene,
III, 388.

SAVAGE, W. . G. 1905. Bacteriological Examination of Tidal Mud
as an Index of Pollution of the River. Journal of Hygiene, V, 146.

SAVAGE, W. G. 1906. The Bacteriological Examination of Water
Supplies. Philadelphia, 1906.

SAWIN, L. R. 1907. Experience with Lactose-Bile Medium for the -
Detection of B, coli in Water, Journal of Infectious Diseases,

REFERENCES 301

Supplement No. 3, 33. Also Transactions of the American Public
Health Association, 1906 Meeting, XXXII, Part II, 33.

SAWIN, L. R. 1910. Pancreatin-bile salt medium for the detection
of B. coli in water. American Journal of Public Health, XX, 672.

SCHEPILEWSKI, E. 1903. Ueber den Nachweis der Typhusbakterien
nach der Methode von Dr. A. W. Windelbandt. Centralblatt fur
Bakteriologie, XXXIII, 394.

SCHEURLEN. 1891. Ueber die Wirkung des Centrifugirens auf Bak-
teriensuspensionen, besonders auf die Vertheilung der Bakterien in
der Milch. Arbeiten aus dem Kaiserlichen Gesundheitsamte, VII,
269.

SCHOTTELIUS, M. 1885. Zum mikroskopischen Nachweis von Cholera-
bacillen in Dejectionen. Deutsche medicinische Wochenschrift,
XI, 213.

SCHUDER. 1903. Zum Nachweis der Typhusbacillen in Wasser.
Zeitschrift fur Hygiene, XLII, 317.

ScHULTZ-ScHULTSENSTEiN. 1903. Ueber die nitrificierenden Mikro-
organismen der Filterkorper biologischer Abwasser-Reinigungs-
anlagen. Mitteilungen a. d. K. Priifungsanstalt fur Wasser-
versorgung und Abwasserbeseitigung zu Berlin, II, i.

SCHUMACHER. 1905. Die Desinfektion von Krankenhausgruben mit
besondere Beriicksichtigung des Chlorkalkes und ihre Kontrolle.
Gesundheits-Ingenieur, XXVIII, 361.

SEDGWICK, W. T. 1889. Recent Progress in Biological Water Analysis.
Journal of the New England Water- works Association, IV, i.

SEDGWICK, W. T. 1902. Principles of Sanitary Science. New York,
1902.

SEDGWICK, W. T., and PRESCOTT, S. C. 1895. a. On the Bacterial
Contents of Certain Ground- waters, including Deep-Wells.
Twenty-sixth Annual Report of the State Board of Health of Massa-
chusetts for 1894, 435.

SEDGWICK, W. T., and PRESCOTT, S. C. 1895. b. On the Influence
of Slight Variations in the Composition of Nutrient Gelatin upon
the Development of Water Bacteria. Transactions of the American
Public Health Association, 1894 Meeting, XX, 450.

SEDGWICK, W. T., and WINSLOW, C.-E. A. 1902. Experiments on
the Effect of Freezing and other Low Temperatures upon the
Viability of the Bacillus of Typhoid Fever, with Considerations
regarding Ice as a Vehicle of Infectious Disease. Memoirs of the
American Academy of Arts and Sciences, XII, 467.

SHUTTLEWORTH, E. B. 1895. Some Deductions from Bacteriological
Work on the Water of Lake Ontario. Transactions of the American
Public Health Association, 1894 Meeting, XX, 58.

302 REFERENCES

SMITH, T. 1893. a. 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.
SMITH, T. 1893. b. The Fermentation-tube with Special Reference

of Anaerobiosis and Gas Production among Bacteria. The Wilder

Quarter-century Book, 187.
SMITH, T. 1895. Ueber den Nachweis des Bacillus coli communis

im Wasser. Centralblatt fur Bakteriologie, XVIII, 494.
SPITTA. 1903. Weitere Untersuchungen iiber Flussverunreinigung.

Archiv fur Hygiene, XL VI, 64.
SPITTA and IMHOFF. 1906. Apparate zur Entnahme von Wasser-

proben. Mitteilungen aus der Koniglichen Priifungsanstalt fur

Wasserversorgung und Abwasserbeseitigung zu Berlin, VI, 57.
STAMM, J. 1906. Ueber die Bedeutung des von einigen pathogenen

Bakterien der Typhus- Coli- Gruppe unter anaeroben Bedingun-

gen produzierten Gases fur die Differentialdiagnose. Centralblatt

fur Bakteriologie, Abth. I, Orig., XLII, 590.
STARKEY, T. A. 1906. A method for the Isolation of Typhoid and

Colon Bacilli from Drinking Waters, etc. American Journal of

the Medical Sciences, CXXXII, 109.
STARKEY, T. A. 1909. New Bacteriological Standard for Water

Analysis. Journal of the Royal Sanitary Institute, XXX, 391.
STARKEY, T. A. 1911. Typhoid-like Infections. Journal of the Royal

Sanitary Institute, XXXII, 467.
STERNBERG, G. M. 1892. Manual of Bacteriology. New York,

1892.
STILES, G. W. 1911. The Bacteriological Examinations of Shucked

and Shell Oysters. Journal of the American Public Health Asso-
ciation, I, 623.
STILES, G. W. 1912. Sewage-polluted Oysters as a Cause of Typhoid

and other Gastro-intestinal Disturbances. Bulletin No. 156, Bureau

of Chemistry, U. S. Department of Agriculture.
STOKES, W. R. 1904. A Simple Test for the Routine Detection of

the Colon Bacillus in Drinking Water. Journal of Infectious

Diseases, I, 341.
STOKES, W. R., and HACHTEL, F. W. 1910. The Control of Typhoid

Fever in City and Country. Archives of Internal Medicine, VI,

121.
STOKES, W. R., and HACHTEL, F. W. 1912. The Use of a Modified

Hesse's Medium for Isolating the Typhoid Bacillus and the Cholera

Spirillum from Stools. American Journal of Public Health, II,

182.
STOKES, W. R., and STONER, H. W. 1909. Dextrose vs. Lactose for

REFERENCES 303

Detecting the Colon Bacillus. American Journal of Public Hygiene,
XIX, 312.

STOKVIS, C. S., and SWELLENGREBEL, N. H. 1911. Purification of
Water by Infusoria. Journal of Hygiene, XI, 481.

STOUGHTON, G. VAN E. 1905. Characteristics of Colon Bacilli and
the Value of the Presumptive Test. Journal of Infectious Dis-
eases, Supplement No. 2, 147.

STROSZNER. 1904. Typhus bacillen in dem Wasser eines Haus-
brunnens. Centralblatt fiir Bakteriologie, XXXVIII, 19.

THOMANN, J. 1907. Zum Nachweis des Bacterium coli commune
im Wasser vermittels der Eijkmannschen Methode. Hygienische
Rundschau, XVII, 857.

THOMSON, T. 1894. Report on an Epidemic of Enteric Fever in
the Borough of Worthing and in the Villages of Broadwater and
West Tarring. Supplement to the Twenty-third Annual Report
of the Local Government Board, containing the Report of the
Medical Officer for 1893-4, 47.

THRESH, J. C. 1903. Report on an Extensive Epidemic of Diarrhoea
Apparently Due to the Pollution of a Public Water Supply.
British Medical Journal, 1903, II, 760.

THRESH, J. C. 1904. The Examination of Waters and Water-Sup-
plies. London, 1904.

THUMM, K. and PRITZKOW, A. 1903. Versuche iiber die Reinigung
der Abwasser von Tempelhof bei Berlin durch das biologische
Verfahren. Mitteilungen aus der Koniglichen Priifungsanstalt
fiir Wasserversorgung und Abwasserbeseitigung zu Berlin, II, 127.

TIEMANN, F., and GARTNER, A. 1889. Die chemische und mikro-
scopisch-bakteriologische Untersuchung des Wassers. Bruns-
wick, 1889.

TOBEY, E. N. 1906. The Methods for Testing the Indol Reaction.
Journal of Medical Research, XV, 301.

TULLY, E. J. 1912. The Hygiene of the Swimming Pool. American
Journal of Public Health, II, 186.

TWORT, F. W. 1907. The Fermentation of Glucosides by Bacteria
of the Typhoid-coli Group and the Acquisition of New Fermenting
Powers by Bacillus dysenteriae and other Micro-organisms. Pro-
ceedings of the Royal Society, Series B, LXXIX, 329.

VALLET, G. Une nouvelle technique pour la recherche du bacille
typhique dans les eaux de boissons. Archives de Medicine Experi-
mentale, XIII, 557.

VAN BUSKIRK, L. H. 1912. Outfits for Chemical and Bacteriological
Field Determinations. Monthly Bulletin, Ohio State Board of
Health, 11,483.

304 REFERENCES

VINCENT, H. 1905. Sur la signification du " Bacillus coli " dans
les eaux potables. Annales de ITnstitut Pasteur, XIX, 233.

VINCENT, H. 1907. Recherches sur les microbes anaerobies des eaux.
Annales de 1'Institut Pasteur, XXI, 62.

WALKER, L. R. 1901. Bacteriological Investigations of the Iowa
State College Sewage. Proceedings of the Iowa Academy of
Sciences, 1901.

WATHELET, M. A. 1895. Recherches bacteriologiques sur les dejec-
tions dans la fievre typhoide. Annales de 1'Institut Pasteur,
IX, 251.

WEISSENFELD, J. 1900. Der Befund des Bacterium coli im Wasser
und das Thierexperiment sind keine brauchbare Hiilfsmittel fiir
die Hygienische Beurtheilung des Wassers. Zeitschrift fiir Hygiene,
XXXV, 78.

WELCH, W. H., and NUTTALL, G. H. F. 1892. A Gas-producing
Bacillus (Bacillus aerogenes capsulatus, N. S.) Capable of Rapid
Development in the Blood Vessels after Death. Bulletin of the
Johns Hopkins Hospital, July, 1892.

WEST, F. D. 1909. Notes on the Voges and Proskauer Reaction for
Bacillus coli communis. American Journal of Public Hygiene,
XIX, 227.

WHERRY, W. B. 1905. A Search into the Nitrate and Nitrite Con-
tent of Witte's " Peptone " with special Reference to its Influence
on the Demonstration of the Indol and Cholera Red Reactions.
Journal of Infectious Diseases, II, 436.

WHIPPLE, G. C. 1896. Discussion of Paper on Water-borne Dis-
eases, by T. Smith. Journal of the New England Waterworks,
Association, X, 217.

WHIPPLE, G. C. 1899. On the Necessity of Cultivating Water Bacteria
in an Atmosphere Saturated with Moisture. Technology Quarterly,
XII, 276.

WHIPPLE, G. C. 1901. Changes that take place in the Bacterial
Contents of Water during Transportation. Technology Quarterly,

XIV, 21.

WHIPPLE, G. C. 1902. On the Physical Properties of Gelatin with
Reference to its Use in Culture Media. Technology Quarterly,

XV, 127.

WHIPPLE, G. C. 1903. On the Practical Value of Presumptive Tests
for Bacillus coli in Water. Technology Quarterly, XVI, 18.

WHIPPLE, G. C. 1907. Quality of Kennebec River Water. Water
Supply and Irrigation Paper, No. 198, N. S. Geological Survey,
167.

KEFEKENCES 305

WHIPPLE, G. C. 1913. The 37° Bacterial Count. American Journal
of Public Health, III, 36.

WHIPPLE, G. C., and MAYER, A. 1906. a. On the Relation between
Oxygen in Water and the Longevity of the Typhoid Bacillus.
Journal of Infectious Diseases, Supplement No. 2, 72.

WHIPPLE, G. C., and MAYER, A. 1906. b. On the Relation between
Oxygen in Water and the Longevity of the Typhoid Bacillus.
Journal of Infectious Diseases, Supplement No. 3, 76. Transac-
tions of the American Public Health Association, 1905 Meeting,
XXXI, Part II.

WIDAL, F. 1896. Serodiagnostic de la fievre typhoide. La Semaine
Medicale, XVI, 259.

WILHELMI, J. 191 1. Die Einleitung der Abwasser in das Meer.
Wasser und Abwasser, IV, 177 and 221.

WILLSON, H. S. 1905. The Isolation of B. Typhosus from Infected
Water, with Notes on a New Process. Journal of Hygiene, V, 429.

WINSLOW, C.-E. A. 1901. The bacteriological Examination of
Water and its Interpretation. Journal of the New England Water-
works Association, XV, 459.

WINSLOW, C.-E. A. 1903. a. Studies on Quantitative Variations in
Gas Production in the Fermentation-tube. Science, N. S., XVII,

375-

WINSLOW, C.-E. A. 1903. b. The Occurrence of the Colon Bacillus
on the Hands. Journal of Medical Research, X, 463.

WINSLOW, C.-E. A. 1905. The Number of Bacteria in Sewage and
Sewage Effluents determined by Plating upon different Media
and by a New Method of Direct Microscopic Enumeration. Jour-
nal of Infectious Diseases, Supplement No. i, 209; Contributions
from the Sanitary Research Laboratory and Sewage Experiment
Station of the Massachusetts Institute of Technology, I, 41.

WINSLOW, C.-E. A., and BELCHER, D. M. 1904. Changes in the
Bacterial Flora of Sewage during Storage. Journal of Infectious
Diseases, I, 170.

WINSLOW, C.-E. A., and HUNNEWELL, M. P. 1902. a. Streptococci
Characteristic of Sewage and Sewage-polluted Waters, Science,
N. S., XV, 827.

WINSLOW, C.-E. A., and HUNNEWELL, M. P. 1902. b. A Study of
the Distribution of the Colon Bacillus of Escherich and of the
Sewage Streptococci of Houston in Polluted and Unpolluted Waters.
Journal of Medical Research, VIII, 502.

WINSLOW, C.-E. A., and KLIGLER, I. J. 1912. A Quantitative Study
of the Bacteria in City Dust with Special Reference to Intestinal
and Buccal Forms. American Journal of Public Health, II, 663.

306 REFERENCES

WINSLOW, C.-E. A., and LOCHRIDGE, E. E. 1906. The Toxic Effect
of certain Acids upon Typhoid and Colon Bacilli in Relation to
the Degree of their Dissociation. Journal of Infectious Diseases,
HI, 547-

WINSLOW, C.-E. A., and PALMER, G. T. 1910. A Comparative Study
of Intestinal Streptococci from the Horse, the Cow and Man.
Journal of Infectious Diseases, VII, i.

WINSLOW, C.-E. A., and PHELPS, E. B. 1905. The Chemical and
Bacterial Composition of the Sewage discharged into Boston
Harbor from the South Metropolitan District. Journal of Infec-
tious Diseases, Supplement No. i, 175; Contributions from the
Sanitary Research Laboratory and Sewage Experiment Station
of the Massachusetts Institute of Technology, I, 7.

WINSLOW, C.-E. A., and PHELPS, E. B. 1907. Investigations on
the Purification of Boston Sewage in Septic Tanks and Trickling
Filters (1905-1907). Technology Quarterly, XX, 387.

WINSLOW, C.-E. A., and WALKER, L. T. 1907. Note on the Fer-
mentative Reactions of the B. coli Group. Science, N. S., XXVI,

797-
WINSLOW, C.-E. A., and WINSLOW, A. R. 1908. The Systematic

Relationships of the Coccaceae. New York.
ZIKES, H. 1902. Zur Kenntnis der chemischen und biologischen

Schwankungen im Gehalte der Brunnenwasser. Mitteil. d. Oester.

Versuchstationen fiir Brauerei und Malzerei. Wien, 1902. H.

10. Rev. Hygienische Rundschau, XIII, 830.

INDEX OF AUTHORS

Abba, 144, 161.

Calmette, 235, 241.

Abbott, 36.

Cambier, 36, 47.

Adami, 81.

Cameron, 245.

Adams, 30, 44, 45, 46.

Carpenter, 237.

Altschuler, 83.

Chick, 65, 105, 153.

American Committee on Standard

Chopin, 81.

Methods, 136, 137, 193, 208.

Christian, 120.

Amyot, 142.

Clark, H. W., 57, 72, m, 146,

157, 162, 166, 242, 248, 251.

Baker, 203, 204.

Clark, W. M., 118, 199.

Barthel, 147.

Clemesha, 17, 21, 108, 114, H5>

Baton, 107, in, 117.

184, 193, 210.

Bachmann, 64, 231.

Cramer, 54.

Beckmann, 143.

Conn, 245.

Belcher, 213.

Conradi, 76.

Belitzer, 140.

Copeland, 102, 105, 192.

Bergey, 180, 191.

Bettencourt, 141, 142, 147, 172.

Biffi, 176.

Davis, 118, 199.

Blachstein, 144.

Deehan, 180, 191.

Blunt, 16.

Dibdin, 228.

Bolton, 53.

Ditthorn, 85.

Borges, 141, 142, 147.

Doebert, 77.

Boullanger, 241.

Dolt, 130.

Braun, 122.

Downes, 16.

Brezina, 55.

Drew, 13, 36.

Brotzu, 140.

Drigalski, 76, 85.

Brown, 151.

Duclaux, 6.

Bmns, 144.

Duggeli, 147-

Buchan, 248.

Dunbar, 97.

Buchner, 17.

Dunham, 56, 96.

Bulstrode, 225, 248.

Durham, 93, 95.

Burri, 150, 183.

Dyar, 140.

307

308

INDEX OF AUTHORS

Eddy, 232.

Egger, 26.

Eijkman, 120.

Ellms, 36.

Eisner, 75.

Endo, 76.

English Committee, 136.

Escherich, 99.

Eyre, 140.

Fehrs, 15.

Ferguson, 116.

Ferreira, 102, 141.

Ficker, 80, 84.

Fischer, 2, 26, 90.

Flatau, 90.

Fox, 91.

Frankland, 12, 37, 98.

Fremlin, 140.

Freudenreich, 16, 150.

Fromme, 113, 119, 121, 142, 163,

171, 225.
Frost, 1 6, 179.
Fuller, 16, 32, 43, 57, 116, 182.

Gaehtgens, 77.

Gage, 8, 9, 30, 31, 43, 44, 45, 46,
69, 70, 72, 107, 108, 109, in,
122, 133, 146, 157, 162, 166,
177, 186, 187, 203, 218, 230,
242, 249.

Garre, 15.

Gartner, 14, 35, 53, 159, 170, 171.

Gautie, 172.

Gildemeister, 85.

Gordan, 147, 209.

Gorham, 249, 254, 256.

Hachtel, 79, 91, 97.
Hale, 116, 126, 127, 129.
Ham, 244, 249, 250, 264.
Hammerl, 151.
Hammond, 113, 115, 116.

Hankin, 75.

Hansen, 221.

Harrison, 129.

Hazen, 19.

Heider, 226.

Heraeus, 35.

Hesse, 29, 45, 47.

Hilgermann, 120.

Hill, 36, 64.

Hiss, 78, 89.

Hoffmann, 24, 80.

Holman, 213.

Hoover, 102, 192.

Horder, 199, 209.

Horhammer, 15. .

Horrocks, 15, 26, 87, 153, 182, 203.
207, 212.

Horta, 102, 141.

Houston, n, 21, 22, 24, 27, 96,
115, 141, 146, 150, 152, 158,
159, 162, 178, 185, 189, 197,

202, 207, 209, 213, 222, 236,
239-

Howe, 175, 177, 180-199.
Hunnewell, 115, 118, 203.
Huntemiiller, 15.

Irons, 106, 109, 120, 122.

Jackson, 64, 78, 81, 91, 122, 123,

131, 192.
Janowski, 6.

Johnson, 58, 142, 182, 235, 232.
Jordan, 14, 16, 17, 20. 23, 39, 55,

153, 155, 226.

Kabrhel, 25.

Kaiser, 163.

Keith, 140.

Kellerman, 26, 238, 240.

Kimberly, 238, 240.

Kinnicutt, 234, 236.

Kisskalt, 9.

INDEX OF AUTHORS

309

Kloumann, 80.

Klein, 75, 81, 146, 211, 213.

Kligler, 148.

Klotz, 82, 213.

Koch, 97.

Kohn, 38, 53.

Konradi, 24.

Konrich, 101, 102, 121, 141, 147,

148, 163, 169, 170, 171, 176,

183, 187, 188.
Korschun, 15.
Kranepuhl, 239
Kruse, 143, 169.
Kiibler, 90.
Kurpjuweit, 239.

Laws, 92, 202.

Lederer, 64, 231.

LeGros, 203.

Lemke, 77.

Lentz, 77.

Levy, 144.

Lochridge, 19.

Loeffler, 77, 96.

Longley, 107, in, 117, 165.

Lubenau, 87.

MacConkey, 122, 175, 177, 178,

180, 184, 189, 190, 191.
Makgill, 122.
Marshall, 176.
Marvel, 246.
Maschek, 26, 35.
Mass. State Board of Health, 136,

161.

Massol, 241.
Massini, 183.
Mathews, 66, 105.
Mayer, A., 18, 19.
Mayer, G., 56.
McWeeney, 175.

Melia, 81, 91, 116, 126, 127, 129.
Miquel, 6, 37,47, Si, 55-

Moore, 140.

Moroni, 144.

Muer, 131.

Muller, 30, 31, 36, 84, 183.

Neufeld, 90.
Neuman, 13, 120, 147.
Newlands, 244, 249, 250, 264.
Nibecker, 12, 67, 68, in, 133, 204.
Niedner, 45, 47.
Nieter, 85.
Nowack, 77, 121.
Nuttall, 211.

Orlandi, 161.
Osgood, 226.
Otto, 13.

Pakes, 153.

Palmer, 209.

Papasotiriu, 146.

Paredes, 102, 141.

Parietti, 74.

Park, 89.

Pease, 249.

Peckham, 182.

Penfold, 183.

Petruschky, 168.

Phelps, 30, 31, 43, 44, 71, in, 113,

115, 116, 122, 133, 177, 186,

187, 233, 237, 249, 252.
Philbrick, 10, n, 54.
Poujol, 144.

Pratt, 234, 236, 238, 240.
Prescott, 12, 26, 43, 67, 107, 146,

203, 204.
Procaccini, 17.
Pusch, 168.

Rapp, 17, 18.
Refik, 143.
Reinsch, 43.
Remlinger, 93.

310

INDEX OF AUTHORS

168

Reynolds, 119.
Rickards, 248.
Rideal, 66, 229, 237, 242.
Riedel, 38.
Rivas, 1 80.
Rogers, 118, 199.
Rondelli, 161.
Roth, 80.
Rothberger, 122.
Ruediger, 22, 23.
Russell, 13, 16.

Sauerbeck, 183.

Savage, 24, 62, 122, 141, 145

214, 225.
Sawin, 124, 125.
Schepilewski, 83.
Scheurlen, 14.
Schneider, 93.
Schottelius, 96.
Schuder, 84.

Schultz-Schultzenstein, 241.
Schumacher, 239.
Sedgwick, 4, 21, 43, 95, 105.
Shell Fish Commission, A.P.H.A

255-

Shuttleworth, 59.
Smith, 101, 106, 140, 150, 180.
Soper, 246.
Spitta, 14.
Stamm, 117.
Starkey, 86, 95, 96.
Sternberg, 52.
Stiles, 248.

Stokes, 79, 91, 97, 122, 127.
Stokvis, 15.
Stoner, 127.
Swellengrebel, 15.

Thoman, 75.
Thorne-Thorne, 246.
Thomson, 75.
Thresh, 36, 95, 246.
Tiemann, 14, 35.
Tietz, 77.
Tully, 73.
Twort, 183.

Vallet, 84.

Van Buskirk, 50.

van der Leek, 129.

Vincent, 56, 172, 213, 223.

Walker, 101, 147, 180, 232.

Wathelet, 92.

Weissenfeld, 145.

Welch, 211.

West, 180.

Wheeler, 18.

Wherry, 178.

Whipple, 18, 19, 38, 39, 41, 44, 46,
63, 71, 72, no, 118, 181, 221.

Whittaker, 26.

Widal, 89.

Wilhelmi, 248.

Willson, 80, 84, 90.

Winslow, 12, 19, 21, 29, 30, 34,
60, 66, 67, 68, 71, 101, in, 118,
133, 147, 148, 155, !8o, 199- 203,
209, 213, 221, 232, 233, 234, 236.

Wolffhugel, 26, 38.

Woodman, 221.

Wright, 140.

Wurtz, 65, 104.

Zagari, 15.
Zeit, 16, 23.

SUBJECT INDEX

Acid-forming organisms, 65.
Acid wastes, antiseptic effect of,

20.

Aesculin, 129.
Aesculin bile medium, 129.
Aesculin medium, preparation of,

279.

Aesculin test, 129.
Agar, drying of, 78.
Agar, for body temperature count,

63.

Agar, preparation of, 270.
Agglutination of typhoid bacilli,

81, 82, 83.

Anaerobic bacteria, 201.
Anaerobic spore-forming bacilli,

211.

Anglo-American procedure, 172.
Aniline dyes, use of, 77.
Antagonism, 16.

Anthrax, occurrence in water, 98.
Apparatus, treatment of, 279.
Arbitrary standards, fallacy of, 51.
Atmospheric waters, 5.
Atypical colon bacilli, 135.
Azolitmin, 272.

Bacillus, anaerobic spore-forming,

211.

Bacillus acidi lactici, 189.
Bacillus aerogenes, 189.
Bacillus alcaligenes, 94.

Bacillus anthracis, 98.
Bacillus coli, 94, 99.

cultural features, 100.

discovery, 99.

distribution in waters, 152.

effect of temperature on, 113.

fermentation reactions, 101.

importance of numbers, 149.

index of pollution, 219.

in cold-blooded animals, 142.

in ground waters, 113.

in sewage, 230.

in soils, 152.

isolation of, 104, 132.

isolation by Conradi-Drigalski
medium, 106.

isolation by Endo medium, 106.

isolation by synthetic media,
130.

mutations in, 183.

occurrence, 99.

occurrence in animals, 140, 141.

occurrence on plants, 146.

pathogenicity, 100.

positive isolations, no.

preliminary enrichment of, 106.

quantitative test, 138.

standard tests for, 115.

streak characteristics, 134.

ubiquity of, 143.

Bacillus coli and sewage strepto-
cocci, 205, 206.

311

312

SUBJECT INDEX

B. communis, 181.
B. communior, 181.
B. dysenteriae, 89, 94.
B. enteritidis, 94.
B. mycoides, 133.
B. sporogenes, 211.

characteristics of, 213.

growth in liver broth, 212.

in sewage, 213.

isolation of, 211.

occurrence of, 212.
B. typhi, 94.

identification of, 87.

comparison with B. coli, 87.

in oysters, 91.

isolation from water, 89, 90.

isolation of, 90.

small numbers in water, 92.
B. welchii, 129, 134, 138, 211.
Bacteria, affected by temperature,

20.

as agents of decomposition, 3.

count of as index of efficiency
of niters, 59.

counts of on different media, 44.

development at high tempera-
tures, 69.

distribution, i.

diurnal variation of in sewage,

233-
effect of composition of media

on, 19.

effect of light, 16, 17.
effect of rainfall on, 7.
effect of storage on, 10.
estimation of, 48.
expression of counts of, 49.
factors influencing diminution,

13-

field determinations of, 49.
food requirements, 2.
in contact effluents, 240.
in disinfected effluents, 240.

Bacteria in driven wells, 27.
in dust and air, 6.
in earth, 6.
in filtered waters, 57.
in ground waters, 25.
in lakes and ponds, 12.
in ocean, 13.
in oysters, seasonal variation of,

254, 256.

in polluted streams, 7.
in rain and snow, 6.
in sand effluents, 234, 240.
in septic effluents, 240.
in sewage, 232.
in sewage effluents, 233.
in shallow wells, 26.
in springs, 26.
in surface waters, 54.
in trickling filter effluents, 236.
in unpolluted streams, 7.
in water, 5.

in water and shellfish, 249, 250.
metatrophic, 2.

microscopic enumeration of, 29.
mineral nutrients for, 53.
multiplication in stored waters,

37-

nitrifying, 4.

number of as index of purity, 60.

numbers of in sewage, 232.

occurrence, i.

paratrophic, 2.

pathogenic in water, 74.

proto trophic, 2, 29.

relation to character of food, 29.

quantitative methods of deter-
mination, 29.

seasonal variation of, 7.

sedimentation of, 14.
Bacteriological examination of

shellfish, 244, 248.
Bacteriological examination of
sewage, methods of, 229.

SUBJECT INDEX

313

Bacteriological examinations of

water, significance of, 217.
Bacteriological methods for super-
vision of filtration plants, 227.
Bacteriological methods for super-
vision of water supplies, 227.
Bacteriological methods in detect-
ing sewage distribution, 226.
Bacteriology of sewage, 228.
Bacteriology of sewage effluents,

228.

Bacteriology of sewage filters, 241.
Bacteriological examination, 220.
advantages of, 220.
certainty of, 223.
delicacy of, 221.
Bile, importance of, 127.
Bile media, 81, 122.
Bile salts, 122.

selective action of, 123, 128.

Body temperature count, 42, 61.

relation to hot weather, 70.

Caffein, selective action of, 80.
Calcium hypochlorite, 238.
Carbon dioxide, absorption of,

117.

Chemical disinfection, 238.
Chlorine disinfection, relation to

counts, 72.

Cholera red reaction, 97.
Cholera spirillum, isolation from
water, 96.

media for, 96.
Clams, 244, 263.
Cold, action of on bacteria, 21.
Colon bacilli, 64, 99.

as index of pollution, 168.

as index of self-purification,

153-

atypical, 135, 177.
colonies of, 133.
comparison with B. typhi, 87.

Colon bacilli, distribution in
waters, 152.

effect of temperature on, 22.

" flaginac," 185.

importance of numbers, 149.

increased survival in cold
weather, 22.

in cultivated soil, 170.

in dust, 148.

in filter effluents, 166.

in filtered waters, 163.

in foods, 147.

in fruits, 147.

in grains, 146.

in ground waters, 161.

in sewage, 231.

in sewage effluents, 231.

in shallow wells, 162.

in shellfish, 252.

in soil, 148.

in surface waters, 157.

in unpolluted waters, 155.

isolation of, 104.

isolation by bile media, 122.

on plants, 146.

standard tests for, 115.

" typical," 176.

ubiquity of, 143.

varieties of, 174.

viability at different tempera-
tures, 128.
Colon group, 99.

characteristics, 104.

distribution in water, 185.

Jackson's classification of, 192.

McConkey's classification of,
189.

statistical classification, 198.

tests for, 174.

variations in, 181.
Colon test, 101.
Colon typhoid group, 94.

reactions of, 95.

314

SUBJECT INDEX

Comparison of sand and mechan-
ical niters, 58.

Composition of medium, impor-
tance of, 43.
Confirmatory tests, 136.
Conradi-Drigalski medium, 76.

for B. coli, 106.

preparation of, 275.
Contact beds, 235.
Counting, 48.
Crystal violet, 76.
Culture media, ingredients for,
265.

preparation of, 265.

reaction of, 267.

sterilization of, 266.

titration of, 267.

uniform methods for, 265.

Deep wells, bacteria in, 27.
Dextrose broth, 107.

advantages of, 108.

comparison with lactose bile,
125.

disadvantages of, 108.
Dextrose test, failure of, 116.
. Diluting samples, 40.
Disinfection of sewage, 237.
Disinfection of sewage effluents,

237-
Distribution of types of colon

group in waters, 184.
Division of colon group, 1 74.
Dunham's solution, 97.
Dysentery, spread by water, 95.

Eijkman test, 120.
Endo medium, 76.

for B. coli, 106.

preparation of, 276.
Examination of shellfish, standard

methods for, 255, 257.
Examination of shell water, 251.

Expression of quantitative re-
sults, 49.

Fermentation of lactose, 114.
Fermentation test, 105, 109, 179.

effect of temperature on, 112.

exceptions to, in.

interpretation of, no.
Field kits, 49.
Field methods, 50.
Filter plants, routine control of,

230.

Filtered waters, bacteria in, 57.
Filtration in Japan, 58.
" Flaginac " B. coli, 185.
Food supply, importance of, 20.

Gartner bacillus, 95.
Gas-forming bacteria, growth in

liver broth, 131.
Gas production in vacuo, 118.
Gas ratio, 100, 109.

effect of age on, 117.

unreliability of, 118.
Gelatin liquefaction, 177.
Gelatin plates, use of, 41.
Gelatin, preparation of, 270.
Green plants, food requirements,

Ground waters, 6.
bacterial content of, 56.
B. coli in, 113.

Hesse medium, 78.

drying of, 78.

preparation of, 275.
High temperatures, significance

of, 69.
Hiss agar medium, 78.

preparation of, 277.
Hog cholera bacillus, 94.

Incubation, 46, 280.

SUBJECT INDEX

315

Incubation period, 41, 47.

for body temperature count,

64.
Incubator, necessity for moisture

in, 46.

Indol test, 176.
Infusoria, destruction of bacteria

by, 15-

Interpretation of results, 51.
Intestinal bacilli, 103, 201.
Isolation of B. coli, 104, 132.

by bile media, 122.
Isolation of cholera spirillum, 97.
Isolation of streptococci, 203.

Lactose bile, 81, 122, 126.
action of bacteria on, 102.
advantage of, 127.
decomposition of, 64.

comparison with dextrose

broth, 125.

preparation of, 269.
Lactose fermentation, 114.

importance of, 115.
Lactose fermenting bacilli, 102.

effect of storage on, 115.
" Lamirascsal," streptococci, 210.
" Larasacsal " streptococci, 210.
Light, destructive effect on bac-
teria, 16, 17.
Litmus lactose agar, 64, 104.

counts, 65.

incubation of, 132.

preparation of, 272.
Liver agar, preparation of, 274.

broth, 131.

preparation of, 273.
Liver gelatin, preparation of, 274.

Malachite green agar, 77.
McConkey's group in filtered
water, 198.

McConkey's groups in raw waters,

198.
McConkey's groups in stored

waters, 198.

Mechanical filtration, 57.
Methods of estimation, 48.
Middletown outbreak, 245.
Milk, preparation of, 278.
Mussels, 244.
Mutations in B. coli, 183.

Nahrstoff agar, ratio, 31.
Nahrstoff gelatin, ratio, 31.
Nahrstoff-Heyden agar, 29.
Neutral red, 122.
Neutral red reaction, 121.
Neutral red test, 178.
Nitrate broth, preparation of, 287.
Nitrates, 3.
Nitrite test, 177.
Nitrogen cycle, 4.
Nitroso-indol reaction, 97.
Nutrient broth, preparation of,

268.
Nutrose, 76.

Obligate parasites, 29.
Overgrowth, effect of, 119.
Oysters, 244.

B. coli in, 259.

bacterial counts of, 258.

bacteriological examination of,

257-

opened, examination of, 262.
rating for B. coli, 260.
sampling, 257.
Oysters and typhoid, 246.

Para-colon bacilli, 187.
Para- typhoid bacilli, 94.

316

SUBJECT INDEX

Pathogenes in water, 74.

Pathogenic bacteria, 74.

Peptone, importance of, 44.

Petri dishes, 105.

Phenol agar, 105.

Phenol broth, 119.

Phenol dextrose broth, 108.

Plate method, 105.

Plating, 40.

Polluted shellfish, effect of cook-
ery on, 248.

Polluted waters, isolation of B.
coli from, 107.

Pollution, progressive, 223.
of shellfish, 245.
temporary, 225.

Porous tops, 64, 105.

Precipitation of typhoid bacilli,
81, 83, 84.

Preliminary enrichment, 106.

Presumptive test, no.

Presumptive tests, various, 126.

Presumptive tests with bile media,
124.

Progressive pollution, detection
of, 223.

Prototrophic bacteria, 29.

Protozoa, reduction of bacteria
by, 15.

Pumping, effect of on bacteria, 35.

Quantitative bacteriological de-
termination, 29.
interpretation of, 51.

Quantitative results, expression
of, 49.

Rainfall, effect of on bacteria, 7.
Reaction, importance of, 43.
Reaction of culture media, 267.
Reaction optimum, 43.

Reducing bacteria, importance of,
242.

Relation between room tempera-
ture and body temperature
counts, 61, 67, 71.

Room temperature counts, 61.

Saccharose, fermentation of, 180.
Samples, dilution of, 40.
Samples, icing of, 39.
Sampling, 33.
Sand nitration, 57.
Sanitary chemical analysis, sig-
nificance of, 216.
Sanitary inspection, 215.
importance of, 172.
Seasonal variation, 72.
Sedimentation, 7.
Sedimentation of bacteria, 14.
Self-purification, 20.
Selective media, 219.
Selective temperatures, 219.
Sewage, bacteria in, 214.

bacteriology of, 228.

colon bacilli in, 231.
Sewage effluents, bacteriology of,
228.

colon bacilli in, 231.

standards for, 239.
Sewage examination, use of bile

medium in, 231.
Sewage sampling, error of, 231.
streptococci, 134, 201.

as index of pollution, 202.

isolation of, 204.

Sewage streptococci and B. coli,
growth in dextrose broth, 205,
206.

detection of, 206.
Sewage and sewage effluents, 228,

229.
Shallow wells, bacteria in, 26.

SUBJECT INDEX

317

Shallow wells, body temperature

count in, 67.
Shellfish and disease, 244.

bacteriological examination of,
248, 251.

bacteria in shucked, 254.

careless handling of, 253.

colon bacilli in, 252.

self-purification of, 252.

standards of interpretation, 264.

streptococci in, 252.
Shell water, examination of, 251.
Significance of 37° count, 62.
Specific sewage organisms, tests

for, 230.

Springs, bacteria in, 26.
Standard msthods, 32, 33, 41.

necessity of, 45.
Standard reaction, 268.
Standards for sewage effluents,

239-

Staphylococci in sewage, 202.
Storage, effectiveness of, 25.

effect of on bacteria, 10, 21.

effect of duration of, 23.

effect on lactose fermenters, 115.
Storage of samples, effect of, 37.
Stored waters, 6.
Streptococci, 64, 133.

antagonism to colon bacilli, 208. |

comparative fermentations by,
210.

from different animals, 209.

in sewage, 202.

in saliva, 204.

in shellfish, 252.

in polluted waters, 203.

in stored sewage, 207.

indicative of recent pollution,
207.

index of pollution, 220.

isolation of, 203.

on animal bodies, 204.

Streptococci, varieties of, 208,

209.

Streptococci, " lamiracsal," 210.
Streptococci, " larasacsal," 210.
Streptococcus equinus, 209.
Sugar broths, preparation of, 269.
Sugar reactions, 179.
Sugars, action of bacteria on, 102.
Surface waters, 5.

bacterial content of, 54.
Swimming pools, bacteria in, 72.
Synthetic media, 130.

Temperature, effect on B. coli, 113.

effect on bacteria in water, 20.

effect on fermentation test, 112.
Temporary pollution, detection of,

225.

Titration of culture media, 267.
Toxic products, effect of, 15.
Trickling filters, 235.
Typhoid, occurrence in cold

weather, 22.

Typhoid and shellfish, 245.
Typhoid bacilli, agglutination of,
81, 82, 83.

artificial infection of water
with, 24.

developing on malachite green
media, 77.

effect of oxygen on, 19.

enrichment in caffein media, 79.

enrichment of, 75.

examination of water for, 74.

in polluted waters, 23.

in pure culture, 16.

in tap water, 24.

in unsterilized waters, 23.

isolation by lactose bile, 81.

isolation of, 7.6.

media for, 76, 77, 78, 79.

precipitation of, 81, 83, 84.

318

SUBJECT INDEX

Typhoid, preliminary enrichment
of, 79.

separation by motility, 85.

small numbers in water, 92.

summary of isolation methods,
86.

uncultivated strains in water, 24.

viability in mud, 24.

viability in sewage, 16.

viability in water, '15.
" Typical" B. coli, 176.

Unpolluted waters, body tempera-
ture count in, 68.
Urea, decomposition of, 3.

Voges-Proskauer reaction, 180.

Waters, classification of, 5.
Wells, bacteria in, 26, 56.

B. coli in, 26, 162, 163.

deep, 27.

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Elements of water bacteriology, with special reference to sanitary water analysis"

See other formats