CAMBRIDGE PUBLIC HEALTH SERIES

UNDER THE EDITORSHIP OF
G. S. Graham-Smith, M.D. and J. E. Purvis, M.A.

University Lecturer in Hygiene and University Lecturer in Chemistry
Secretary to the Sub-Syndicate for and Physics in their application
Tropical Medicine to Hygiene and Preventive Medi-

cine, and Secretary to the State
Medicine Syndicate

THE BACTERIOLOGICAL EXAMINATION
OF FOOD AND WATER

CAMBRIDGE UNIVERSITY PRESS

C. F. CLAY, MANAGER
Itontion: FETTER LANE, E.G.

TOO PRINCES STREET

!LontJ0n:

H. K. LEWIS, 136, GOWER STREET, W.C.
WILLIAM WESLEY & SON, 28, ESSEX STREET, STRAND

$rf» gorfe: G. P. PUTNAM'S SONS
Calcutta an* fHaUras: MACMILLAN AND CO., LTD.
^Toronto: J. M. DENT AND SONS, LTD.
THE MARUZEN-KABUSHIKI-KAISHA

A II rights reserved

THE

BACTERIOLOGICAL EXAMINATION
OF FOOD AND WATER

WILLIAM G. SAVAGE

B.Sc., M.D. (Lond.), D.P.H.

County Medical Officer of Health, Somerset
Late Medical Officer of Health and Public Analyst, Colchester

Lecturer on Bacteriology, University College, Cardiff

Assistant in charge of the Bacteriological Department, University College, London

etc. etc.

Cambridge :

at the University Press

1916

I /\ — '

First Edition 1914
Second Edition 1916

EDITORS' PREFACE

IN view of the increasing importance of the study of public
hygiene and the recognition by doctors, teachers, adminis-
trators and members of Public Health and Hygiene Committees
alike that the salus populi must rest, in part at least, upon a
scientific basis, the Syndics of the Cambridge University Press
have decided to publish a series of volumes dealing with the
various subjects connected with Public Health.

The books included in the Series present in a useful and
handy form the knowledge now available in many branches
of the subject. They are written by experts, and the authors
are occupied, or have been occupied, either in investigations
connected with the various themes or in their application and
administration. They include the latest scientific and practical
information offered in a manner which is not too technical.
The bibliographies contain references to the literature of each
subject which will ensure their utility to the specialist.

It has been the desire of the editors to arrange that the books
should appeal to various classes of readers : and it is hoped that
they will be useful to the medical profession at home and abroad,
to bacteriologists and laboratory students, to municipal engineers
and architects, to medical officers of health and sanitary in-
spectors and to teachers and administrators.

Many of the volumes will contain material which will be
suggestive and instructive to members of Public Health and
Hygiene Committees ; and it is intended that they shall seek
to influence the large body of educated and intelligent public
opinion interested in the problems of public health.

405779

PREFACE

TEXT-BOOKS and Manuals upon Bacteriology are, for the
most part, written by Bacteriologists whose investigations
and routine work are in the field of Pathological Bacteriology.
The descriptions in these text-books of methods suitable for
the examination of morbid human materials and for the study
of the pathogenic bacteria leave little to be desired on the
score of either lucidity or completeness.

In marked contrast is the inadequate treatment given to
the bacteriological examination of water, air, foods and the
like. The description of their examination is either relegated
to a few pages at the end of the volume, where their treatment
is both scanty and incomplete, or, not infrequently, is omitted
altogether. This branch of bacteriology is of immense practical
importance and justifies a more extended treatment.

The aim of this volume is to remedy this defect and to
make available a practical manual dealing not only with the
examination of these substances but also with the deductions
to be drawn from the bacteriological data obtained from their
examination. Much of the available information is only at
present to be found in original papers not always readily
accessible.

The methods selected are, in general, those which I have
found to be of practical value and proved utility, while special
attention has been given to the difficult matter of the deductions

viii PREFACE

to be drawn from bacteriological examinations, and how far
they may be used as a basis of administrative action.

Figures I and 16 and figures 8, 9, 10, n and 13 are taken
from my books The Bacteriological Examination of Water
Supplies and Milk and the Public Health respectively, and 1 am
greatly indebted to Mr H. K. Lewis and Messrs Macmillan
and Co. for their permission to use these illustrations.

I have also to thank Messrs W. B. Saunders and Co. for
figure 3 (from Eyre's Bacteriological Technique), Messrs Long-
mans, Green and Co. for figure 12 (from Curtis's Essentials of
Practical Bacteriology), Messrs Baird and Tatlock for figures
5 and 15, and Messrs A. Gallenkamp and Co. for figure 7.

WILLIAM G. SAVAGE.

WESTON-SUP.ER-MARE.
February 1914.

CONTENTS

CHAP. PAGE

I. General considerations ....... I

II. General methods for the isolation and identification of

Indicator Organisms ...... 16

III. Water 27

IV. Soil and Sewage 61

V. Shellfish 74

VI. Milk .......... 85

VII. Modified milk and milk products in

VIII. The bacteriology of meat and meat products . . 122

IX. Air . . . • 136

X. The determination of antiseptic and germicidal power 148

Appendix .- - 158

Addendum - 171

Index . 189

LIST OF ILLUSTRATIONS

FIG. PAGE

1. Bacillus enteritidis sporogenes. Typical milk cultures, after 24

hours' incubation at 37° C. . . . . . . 10

2. Bent glass distributing rod 20

3. Apparatus arranged for filtering— aspiration . . . 24

4. Collecting bottle and tin . . . . . . . 29

5. Savage's plate-cooling apparatus . . . • . . 33

6. Wilson's apparatus for concentration under reduced pressure 41

7. Fraenkel's borer . . . . . . . •..•** 68

8. Collecting bottle and ice box . 86

9. Dilution of milk samples . . . . . , . 88

10. B. enteritidis sporogenes enumeration jar. . . ... 94

11. Apparatus for estimating milk sediments 96

12. Experimental tuberculosis in a guinea-pig (about the third week

after inoculation) inoculated subcutaneously in neighbour-
hood of the left knee-joint ... . . . 102

13. Centrifugal tube for leucocyte enumeration .... 106

14. Agar plates exposed for 30 minutes in a school . . 141

15. Sedgwick-Tucker tube . . t 143

1 6. Frankland's tube ~ 144

CHAPTER I

GENERAL CONSIDERATIONS

Although water, food, air, etc. are not all examined bacterio-
logically with the same objects, broadly there is a similarity of
procedure and certain general considerations may be conveniently
considered together, in this way avoiding repetition in the
chapters dealing with the examination of the different substances.

Bacteriological examinations of food and other substances
are undertaken essentially for one or other, or for all, of the
following purposes.

(a) To examine for the presence of definite disease-pro-
ducing organisms. This is done either with the object of
detecting their presence or, on the negative side, to judge by
the failure to find them, whether they are absent or, if at one
time present, have been eliminated.

(b) To measure the extent to which the substance under
examination has been polluted by material derived from
undesirable or harmful sources.

(c) To assess the value and completeness of any purification
processes to which the substance under examination has been
subjected.

The examination for specific disease-producing organisms is
a procedure of great importance but is of limited applicability
and practical utility in connection with the examination of water
and foods. The detection of pathogenic bacteria in the human
or animal body and in many of their excretions is, with modern

s. w. i

-CONSIDERATIONS

methods, "o'ftfen'iat.prcicsdtiifeVof-'no great trouble or complexity.
On the other hand their isolation from water, food, air, etc. is
frequently a matter of the utmost difficulty for reasons which
readily present themselves. In these substances any pathogenic
bacteria are frequently present in but small numbers compared
with the total number of bacteria present, while these competing
bacilli, by growing more readily and abundantly in the media
used for the isolation of the pathogenic bacteria, may materially
enhance the difficulties of isolation. Further, there is always the
possibility that any pathogenic bacteria originally present may
have been completely eliminated by the time attention is called
to the need for bacteriological examination. These pathogenic
bacteria in water, soil or even in milk are, or may be, in an
environment either in itself unsuitable to their multiplication
and sustained life or in one relatively unsuitable compared to
that enjoyed by the horde of naturally present, or added,
competing bacilli. This may cause their actual or relative
disappearance when the food or water is submitted to bacterio-
logical examination. For these reasons the isolation of specific
pathogenic bacteria plays but a small part in this class of
bacteriological analysis.

The second object of a bacteriological examination — to
measure undesirable bacterial contamination — is by far the
most valuable part of the bacteriological examination of the
materials under consideration. By modern methods it is
possible not only to say whether such pollution exists but,
with considerable approach to accuracy, to determine the extent
of such pollution and to furnish data as to the degree of its
harmfulness.

Although the substances to be examined are various while
the sources of pollution are many and not all of the same
character there is for most of them a broad relationship as to
the nature of the pollution to be detected and measured. For
example, for water supplies bacteriological analysis is mainly
concerned with the detection and estimation of excretal and
sewage pollution. For shellfish the degree of sewage pollution
is again the object of the examination. For milk, apart from
the detection of pathogenic bacteria, the examination is chiefly

GENERAL CONSIDERATIONS 3

to study bacterial contamination mainly from cow manure or
excreta-laden dust. Although the objects of the bacteriological
examination of air are somewhat different it is in general true
to state that excreta (human and animal) and sewage are the
most important forms of pollution which the Public Health
Bacteriologist has to detect and estimate in connection with
his work. To do this the bacteriologist has to determine the
organisms which are prevalent in sewage and excreta and which
can be used by him to estimate the extent of the contamination
by these substances.

These indicator organisms are of great importance. Those
which have been most widely used for this purpose are the
B. coli group, streptococci, and B. enteritidis sporogenes and
closely allied anaerobic bacteria. It will be convenient to
consider first the general and special characters of these groups
of bacteria and subsequently to deal with the grounds on which
they have been selected as suitable to measure excretal and
sewage contamination.

B. coli group.

A study of the characters and distribution of this group is of
the utmost importance to the bacteriologist who has to deal with
the examination of water, milk and other foods. Difficulty is
met with at the outset in defining the characters of the group
and it is easy to err on the one hand by adopting too compre-
hensive a definition, including organisms essentially different
and with different distributions in nature, and on the other, by
limiting the group by too narrow a characterisation, excluding
organisms of material importance and significance.

All the members of the group are small non-sporing bacilli,
decolorised by Gram's method of staining, which grow well at
both 37° C. and 20° C, which ferment both glucose and lactose,
with the production of acid and gas, and which fail to liquefy
gelatine (within a reasonable period). Motility is present or can
be demonstrated by the use of proper media and technique but
it is not always exhibited and its absence in ordinary examina-
tions cannot be taken as excluding organisms from the group.

1—2

4 GENERAL CONSIDERATIONS

Such a characterisation is a very wide one and will include
a large number of organisms which certainly are not all alike,
while many of them will not be identical with the B. coli
communis of Escherich. On the other hand these characters
will separate the group from the Proteus, Gaertner, Dysentery
groups, etc.

All the members of this large group have not an equal
significance as indicators of excretal contamination so that it
is important to carry the classification further.

The cultural characters which have been employed for the
differentiation of the bacilli of this group are very numerous but
in regard to a number of them experience has shown that no
differentiation of value is likely to result from their employment
and they may be disregarded. Of the differentiation tests still
considered by many bacteriologists to be of value the following
are the most important : — Production of indol, growth in litmus
milk, characters of the growth on gelatine slope, production of
fluorescence in neutral red media, Vosges and Proskauer's
reaction, and the fermentation of various sugars and alcohols.

The value of these tests can be best gauged not only by their
ability to adequately differentiate the group as a whole but also
by the light they throw upon the distribution of the different
members of the group. If by their use differences of distribution
are made manifest, so that the finding of the one variety indicates
one kind of contamination while the presence of another variety
points to a different source of contamination it is obvious that
the tests which enable this to be done are of the utmost value.

Excreta — human and animal — is the most important form of
pollution for which this group is used as a test and the value of
these differentiating reactions must largely depend upon how far
they satisfactorily differentiate the strains found in excreta from
those which have not recently been derived from the animal
intestine.

The characters of B. coli as found in excreta have been
investigated by several workers of whom the following may
be mentioned.

Houston1 studied the characters of 101 Coli-like organisms
1 Local Government Board, Medical Officer's Report^ 1902-3, p. 511.

GENERAL CONSIDERATIONS 5

isolated from normal stools of healthy persons. Of these 8 did
not ferment lactose although they fermented glucose. Of the
remaining 93, 72 were completely positive as regards fermenta-
tion of glucose and lactose, positive neutral red reaction, pro-
duction of indol, acid and clot in litmus milk and reduction of
nitrates, 12 were non-motile, otherwise typical, 2 produced no
indol as their only abnormality, 2 failed to give the neutral red
reaction, 4 were typical except that they did not clot milk.
That is leaving out motility as a distinguishing character
84 per cent, were completely positive in their characters.

Of these organisms about half fermented dulcite while
considerably less than half fermented saccharose.

Houston1 obtained very similar results with 60 B. coli
organisms isolated from fresh cow-dung. 95 per cent, gave
acid or clot in milk, 967 per cent, gave indol, 967 per cent,
acid and gas in lactose media, 98-3 per cent, a positive neutral
red reaction.

53 out of the 60 (88 per cent.) were quite typical as regards
all these tests.

MacConkey2 records the examination of 178 samples of
human faeces and 131 samples of animal excreta, derived from
the horse, calf, goat and pig. Ignoring for the moment most of
the sugar-alcohol tests which were also employed, and excluding
a few yellow liquefying organisms certainly not members of the
group the results obtained may be tabulated as follows : —

Strains isolated Percentages

Human Animal Human Animal

Milk clotted ... ... 178 126 100 100

Indol formed 164 122 92 96*8

Gelatine not liquefied... 175 124 98*2 98*4

Saccharose fermented 76 56 42 44*4

,, not fermented 103 70 58 55^6

All the organisms fermented both glucose and lactose.

MacConkey strongly advocates the use of fermentation tests
including lactose, saccharose, dulcite, adonite, inulin, inosit and
mannite, and the abandonment of many of the old tests. In

* Ibid. 1904-5, p. 358.

2 Journ. of Hygiene, 1909, IX, p. 86.

6 GENERAL CONSIDERATIONS

this way he has differentiated the strains isolated from different
sources into a large number of different varieties. His lines of
classification have been followed by other workers. The value
of their employment must depend upon the light they shed upon
the differential origin of the different strains. For example, to
demonstrate that the fermentation of saccharose test is of value
it has to be shown that organisms which ferment this sugar
have a greater or a lesser significance than those which do not
ferment saccharose. If all B. coli in excreta, or even a pre-
ponderance of them, fermented saccharose, then the presence of
a B. coli in water or food which was without this property wrould
have but little significance as an indicator of excreta. Nothing
of the sort, however, has as yet been established, and the same
may be said for the other sugar-alcohol tests, and interesting
as these tests are and valuable for research work, it cannot be
said that their employment for routine work adds information
of material value.

The testing of the pathogenicity of isolated strains of B. coli
has not proved of any practical utility for public health work.

For routine practical work it is important to have some
definite and easily carried out system for recording B. colt
organisms and noting how far they conform to the type most
frequently met with in sewage. In 1905 the writer suggested1
that the expression excretal B. coli should be used for organisms
giving all the following characters :

A short rounded bacillus.

Translucent non-corrugated growth on gelatin slope.

Non-liquefaction of gelatin (two weeks).

Acid production in litmus milk with coagulation (within two-
weeks).

Fermentation of lactose with production of acid and gas.
„ glucose

Neutral red reaction (in glucose media).

Production of indol in peptone water.

Since practically all organisms which ferment lactose also
ferment glucose the fermentation of glucose may be omitted>

1 Lancet, Feb. 4, 1905.

GENERAL CONSIDERATIONS 7

while the neutral red reaction is not of much assistance : other-
wise these tests are all valuable and satisfactory.

Used in this way the term excretal B. colt would imply without
further description that all these characters were present. These
characters are certainly those possessed by the vast majority of
B. colt isolated from excreta and the term is a much more con-
venient one than typical since to the latter so many different
interpretations have been given. If any characters are negative,
the fact can be mentioned in brackets. Thus an organism
having all these attributes, except that it gives no indol reaction,
can be readily written excretal B. coli (indol — ).

Streptococcus group.

The term streptococcus being only a description of a mor-
phological type, must obviously include a number of different
organisms, although Marmorek and other investigators hold that
all human streptococci are identical. Many different kinds of
streptococci have been described and much discussion has taken
place as to how far the described varieties are separate species
or mere variants of a common type. The methods in use to
classify streptococci are very numerous, but it cannot be said
that their differentiation by any method is completely satis-
factory nor is the basis of classification uniform for different
workers.

Some investigators rely solely upon the characters most used
in earlier bacteriological investigations, such as morphology,
pigment-production, characters of the gelatine plate colonies,
growth in broth, milk, potato, etc. Other workers place in
addition to, or in substitution for, these tests, great reliance
upon pathogenicity, agglutination tests, or the haemolytic tests
introduced by Schottmuller. A further group of workers base
their differentiating characters in large part upon the ability of
streptococci to produce acid in certain sugars or alcohols.

It cannot be said that any one series of tests is satisfactory
or sufficient, but speaking generally the sugar-alcohol tests,
together with morphology, growth in broth and milk, and patho-
genicity are in the writer's experience of most utility.

8 GENERAL CONSIDERATIONS

It should, however, be stated that several investigators have
questioned the reliability and value of these sugar-alcohol tests
and have not found them to be sufficiently stable.

These tests were introduced by Gordon1 and included the
following nine tests : Litmus milk for clot (three days at 37° C.),
reduction of neutral red broth during anaerobic incubation for
two days at 37° C., production of acid in three days at 37° C. in
slightly alkaline broth containing respectively I per cent, of
saccharose or lactose, raffinose, inulin, salicin, coniferin, or
mannite. With the exception of the use of coniferin these tests
have been extensively employed.

Houston in 1902 studied the biological characters of 300
human faecal streptococci2, in 1903 the characters of 100
streptococci from cow-dung3, and in 1908 he investigated the
characters of 100 streptococci isolated from lumps of faeces
taken at a sewage outfall works4.

These 500 streptococci showed the following percentage
positive reactions:

Human

Test Faeces Cow-dung "Sewage"

Salicin 927 93 97

Saccharose 86-3 89 49

Lactose ... ... ... 76^3 85 ico

Litmus milk ... ... 617 73 ico

Neutral red broth /.. 39*3 o

Raffinose ... ... ... 32 74 ico

Mannite ... ... ... 24*3 o 4

Inulin 47 13 —

The biological characters of streptococci from excreta have
also been studied by Winslow and Palmer5. They examined
302 streptococci, 116 from human, 100 from horse, 86 from cow
excreta. They employed similar methods but the amount of
acid was in each case estimated by titration, using phenol-
phthalein as indicator. Each culture was incubated 72 hours at

1 Local Government Board, Medical Officers Report, 1903-4, p. 388.

2 Ibid. 1903-4, p. 484.

3 Ibid. 1904-5, p. 326.

4 Metropolitan Water Board, Fifth Research Report*
6 Jonrn. of Inf. Diseases, 1910, vol. vn, p. i.

GENERAL CONSIDERATIONS 9

37° C. Only glucose, lactose, raffinose and mannite were em-
ployed. Their percentages of positive results were as follows :

Human Equine Bovine

Glucose 89 84 65

Lactose ... ... 62 8 52

Raffinose ... 6 4 28

Mannite ... ... 28 2 6

Streptococci from sources other than excreta tested in the
same way do not show differences sufficiently distinctive to
enable the excretal origin of any streptococcus to be ascertained
from an examination of its biological properties. v

For the examination of water, soil, etc. the differentiation of
the isolated streptococci has not up to the present proved to be
of any immediate practical value.

A classification of streptococci on this basis has on the other
hand proved to be of value in differentiating certain human
streptococci and for the identification of streptococci in air.
The classification adopted is discussed in Chapter IX.

B. enteritidis sporogenes.

This bacillus was isolated in October, 1895, by Klein from
the intestinal discharges of patients at St Bartholomew's Hospital
suffering from an epidemic of diarrhoea. It was also isolated
from the milk supplied to these patients.

It is a strictly anaerobic organism with fairly definite char-
acters, of which the most important are the following : A fairly
large bacillus r6 to 4-8 p long and about 0*8 //, thick. Motile,
and stains by Gram's method. It readily forms spores, which
are usually present near the ends of the rods.

Under anaerobic conditions it grows well in milk, glucose
agar, and upon blood serum and other media. Milk cultures are
especially characteristic, and this medium is usually used for its
isolation. The milk is coagulated and separated into a stringy
material (coagulated casein) and a clear or somewhat turbid
whey, a scattered layer of cream remaining on the top, which is
broken up by the gas which is readily liberated by shaking the
tube. The whey is acid, and the tube-contents smell of butyric
acid. The whey examined microscopically shows the bacilli.

10 GENERAL CONSIDERATIONS

The above cultural and microscopic characters are very
similar to those of B. butyricusy and probably some other
organisms, so that it is necessary to make inoculation experi-
ments if it is required to accurately differentiate these bacilli,
B. enteritidis sporogenes being highly pathogenic, and B. butyricus
not pathogenic.

Fig. i. Bacillus enteritidis sporogenes. Typical milk cultures,

after 24 hours' incubation at 37° C.
(From photographs kindly lent by Dr Klein.)

Klein recommends that I c.c. of the whey from the milk
culture be injected subcutaneously into the groin of a guinea-pig.
The animal will usually be found dead within twenty-four hours.
On making a post-mortem examination the hairs near the site
of inoculation are readily stripped off, the skin underneath being
green and gangrenous. Beneath the gangrenous area extensive
sloughing of the subcutaneous tissue has taken place, and a con-
siderable blood-stained, evil-smelling fluid is present, containing
numerous bacilli with the characters given above. The bacilli

GENERAL CONSIDERATIONS II

never show the smooth unseptate filaments characteristic of
the bacillus of malignant cedema. If cultures of diminished
virulence are used the pathogenic effects are somewhat modified.
The virulence of this organism is subject to considerable varia-
tion, to some extent dependent upon its source.

The changes in milk cultures and the demonstration of the
pathogenicity of the whey, along with the post-mortem appear-
ances described, are quite sufficient for the identification of this
organism.

Since B. butyricus and the other culturally nearly identical
organisms have a very similar distribution in Nature, it is not
usually considered necessary to carry out the pathogenicity
test in every instance. The diagnosis is then based upon the
characteristic "enteritidis change" in the milk.

Bacterial indicators of excretal and sewage^utamination.

These three groups — B. colt, streptococci ^B B. enteritidis
sporogenes — are very extensively employed to estimate the extent
of excretal or sewage contamination and it is of great importance
to consider critically the value of these determinations for this
purpose.

The conditions of a perfect bacterial indicator are fairly-
definite and obvious. They are :

1. It should be abundant in the substances, for which its
presence serves as an indicator.

2. It should be absent, or at least relatively absent, from all
other sources.

3. It should be easily isolated and numerically estimated.

4. Its characteristics should be definite and not liable to
variation, whereby its distinctive characters might be impaired.

All three groups are extremely abundant in both human
and animal excreta and in sewage.

As regards human excreta B. coli group organisms are
present to the extent of 100 million to 1000 million or more
per gramme, streptococci are usually equally numerous, while
about I million to 10 million B. enteritidis sporogenes are present
per gramme.

12

GENERAL CONSIDERATIONS

These organisms also occur in immense numbers in the
intestines of all the domestic animals and apparently of all
mammals. The following table shows some results obtained
by the writer with quite fresh specimens of animal excreta.

Source

B. coli

Streptococci.

B. enteritidis
sporogenes Spores

Approximate

number per Gramme

D£ Excreta

Horse No. i

over i million

TV to i million

10 to TOO

, No. i

TV to i million

over i million

100 to 1000

, No. 3

1,000 to 10,000

over i million

100 to 1000

Cow No. i

•fa to i million

10,000 to 100,000

100 tO 1000

, No. 2

10,000 to 100,000

TV to i million

to to 100

, No. 3

i to 10 millions

over 10 millions

10 tO 100

. No. 4

i to 10 millions

iV to i million

100 tO 1000

Pig No. i

over 100 millions

absent

TV to i million

,, No. 2

10 to 100 millions

absent

10,000 to 100,000

„ No. 3

70 millions

absent

1000 to 10,000

Sheep No. i

ro to 100 millions

i to 10 millions

10 to too

„ No. 2

10 to 100 millions

10 to 100 millions

10 tO IOO

The distribution of streptococci and B. enteritidis sporogenes
in animal excreta has not been so fully worked out, but as
regards B. coli this group has been found to be abundant in the
excreta of many birds and fishes. Their occurrence in birds
and fishes is of importance in connection with the bacterial
content of upland waters which may be contaminated with these
organisms from this source.

Houston has examined the excrement of gulls and found
B. coli in enormous numbers. Streptococci and spores of B.
enteritidis sporogenes were present in the excreta of two only
(out of eight) of the gulls and in only small numbers. Eyre has
also found B. coli abundant in a gull and a number of other birds.

Eyre and Johnson have independently examined a number
of fishes, the former from the sea, the latter from river waters
and both found B. coli to be present in the intestinal contents.

All three groups of organisms are abundant in crude sewage.
The actual numbers found will of course vary greatly with the
strength of the sewage, but the data given by Houston give
average figures. These are B. coli and allied forms about

GENERAL CONSIDERATIONS 15

100,000 per c.c., streptococci 1000 — 10,000 or more per c.c. and
spores of B. enteritidis sporogenes 100 to IOOO per c.c.

The second condition is that these indicator organisms should
be absent or relatively absent from sources other than those for
which they are to serve as an index. A great deal of work has
been carried out upon the distribution of these organisms in
saprophytic surroundings and while the results are not in com-
plete accord the work done has established certain general facts
in favour of the use of these organisms (and more particularly
the B. coli group) as a satisfactory measure of excretal or sewage
contamination.

Dealing first with B. coli in soil the general facts set out
in Chapter IV show that members of this group are only
present when the soil has been contaminated with excremen-
titious matters and that virgin soil and soil not manured are
free from these organisms. Further it has been shown that
these organisms gradually die out in soil.

Members of the B. coli group have also been found to be present
on wheat, rye and other grain and a number of investigations in
this direction have been made. The results, considered broadly,
show that when the grain is carefully collected directly from the
fields no B. coli are present, but that if the examinations are
made after ordinary storage members of this group are frequently
found in small numbers. It is to be anticipated that stored
grain will often show B. coli on examination in view of the fact
that these organisms are abundant in the excreta of rats and
mice, animals generally numerous in grain ships and storage
places.

In quite pure water, pure air, etc. B. coli group organisms do
not occur.

Briefly stated, we have, in the B. coli group, organisms which
are extremely abundant in excreta and sewage, but which do
not occur in air, water, soil or other substances unless they have
been in contact with excrementitious matter, while they do not
multiply to any extent under ordinary natural conditions, outside
the animal body.

As regards the saprophytic distribution of streptococci there
is no evidence that streptococci have any true home, under

14 GENERAL CONSIDERATIONS

natural conditions, apart from the animal body. As Andrewes1
points out, " They are certainly abundant in the air of London,
but these common air-streptococci are, as I have elsewhere
shown, identical with the forms most abundant in the horse-
dung which is amongst the chief constituents of London street
dust. They occur, too, in water, but in proportion to its con-
tamination with sewage, or at least, with animal excreta. Much
the same is true of earth and soil. In all these situations strep-
tococci may, indeed, be found, but always, it would seem, as the
result of pollution with organic matter."

Their presence in contaminated and absence from pure soil
is shown in Chapter IV.

Outside the animal body they may survive for considerable
periods but do not thrive. The evidence as to the duration
of viability and vitality outside the animal body is somewhat
conflicting, but in general it would appear that the majority are
delicate organisms and rapidly die out, but that a small number
of hardy strains may persist for very long periods.

In relation to their viability in water one of the most recent
investigations is that of Houston2, who studied the viability of
100 streptococcus strains, isolated from lumps of excreta in
sewage, in Thames river water sterilized by passing through
a Pasteur filter. About 100,000 of each of the 100 varieties
were added separately to 100 tubes, each containing 10 c.c. of
the filtered water. Cultures (i c.c. of water) were made from
them each week. The tubes were kept in a dark cupboard at
a temperature varying from 8° to i6°C. The following is a
summary of the results:

3 per cent, died in I week.
II „ ,,2 weeks.
32 „ „ 3 „
38 „ „ 4 „

8 ,,5 „

4 » » 6
2 9

I „ still alive in 9 weeks,

(one tube became contaminated.)

1 Lancet, 1906, n, Nov. 24 (Horace Dobell Lecture).

2 Metropolitan Water Board, Fifth Research Report.

GENERAL CONSIDERATIONS 15

Our knowledge as to the distribution and viability of the
different varieties of streptococci is very incomplete.

As regards the distribution of B. enteritidis sporogenes under
saprophytic surroundings the investigations made have usually
not accurately differentiated this organism from very closely
allied forms, and the results which have been obtained for
the most part refer to the distribution of the spores of the
organisms which cause the so-called "enteritidis change" in
milk. This change is, however, usually due to B. enteritidis
sporogenes, and for practical purposes the results obtained may
be taken as showing the distribution of this bacillus.

The spores of this bacillus are present in soil and, while
they are more abundant in polluted than in virgin soils, they
are present to some extent in soils not recently contaminated
(see Chapter IV).

Klein and Houston found virulent B. enteritidis sporogenes
spores present in most of the samples of grain (wheat, oats,
rice, oatmeal and wheat-flour) examined. One gramme had
usually to be examined before positive results were obtained.
Balfour Stewart also found the spor.es prevalent in grain.

In considering the value of this organism as an excretal
indicator, it must be remembered that it is a spore-bearing
bacillus and that its spores are very resistant. Animal ex-
cretal pollution is so widespread that it is not a matter of
surprise that such a highly resistant organism should be widely
distributed in nature. The available evidence shows that it is
absent, or relatively absent, from sources which have never been
•contaminated, but that it is fairly prevalent in sources the pollu-
tion of which had taken place possibly at a long antecedent
period. These considerations obviously place a considerable
limit to its usefulness.

The other conditions of suitable bacterial indicators refer to
the readiness with which they can be isolated, estimated and
accurately defined. All three groups, although they possess
drawbacks, fulfil these conditions. From this point of view
the streptococci are the least satisfactory, and really satisfac-
tory methods for their isolation and differentiation are greatly
needed. Thanks to the numerous investigations which have

l6 GENERAL CONSIDERATIONS

been made and particularly to the introduction by MacConkey
of the use of bile salt, the isolation and identification of B. coli
has now been placed on a very satisfactory footing.

In general it may be said that the B. coli group is by far
the most reliable indicator of excretal pollution which we
possess and its use is equally applicable for water, shellfish,
soil and other substances. The other two indicators are chiefly
of value for confirmatory purposes.

These bacteria not only indicate excretal contamination, but
they can be used to measure its extent by means of careful
numerical estimation. This is a matter of extreme importance,
since excretal pollution is so widespread that to some degree
evidence of it must also be widespread, and measurement, not
mere detection, is required.

CHAPTER II

GENERAL METHODS FOR THE ISOLATION AND
IDENTIFICATION OF INDICATOR ORGANISMS

While for the isolation and identification of indicator organ-
isms considerable variations are required in detail, according to
the material submitted for examination, the general procedures
are identical whatever the source. It is therefore convenient and
saves repetition to discuss them together in a single chapter.

The methods used must be capable not only of detecting
the indicator organism but must be able to estimate the number
present in a given sample. In this they differ from procedures
used for the isolation of pathogenic bacilli such as B. typhosus
or Sp. cholerae, since for these detection is all that really
matters.

In the previous chapter it has been explained that the three
most useful groups of organisms for the estimation of excretal
contamination are B. coli and allied organisms, streptococci and
B. enter itidis sporogenes..

ISOLATION OF INDICATOR ORGANISMS I?

The examination and isolation of these organisms involve
three considerations :

(a) A preliminary determination of their probable presence.

(b} Their isolation in pure culture.

(c) The application of determining tests.

Sometimes a and b are combined.

Bacillus colt Group.

To prevent the separate examination and plating of all the
different quantities of water, milk, or other substance selected
for examination it is necessary to have some means by which
those amounts which contain B. coli organisms can, with
some measure of probability, be distinguished from those which
are free from this organism. Authorities are not in complete
agreement as to the respective value of glucose and lactose
fermentation, but the great majority of workers select the fer-
mentation of lactose as the essential primary means of differentia-
tion since this character is a convenient and satisfactory one to
employ.

By the employment of some agent such as litmus or neutral
red in conjunction with lactose to indicate when the acid fer-
mentation of this sugar has taken place, or by the use of double
tubes so that the formation of gas can be seen, it is possible to
differentiate between liquid media tubes (or individual colonies
when solid media are employed) which respectively do not or
which do contain lactose-fermenting bacilli. Such means of
differentiation are of immense importance in routine work as
they enable tubes or plates not containing lactose-fermenting
bacilli to be at once discarded. All modern methods make use
of some such means of differentiation.

For liquid media the fermentation of lactose with production
of gas as well as acid is by far the best means of differentiation,
with solid media the production of acid alone has to be relied upon.

In regard to the comparative value of liquid and solid media,
for this primary differentiation there are considerable differences
of opinion.

S. W. 2

1 8 ISOLATION OF INDICATOR ORGANISMS

The use of tubes of liquid media involves two possibilities of
considerable error. One is that the estimations obtained only
give very widely spaced results. For example, with water samples,
O'l, I, 10 and 100 c.c. are the amounts usually selected for exami-
nation, and according to the positive findings the number of B. coll
organisms will be 10,000, 1000, 100, 10, etc. to the litre. There is
obviously a considerable spacing between these results. This
wide spacing can be largely obviated by the use of more
tubes but this greatly increases the labour and therefore in
practice this method of diminishing the possible error is limited.
The other objection is that bacteria are by no means uniformly
distributed and if the bacilli under examination happen to be
present in the smaller amounts used for examination, misleading
results may be obtained and a much greater prevalence recorded
than is warranted by the facts.

This second objection also largely applies to solid media
enumeration methods, while these methods have several other
decided drawbacks of their own. One of the chief of them
is that they enable small quantities (i.e. O'l to i c.c.) only of
the material under examination to be dealt with. It is not
very convenient to concentrate the bacteria into a smaller bulk,
by filtration or other method, so that this method is not suitable
for many substances.

A further decided objection to their use is that they give
unreliable colony differentiation. Theoretically the medium
used should clearly differentiate the B. coli group colonies from
the rest of the bacteria by their being coloured red or otherwise
distinguished ; in practice there are many intermediate colonies
which it is not possible to say, without further cultural differ-
entiation, whether they belong to the B. coli group or not. In
particular if the plates are crowded true B. coli group bacilli will
frequently not develope into characteristic colonies until after
several days so that there is great danger of their being un-
enumerated as B. coli colonies. The presence of these indeter-
minate colonies seriously reduces the accuracy and value of the
direct plating method and makes the results obtained dependent
to some extent upon the recorder. From extended comparative
work with the examination of milk and other substances the

ISOLATION OF INDICATOR ORGANISMS Ip

writer decidedly prefers liquid media for primary enumeration.
It should, however, be said that a number of authorities prefer
the solid plate method1.

A second object to be aimed at in the detection of B. coli and
other indicator organisms is to use means whereby the growth
of these organisms is favoured while that of other bacteria is
retarded. A simple way of doing this is to employ a temperature
which selectively favours the growth of the bacilli required.
Vincent suggested a temperature of 42° C, MacConkey used
at first a temperature of 43° C., while in Eijkmann's method a
temperature of 46° C. was employed. B. coli organisms grow well
at 37° C. and the advantage of these higher temperatures is not
commensurate with the trouble of using special incubators.

Growth under anaerobic conditions was for similar reasons
advocated by Fakes, but the advantages are not great enough
to balance the extra trouble and other methods are better.

Selective growth may also be favoured by the addition of
certain chemicals. Studies along this line have been very valu-
able and fruitful. Phenol was at one time extensively employed
but is now largely given up as it does not satisfactorily differ-
entiate.

We owe to MacConkey the most valuable substance for
selective differentiation in the bile salts, usually employed as
sodium taurocholate. This substance is a satisfactory inhibitory
substance, while in the amounts used it does not interfere with
the growth of B. coli.

The inhibitory action of certain colouring media, such as
crystal violet, have also been successfully employed.

Of all these agents it may be said that the use of lactose
bile salt broth in double tubes is the most valuable when the
liquid enumeration method is employed, while for solid media
any of those mentioned in the next stage may be employed.
The composition of these media is given in the Appendix.

The second stage of the examination consists in the isolation
of the bacillus in pure culture. If solid media were employed

1 For a good discussion of this question and in favour of solid media see Gartner,
" Das Bacterium coli als Indikator fiir fakale Verunreinigung eines Wassers," Zeit.f.
Hyg. 1910, vol. 67, p. 55.

2—2

20 ISOLATION OF INDICATOR ORGANISMS

for the first stage this is already carried out, but if liquid differ-
entiating tubes are employed the bacillus has to be isolated from
the tubes selected.

The principles enunciated above apply equally here and it is
a great advantage to employ media containing both a differen-
tiating agent and a substance which retards the growth of
bacteria other than those to be isolated.

On these grounds the use of ordinary gelatine and agar have
been given up and coloured media employed.

Of such media lactose bile salt neutral red agar (L.B.A.)}
nutrose agar (Drigalski-Conradi agar), and fuchsin agar may be
specially mentioned.

The ingenious aesculin agar may also be mentioned here.
This contains the glucoside aesculin and ferric citrate. The action
of B. coli causes the aesculin to combine with the iron citrate
and form a dark brown salt, the B. coli colonies being black.

Fig. 2. Bent glass distributing rod.

'In brushing these plates of coloured media it is important
that the media should be thoroughly dried. A convenient
general procedure is to add one loopful of the tube selected for
examination to a tube of sterile water. After mixing well place
two loopfuls of the latter upon the surface of the L.B. A. or other
material selected and distribute uniformly over the plate with a
sterile bent glass rod. The medium in the plate will be wet from
condensed water on the surface. To dry incubate it for \\ to
2 hours uncovered in the 37° C. incubator. Then cover and

ISOLATION OF INDICATOR ORGANISMS 21

invert. Incubate for 20 — 24 hours at 37° C. before the plates
are examined.

At least three characteristic colonies should be subcultivated
even if they all appear alike. Only one if typical need be
worked out, but the others should be subcultivated, so that if the
one selected is atypical in any of its characters, these can in
turn be fully examined.

The third stage is the application of determining tests. The
value of the different tests has already been considered and
those advocated for employment are mentioned under the
respective substances to be examined.

Streptococci.

In the determination and enumeration of streptococci many
of the principles involved are identical with those concerned in
the B. coli enumeration and explained above so that a shorter
consideration is sufficient. As with other indicator organisms a
numerical estimation is required, not merely their detection.

There are no known chemical agencies which will enable,
by macroscopic appearances alone, tubes of liquid media contain-
ing streptococci and other organisms to be differentiated from
those which do not contain streptococci. For their preliminary
determination in liquid media reliance has therefore to be placed
upon other procedures.

When suitable liquid media are employed the presence or
absence of streptococci can be judged with considerable accuracy
by a careful microscopic examination of the liquid, particularly
of any sediment.

Using this basis of enumeration various quantities- (O'l, ro,
10 c.c., etc.) of the water, milk, or other substance under exami-
nation are added to liquid media (such as glucose neutral red
broth) which allow the streptococci to grow well and the tubes
are carefully microscopically examined after incubation at 37° C.
for the presence of streptococci.

It is obvious that such a method is only capable of enabling
streptococci as a class, and not the prevalence of any particular
kinds, to be enumerated. This method has both the drawbacks
of widely spaced results and the errors from unequal distribution

22 ISOLATION OF INDICATOR ORGANISMS

discussed under B. coli, while in addition there is possibility of
error in that the streptococci may fail to grow or be overgrown
or that they may fail to be detected when present.

The last is of some importance and can only be diminished
by careful examination in hanging drop preparation, followed in
doubtful cases by examination of the stained centrifugalised
deposit.

An alternative method, as for B. coli enumerations, is to use
solid media over which definite fractions of the substance under
examination are brushed and then incubated. All colonies
likely to be those of streptococci are subcultivated and carefully
investigated. Several media have been used as suitable for this
purpose, the most useful being ordinary agar and the medium
of Drigalski and Conrad i.

The enumeration of streptococci by this means is not very
satisfactory. Only small quantities (o-i to ro c.c.) can be used for
plating since concentration methods are time consuming and
not satisfactory. Also none of the media employed are capable
of differentiating in any satisfactory way the probable strepto-
coccus colonies so that in practice a very large number of
colonies have to be subcultivated and investigated, very few of
which are streptococci. This makes the procedure very tedious.
It is also probable that not all streptococci will grow on such
plates, particularly when the other colonies are numerous. When
it is a question of studying the biological properties of the
isolated streptococci this method must be adopted.

The determining tests to apply to isolated streptococci have
already been discussed in Chapter I and are further referred to
in the subsequent chapters.

Bacillus enteritidis sporogenes.

The estimation is based, not upon the number of these
bacilli present, but upon the number of their spores. The
cultural feature which is used as pointing to the presumptive
presence of this organism is the characteristic change in milk
described in Chapter I.

Various amounts of the substances under examination or,
if concentration of the bacteria has been effected, definite

ISOLATION OF INDICATOR ORGANISMS 23

equivalent amounts, are added to tubes of sterile whole milk,
recently boiled. The destruction of all but spores is ensured
by heating to 80° C. and maintaining at that temperature for
ten minutes. The milk tubes are then cooled and incubated under
anaerobic conditions. The most convenient method of doing
this is to employ the pyrogallic acid and caustic potash method.
The milk tubes are placed in jars or larger tubes containing
pyrogallic acid in powder to which is added, when all is in place,
a little 10 per cent, caustic potash solution. The tubes or jars
must then be immediately closed by tightly fitting stoppers. The
alkaline mixture absorbs the oxygen. The numerical estimation
is arrived at by recording the tubes, which show after incubation
the characteristic " enteritidis change," as containing the spores
of this bacillus while those which do not show it are recorded as
negative.

As a rule no attempt is made by workers to isolate the
bacillus in pure culture or to definitely prove its presence by
animal inoculation. The estimation is therefore not really one
of a definite bacillus but that of a group of closely allied bacilli
which, as far as is known, have a very similar distribution in
nature and a very similar significance.

The estimation as carried out is a fairly simple one and
involves no great variations in the hands of individual workers.
There are two points in connection with the test which require
further consideration. Both deal with the question of the
amounts to examine.

It is easy to handle O'l, I and 10 c.c. of the liquid under
examination (water, milk, etc.), but when larger amounts have
to be examined, as is necessary with water, some method of
concentration has to be practised. These concentration of
bacteria methods have been mentioned under Streptococci
and B, coli, but are more conveniently considered here.

To obtain the bacteria in a convenient bulk of fluid it is
possible to evaporate off the water at a low temperature under
reduced pressure, as in the method for B. typJiosus in water
described in Chapter III, but this is a procedure very rarely
employed. The usual plan is to filter the water through a
sterile porcelain filter, the water passing through and leaving

ISOLATION OF INDICATOR ORGANISMS

the contained bacteria deposited upon the filter. In most cases
it is more convenient to filter from within out, so that all the
bacteria are deposited upon the inside of the candle. After
filtration a definite quantity of sterile water is added and an
emulsion made. Usually 10 c.c. represents the final bulk. If,
for example, 1000 c.c. of water is concentrated in this way into
10 c.c., it is assumed that each c.c. of the emulsion will contain
the bacteria in looc.c. of the original fluid and other fractions
in proportion. Of course this assumption is not altogether

' TrT ^ To exhaust
LJ**1 v pump

Fig. 3. Apparatus arranged for filtering — aspiration.
From Eyre's Bacteriological Technique.

correct, but it is sufficiently accurate for practical purposes.
Definite fractions of the emulsion are added to milk tubes
and treated as above described.

Filtration is extremely slow through such fine pore filters,
so that to accelerate it aspiration or pressure or both may be
employed. As a rule aspiration through an exhaust pump is
sufficient.

The apparatus figured (Fig. 3) is a convenient form to
use. In this apparatus the water is placed in a separating

ISOLATION OF INDICATOR ORGANISMS 2$

funnel, which is connected with the filter candle by means of
a perforated rubber stopper. The filter flask is connected to
an exhaust-pump. In this arrangement the filtration is from
within out, so that all the bacteria are deposited upon the
inside of the candle. After filtration sterile water is added,
and an emulsion made. The porcelain filter tube should hold
rather over 10 c.c. ot water.

The apparatus must be sterilized in steam for several hours
before use.

The actual amounts to select for examination will vary
somewhat with the nature of the material being examined.
For sewage or milk, for example, only small amounts are
required, as these spores are likely to be abundant in such
material, while for water 100 or 1000 c.c. are often suitable
amounts to examine. As a rule the amounts taken are not
nearer spaced than -^, i, 10, ico, 1000 c.c. etc. This wide
spacing, while sufficient for some substances, such as water, is
not satisfactory when the spores are likely to be fairly numerous.
An alternative and more satisfactory procedure is given under
milk (Chapter VI).

Methods of dilution and of recording results.

As has already been pointed out, numerical estimations,
not mere detection, are essential for the indicator organisms
discussed in this chapter. Using the methods recommended,
this involves the examination of a series of carefully measured
quantities of the substance. Considerable attention to detail is
required if reliable and comparable results are to be obtained.
When solid substances, for example soil or the meat foods,
have to be examined, accurate enumerations of the number of
bacteria present or of the number of any special organism are
very difficult. The most satisfactory procedure devised up to
the present is to add definite weights of the substance to bottles
or flasks containing definite amounts of sterile water. As a
rule this can most conveniently be done by cautious addition
of the soil or other substance to the diluting flask, in equi-
librium upon a balance, until the addition of the quantity
required is registered. The diluting vessels should have glass

26 ISOLATION OF INDICATOR ORGANISMS

or india-rubber stoppers to enable the solid and the water to-
be very thoroughly shaken and mixed. Definite fractions 01
the water are then examined in the same way as for an
ordinary fluid and as described below.

The assumption has to be made that the diluting water,
after thorough mixture, contains all the bacteria in the solid
substance added. This is an inaccurate assumption, since it
is not to be anticipated that all the bacteria can be dislodged,
the extent of the error varying with the nature of the sub-
stance examined. If, however, it is realized that only rough
comparable results are in this way obtainable, this method is
of great value. The results are most conveniently returned
as per gramme of substance.

The procedure is much simpler when liquids have to be
directly dealt with, but certain precautions are necessary.
The amounts to be examined are removed by sterile pipette.
For fractions of a cubic centimetre it is obvious not only
that the gradations must be wide apart (i.e. the bore must
be narrow) but also that too small fractions must not be
sampled. When no very special accuracy is required, it is
possible to measure sufficiently down to -^ c.c. for water and
some other liquids, if they are not heavily contaminated, but
with fluids such as milk it is not reliable to take so small a
quantity as O'l c.c. and errors as great as thirty per cent, may be
made in this way. In all cases where possible it is best to obtain
these small fractions by making up preliminary dilutions, and this
course is absolutely necessary for quantities less than O'l c.c.

A whole series of such dilutions may have to be prepared.
They are most conveniently made by adding 10 c.c. of the
original fluid to 90 c.c. of sterile water in a stoppered bottle,
mixing very thoroughly. One c.c. of such a first dilution will
represent O'l c.c. of the original fluid. In the same way second
dilutions can be made from the first, and so on as many as
are required. It is preferable to dilute as above rather than
to add I c.c. to 99 c.c., as obviously there is far less likelihood
of errors from slight inaccuracies in measurement.

One mark 10 c.c. pipettes are convenient for making the
dilutions in bulk. Such pipettes should be made short (9 inches

ISOLATION OF INDICATOR ORGANISMS 2/

or so) for convenience of sterilization, but care should be taken
to obtain those in which the lower part is of sufficient length to
reach well into the diluting bottles.

While solids are most conveniently recorded as per gramme
of material, data as to bacterial content of liquids are returned
as per c.c. or litre.

For B. coli, streptococci, etc. when a series of dilution tubes
are used for enumeration purposes the results are not definite
amounts, but can only be returned between the limits of the
dilutions employed. For example, if a given water sample
showed B. coli present in 10 and in 50 c.c. but absent in i
and O'l c.c., the deduction would obviously be that there were
more than 100 but less than 1000 B. coli per litre of the water
and the result would be recorded as B. coli 100 — 1000 per litre.
For routine work and reporting this is a preferable method of
recording to stating the amounts examined which were positive
and those which were negative.

CHAPTER III

WATER

Water under natural conditions invariably contains bacteria,
the actual numbers present varying enormously with the class
of water supply, the degree of soil filtration, opportunities of
bacterial contamination, etc.

Natural water supplies as regards their bacterial content are
never in a state of equilibrium. On the one hand bacteria-
holding substances are constantly being added, thereby raising
their bacterial content, while on the other hand certain natural
purification agencies are constantly at work tending to reduce
the bacteria present. Of these agencies the forces conveniently
summed up as sedimentation are the most important, while light,
sunlight and deficiency of food material are all factors playing
a part. The effects of these different agencies are most readily
studied in regard to rivers and streams. An ordinary stream

28 WATER

is constantly fed by small streams containing soil washings,
drainage of cultivated and uncultivated land and often drainage
from human habitations. In this way vast numbers of bacteria
are being added to the water and if samples are collected at
favourable situations they show a great increase in the bacterial
content. Acting all the while, however, are the agencies men-
tioned above and making for the diminution of the bacteria
present, so that samples taken lower down the same stream
may show a much smaller number of bacteria with an absolute,
or at least a relative, diminution of the unsatisfactory organisms
possibly very prevalent higher up.

Water, unless highly charged with organic matter or unless
other conditions are favourable, is not a medium which readily
allows bacterial multiplication.

It will be shown later on that the numerical estimation of
certain special bacteria is of far greater importance than general
bacterial enumerations. The presence of these special bacteria
in water supplies is subject to the same influences as affect the
total number of bacteria.

In view of the varying influences which affect the bacterial
content of water and which are at work unequally for different
classes of water it is essential in all cases to carefully consider
the nature of the water supply under investigation.

Analysis of water supplies is undertaken for two purposes:

(a) To obtain reliable information in regard to the fitness
of a given water supply for drinking purposes.

(b) To judge the suitability of a supply for domestic or
medicinal purposes.

Bacteriology plays no part in the second object, which has
to be judged on chemical grounds, but is of very great value
in ascertaining suitability for drinking purposes.

From the bacteriological examination of water data can be
obtained which not only give information as to existing con-
ditions, but from which deductions can be drawn as to recent
harmful pollution.

Bacteriological examinations are also very useful in connec-
tion with the testing . of the efficient working of sand and
other filters.

WATER

Collection and transmission of samples.

One of the drawbacks of the bacteriological examination
of water is the great care which is required in the collection
of the samples ; even apparently trivial errors or omissions
may entirely vitiate the result. Very precise and seemingly
trifling directions must be given, unless the sample is collected
by an expert.

For the ordinary examination 2-ounce (57 c.c.) glass-stoppered
bottles are sufficient.

Fig. 4. Collecting Bottle and Tin,

Sometimes it is advantageous to examine greater quantities
of the water and collecting bottles of larger size must be used.
It is rarely necessary to examine more than rooc.c., and this
can be most conveniently done by using two of the small
bottles for collecting the sample.

If the specimen cannot be examined at once, and delay is
unavoidable, the sample should be packed in ice, and then

30 WATER

transmitted to the laboratory. Special forms of apparatus have
been designed for this purpose. That figured in Chapter VI
under Milk Examination is a very convenient form. As used
by the writer, the inner chamber is made to take exactly four
bottles and tins as described below.

The bottles fit tins into which they can just slip (Fig. 4).
The bottom of the tin is provided with a layer of cotton wool
and then a piece of asbestos cardboard. Several thicknesses of
asbestos cardboard are also fitted into the cover of the tin so
that, when in place, the bottle is firmly in contact with the
asbestos above and below. The bottles, with their stoppers
rather loosely inserted, should be sterilized in their tins. The
tins, after sterilization, are not opened until immediately before
the sample is collected.

To take samples from various depths, several different forms
of apparatus have been devised. The ordinary collecting bottle
may also be used for this purpose. It is tied into a leaden
cage, and lowered to the required depth by catgut or string
attached to the cage. The loosened stopper is then removed
by a jerk upon a second string, previously tied to the stopper,
and the sample collected.

In collecting samples from a reservoir, lake or river, plunge
below the surface before removing the stopper, thus avoiding
scum and surface contaminations. If from wells with pumps,
pump away a considerable quantity of water before collecting
the sample ; while, if a complete investigation is required, a
second sample should be obtained after several hours' pumping.
If from a tap, allow the water to first run to waste for five to ten
minutes. In collecting samples from a tap it is necessary to
remember that the interior of the tap may not be clean and care
must be taken to obviate this source of error. As far as possible
samples should be taken from rising mains and always with
as little intervention of pipes as is practicable.

The stopper should only be removed at the actual time of
collection, great care being taken that the part of the stopper
which goes into the bottle is neither touched nor brought into
contact with anything, apart from the bottle or water. It must
be held by its free end, and at once replaced and screwed in

WATER 31

firmly when the sample has been collected. The collected
water should not quite fill the bottle.

It is essential that full particulars as to the source of the
water be supplied with the sample. It is impossible to give
a satisfactory opinion without such information.

The following should be recorded :

1. Date of sampling.

2. Nature of the water — spring, upland surface, etc.

3. Whence obtained — pump, draw-well, river, tap, etc.

4. Precise particulars of sampling — e.g. depth below sur-
face, from middle or sides. If from tap or pump, time during
which the water was allowed to run to waste. Filtered or
unfiltered.

If from a tap, note if it was directly connected to a main,
•or if it was connected with a storage cistern, or other form
of supply.

5. Details as to previous rainfall.

The above particulars should all be supplied to the bac-
teriologist conducting the examination, and in addition to the
careful topographical investigation which should be made and
recorded in the case of all water supplies.

The general nature of the examination.

Bacteriological water examinations, as conducted at the
present day, usually involve three lines of investigation, although
all three are not, of necessity, carried out for every sample.
These are :

(a) The quantitative estimation of the total number of
organisms present capable of developing upon the nutrient
media used.

(b) The isolation and numerical enumeration of organisms
not necessarily harmful, but which from their origin are espe-
cially liable to be associated with harmful contamination.

(c) The isolation and identification of actual disease-
producing organisms such as Bacillus typhosus and Spirillum
cholerae.

32 WATER

It will be noticed that from the first two methods deductions
only can be drawn as to purity, while with the last method the
actual disease-producing organisms are isolated.

Quantitative Exam in at ion.

As almost universally practised at the present day, this
consists in adding varying quantities of the water to tubes of
liquefied gelatine and agar, each of which is then thoroughly
mixed and poured out with all precautions into a Petri dish,
and solidified as rapidly as possible. These plates (Petri
dishes) are then incubated, the agar at 37° C., and the gela-
tine at 20° C. to 22° C. Each organism present, capable of
multiplication under the conditions existing, develops into a
mass of bacteria or colony, visible to the naked eye, and as
such readily counted. The total number of colonies gives
the maximum number of organisms capable of development
in the medium used, within the time given and at the particular
temperature of incubation.

In actual practice many details have to be attended to.
For instance, the reaction of the gelatine and agar used has a
marked influence on the number of organisms which develop.
Both the English and American Committees on Standard
Methods of Water Analysis recommend a + I per cent, reac-
tion, and accurately standardized media of this reaction should
alone be used for routine work (see Appendix).

The composition of the agar and gelatine nutrient media
also influences the number of colonies, and care should be
taken to consistently manufacture in the same way and with
the same quality and amount of ingredients.

The amount of water to add to the media tubes must
obviously vary with the suspected degree of contamination of
the water.

For ordinary waters O'2 and 0*5 c.c. are convenient amounts
to add to the gelatine tubes, and O'2 and roc.c. to the agar.
For contaminated waters a considerably greater dilution may
have to be practised. The water is conveniently measured
by i c.c. pipettes graduated in one-tenths of a c.c., previously

WATER 33

sterilized in the hot-air sterilizer after thorough cleaning and
plugging of the upper ends with cotton wool.

In practice proceed as follows : Melt agar and gelatine
tubes, and cool down to 42° or 43° C. for the agar, and 25°
to 30° C. for the gelatine. Mix the sample thoroughly. Add
O'2 and 0*5 c.c. of the sample respectively to two gelatine
tubes, and O'2 and ro c.c. to agar media tubes. Distribute the
water uniformly through the medium by rotation between the
fingers. Pour out, after flaming the cotton-wool plugs, into
sterilized Petri dishes, only raising the upper dish just suffi-
ciently to admit the top of the test-tube. Solidify as soon as
possible either over ice or by using a plate-cooling apparatus

(Fig- 5).

Fig. 5. Savage's Plate-cooling Apparatus.

The inlet pipe is connected by indiarubber tubing to a water tap and the outlet
pipe to a sink. Plates of media placed on the thick plate-glass top are rapidly
solidified.

Incubate the gelatine plates at 20° to 22° C. and the agar
at 37° C. Count the agar plates the next day, and after forty
to forty-eight hours. Count the gelatine plates daily, the final
count being at the end of seventy-two hours.

To count the colonies it is best to count against a dark back-
ground, dividing up the area of the plate, to facilitate counting,
by lines on the back made with a paraffin pencil.

All the colonies on the plate should be counted, but if they
are very numerous, and an approximate estimation only is
possible, then, but only then, some mechanical aid such as
Pakes' disc may be used, a few segments counted, and the
total number deduced.

In stating results, the details of temperature, time of incu-
bation, and reaction of the medium should always be given
and recorded with the analysis.

s. w. 3

34 WATER

Qualitative Exam in at ion*
This may be divided into two parts.

A, The detection and estimation of indicator organisms.

B. The examination for specific disease-producing bacteria.

A. The detection and estimation of indicator organisms.

The most important part of the bacteriological examination
of water is the examination for and estimation of organisms
whose presence serves as an indication of undesirable con-
tamination, such as from excreta or sewage. The value and
significance of such indicators has been discussed in Chapter I.

Examination for B. coli and allied organisms.

The general principles relating to the isolation of these bacilli
together with a consideration of the most suitable methods have
been dealt with in Chapter II, so that only the actual procedures
recommended for water samples need to be described. The
following is recommended as convenient and satisfactory :

Add 0*1 and roc.c. respectively to tubes of lactose bile salt
broth in double tubes. Add 10 c.c. to a similar tube, but con-
taining lactose bile salt broth of double strength. To the
remainder in the bottle, after all the different amounts of water
have been withdrawn for the different parts of the examination,
add the contents (about 10 c.c.) of a tube of four times strength
neutral red broth. Replace the glass stopper. Four times
strength bile salt broth may be used, and, if the examination
is for B. coli alone, is preferable, but by using neutral red broth
the mixture is also available for the examination for strepto-
cocci.

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

The tubes are labelled, incubated at 37° C, and examined
after twenty-four and after forty-eight hours.

WATER 35

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

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

In certain cases it is of value to ascertain the presence or
absence of B. coli in 500 or 1000 c.c. of the water.

This may be done by filtering the water through a porcelain
filter, as described for the examination of B. enteritidis sporogenes
(Chapter II), and using the filter brushings, emulsified in a
little water, to add to tubes of bile salt broth. This method is,
however, cumbersome and unsatisfactory.

A far better procedure is to convert the water sample itself
into a nutrient medium by the addition of four times strength
broth, incubate at 37° C. for twenty-four hours, and then either
inoculate suitable plates direct, or, what is preferable, add I c.c.
by sterile pipette to a tube of bile salt broth, and incubate this
for one or two days. If any B. coli were present in the original
bulk of water, the preliminary incubation in the water broth
would have allowed them to multiply sufficiently to be at least
present in I c.c. of the sample, an amount of fluid readily
examined.

The composition of the media used is given in the Appendix.

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

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

(a) Gelatine slope (for morphology, motility, cultural ap-
pearance, and liquefaction).

3—2

36 WATER

(b) Litmus-milk at 37° C.

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

(d) Peptone water (for indol production).

(e) Saccharose peptone litmus solution (in a double tube).
Houston1 has suggested and used very extensively a number

of modifications which are valuable when a large series of samples
have to be examined. The B. colt organisms are isolated from
the positive gas tubes by inoculating sloped lactose, bile salt,,
neutral red peptone agar (called by Houston rebipelagar) in test
tubes instead of plates. As far as possible five red colonies are
in each case subcultivated and investigated. The confirmatory
tests selected are carried out with the following media : — glucose
litmus gelatine, lactose litmus gelatine, saccharose litmus gelatine
and peptone water. The results are recorded after 24 hours'
incubation at 20 — 22° C., the data recorded being the production
of gas from the three sugars and the presence of indol.

Houston recommends gelatine sugar media (i p.c. sugar,
2 p.c. peptone, I p.c. lemco, 7-5 p.c. gelatine and I c.c. of a 5 p.c.
potassium hydrate solution) for gas production testing, finding
such media more sensitive than the usual liquid media.

Houston has introduced a number of modifications which
save time when dealing with large batches of routine samples
such as are examined daily in the Metropolitan Water Board
Laboratories.

Labelling is reduced to a minimum by the use of coloured
cotton wool plugs. Five colonies from each rebipelagar tube
being subcultivated the media are tested in sets of five. Five
quite small test tubes (2" x J") filled with the same sugar gelatine
or other medium used are placed, in one large test tube, only the
latter having a cotton-wool plug,

To expedite the inoculations the colony selected is picked
off by a sterile straight iron wire (sterilized in batches over
naked bunsen flame). The wire is then placed in a tiny test
tube containing a few drops of sterile salt solution. Into this
little tube are then introduced as many more sterile wires as

1 For a more detailed account see Metropolitan Water Supply Itcport for January,
1907, pp. 46-52.

WATER 37

correspond to the number of tests proposed to be made and these
are used to inoculate the confirmatory tests. This saves time
and trouble, besides avoiding any uncertainty in repeatedly
going back to a colony in order to inoculate a series of cultures.

Examination for streptococci.

The detection of the presence of streptococci and their
numerical estimation in water is a procedure which is only
carried out by some bacteriologists but is one which is decidedly
useful.

The method originally employed by Houston, to whom is
due the credit of first pointing out their importance in water
examinations, was to concentrate the water by passing a litre or
other definite quantity through a sterile porcelain filter and
brush definite fractions over agar plates. The agar plates were
examined after 24 hours' incubation at 37° C. and the minute
colonies subcultivated into broth tubes and further investigated,
if streptococci were found to have grown. In more recent work,
examining London water, Houston brushed I c.c. of the water
•directly over plates containing Drigalski-Conradi medium and
subcultivated all the minute colonies which developed after
incubation at 37° C.

The differentiating tests to apply are discussed in Chapter I.

The method employed by the writer is based upon the
•detection of streptococcus chains in different fractions of the
water sample.

This may conveniently be done by adding 0*1 and roc.c.
respectively to tubes of glucose neutral red broth, and 10 c.c.
to a tube of the same broth, but of double strength. These,
together with the 30 c.c. preparation used in the B. coli examina-
tion, are examined after incubation at 37° C. for 40 to 48 hours
in hanging-drop preparation. Only cocci in quite definite chains
can be taken as evidence of the presence of streptococci, and
several preparations should be examined from each tube.

It is true that in this way the presence of streptococci as a
class only will be ascertained, but with our present knowledge the
streptococcus test for excretal contamination does not rest upon

38 WATER

the presence of any one variety of streptococcus, but upon the
group as a whole.

Neutral red broth is preferred to plain nutrient broth because
the streptococcus chains seem to be more readily detected micro-
scopically in it.

In cases in which it is doubtful whether streptococci in chains
are actually present a definite opinion may often be arrived at
by centrifugalizing the fluid, and microscopically examining a
little of the deposit stained by methylene blue.

Examination for B. enteritidis sporogenes.

This examination is carried out by some bacteriologists and
is included here although not recommended.

Even in waters from contaminated sources this organism
may be present in only small numbers, so that it is usually
necessary to examine a considerable volume of the water. Con-
venient amounts to examine are 10, 100, 500 and 1000 c.c.

To deal with these large volumes of water the sample is
concentrated by filtration through porcelain as described in
Chapter II.

If one litre is concentrated into 10 c.c., then 10, 100 and
500 c.c. of the original water will be represented by o'l, I and
5 c.c. respectively of the concentrated filtrate.

The presence of the spores of this bacillus are determined by
inoculating milk tubes as described in Chapter II.

B. The examination for specific disease-producing bacilli.

Some diseases, such as diphtheria or scarlet fever, appear to-
be never spread through water. The only bacilli associated with
definite diseases for which it is ever necessary to examine water
are the typhoid bacillus, the cholera vibrio, the organisms of
dysentery in the tropics, and, under very exceptional circum-
stances, members of the Gaertner group in suspected outbreaks
of food poisoning or paratyphoid fever. The examination
for Gaertner group organisms is described in Chapter VIII
and modifications for the examination of water will readily
suggest themselves. Only the examination for the bacteria of

WATER 39

typhoid fever and cholera will therefore be dealt with in this
chapter.

Examination for B. typhosus.

The methods available for the examination of water for the
typhoid bacillus are still not very satisfactory, and this is
evidenced by the very large number which have been recom-
mended. Great improvements in procedure have taken place
in recent years resulting from the introduction of selective differ-
entiating media and of more satisfactory methods of concentrating
the bacilli.

The problem is a difficult one, because a comparatively
delicate organism, with no very definite morphological or cultural
characters, has to be detected among a large number of other
organisms which for the most part thrive much better on all
media and at all temperatures than itself.

Further, the organism may be present in very small numbers,
for it tends to rapidly die out in water, so that it is necessary to
examine a large bulk of water.

Moreover, owing to the long incubation period of the disease,
attention is usually not directed to the water as a possible source
of the infection until several weeks after the specific contamina-
tion, and when all the typhoid bacilli may have died out.

The identification of B. typhosus from water naturally divides
itself into three stages :

1. Preliminary methods, whereby any typhoid bacilli present
are obtained in a quantity of fluid small enough to be directly
plated.

2. The isolation of the organism in pure culture.

3. The tests necessary to establish its identity.

i. Concentration. One or other of the following preliminary
methods may be used :

(a) Enrichment. By the addition of concentrated broth or
other substance to the water, it may be converted into a nutrient
medium in which any contained typhoid bacilli can multiply so
as to be, after incubation, numerous enough to be present in a
very small quantity of the fluid. If only plain nutrient broth is

40 WATER

used, there is considerable danger that some of the other bacilli
present will multiply at a greater rate than any typhoid bacilli
present, and the latter be relatively less numerous after incuba-
tion, if not actually suppressed. Numerous chemical substances
have been tried, or advocated, as enabling the typhoid bacillus to
multiply relatively better than competing bacilli, but most of
them, on further investigation, have not been found to be reliable.
Malachite green has been found to be of some service, but a
series of different dilutions must be employed. The water is
placed in flasks, and sufficient malachite green broth is added to
each to make the strength of malachite green in the mixture I in
2000, i in 5000, I in 10,000 respectively. After incubation at
37° C., or preferably at 40° to 42° C., for 24 hours, the mixtures
are plated upon the solid media described below.

On the whole it may be said that enrichment and selective
enrichment methods are less satisfactory than sedimentation or
direct concentration.

(b) Concentration by mechanical precipitation by chemicals.
Chemicals are added which are harmless to typhoid bacilli but
which form a flocculent precipitate which carries down all or
most of the bacilli in the water and enables them to be obtained
in a small and easily handled bulk of material. If a centrifuge
is available precipitation is facilitated.

A number of chemicals have been recommended for this
purpose, e.g. lead acetate and sodium hyposulphite (Vallet and
Schiider), iron sulphate (Picker), alum (Willson), liquor ferri
oxychlorati (Miiller).

Picker's method is perhaps as good as any. In this method,
two litres of the water are placed in a tall cylinder, and mixed
with 8 c.c. of a 10 per cent, soda solution. Seven c.c. of a 10 per
cent, solution of ferrous sulphate are then added, and stirred in
with a glass rod. The mixture is placed in an ice-chest, and
allowed to stand for a few hours ; if a centrifuge is available the
precipitate can be separated at once. The precipitate is trans-
ferred to a sterile tube, and about half its volume of a 25 per cent,
solution of neutral potassium tartrate is added. The tube is
corked and well shaken, until the precipitate is completely
dissolved, more tartrate being added if necessary. From this

WATER

solution large Petri dishes containing Drigalski-Conradi or other
suitable solid media are inoculated.

Ficker claims that by this method there is very little loss of
typhoid bacilli ; 97 to 98 per cent, of those present being carried
down with the precipitate.

(c) Concentration by sedimentation without chemical precipi-
tation. If a powerful centrifuge is available the water may
be directly centrifugalized and the deposit used for plating.
Houston1 in some investigations upon typhoid bacilli in river
water used this method. In dealing with a raw water containing
72,000 bacteria per c.c. he was able to recover the typhoid
bacillus when 47 had been added to 100 c.c. of the water.

(d) Concentration by evaporation wider reduced pressure.
Wilson2 has described a method which enables water to be

f r f~ r~T TTTt "f- H

T rr

Fig. 6, Wilson's apparatus for concentration under reduced pressure.

1 Fifth Research Report, Metropolitan Water Board> 1910.
8 Brit. Mtd. fount. May 18, 1907.

42 WATER

rapidly concentrated without destruction of bacteria. The glass
reservoir containing the sample of water is connected with an
exhaust pump which rapidly lowers the pressure. The reservoir
is immersed in water which is accurately kept at 37° C. The
water boils at this reduced temperature and as a rule 24 hours
suffice to evaporate four litres of water almost to dryness. The
residue is spread over solid media plates as described below.

This method can also be combined with "enrichment" by
the addition of broth or other nutrient medium to the water two
hours after the commencement of the evaporation, this delay
being necessary to avoid frothing.

In a later paper Wilson and Dickson record that they were
able, using this method, to recover the typhoid bacillus from
water containing only 30 typhoid bacilli in four litres, equivalent
to one bacillus in 133 c.c. of water.

2. Isolation. To isolate the typhoid bacilli from the concen-
trated material a large number of special media are now available.
These media do not so much show which colonies are typhoid
bacillus colonies as indicate the colonies which certainly are not
typhoid bacillus colonies. In this way a vast number of organisms
also present are either suppressed or, if they develop, are readily
distinguished by their colour, appearance, etc., from typhoid
bacillus colonies. All colonies possibly the latter must be
further investigated.

Of these special media, the following may be mentioned :
Drigalski-Conradi agar, lactose bile salt neutral red agar (L.B.A.),
fuchsin agar, malachite green agar and brilliant green picric acid
agar as modified by Fawcus.

The composition of these media is given in the Appendix.

Houston in the investigations mentioned above used L.B.A.
modified by the addition of other sugars, etc. In addition to
lactose he added saccharose, adonite, raffinose and salicin, all
0*2 per cent, of each. By their addition the number of white
colonies was still further reduced and without risk of typhoid
bacilli being overlooked.

With all these media the plates before or after inoculation
should be thoroughly dried uncovered, preferably in the incu-
bator. They are then covered, inverted and incubated. The

WATER 43

colonies are sufficiently differentiated after 16 to 24 hours at
37° C.

With Drigalski-Conradi agar the B. coli colonies are red, not
transparent, and have a diameter of 2 to 6 millimetres, but con-
siderable variation in size and degree of colour are met with.
The B. typhosus colonies are blue, with a violet tinge ; they are
transparent and resemble dewdrops, and have a diameter of I to
3 millimetres, seldom larger.

With fuchsin agar the B. coli colonies are bright red, round
and have prominent margins ; the typhoid colonies are round,
colourless, very transparent, and have thin margins.

With L.B.A. B. coli and other lactose fermenters grow as red
colonies while B. typhostis, Gaertner group bacilli and other non-
lactose fermenters form white colonies.

Using Fawcus's brilliant green picric acid agar lactose fer-
menters form opaque colonies, while B. typhosus and other non-
lactose fermenters grow as clear colonies.

Malachite green agar is more suitable for Gaertner group
work. Fuchsin agar does not keep well and Drigalski-Conradi
agar is troublesome to prepare and is not always satisfactory
to use. Fawcus's medium is praised by some workers but has
not been very satisfactory in the writer's hands.

In general L.B.A., with or without the addition of further
sugars, etc., is recommended as the most generally useful, while
fuchsin agar is also decidedly valuable.

3. Identification. All suspicious colonies found on the
different media in the plates are subcultivated into broth, and
incubated at 37° C. until next day. They are all then examined
in hanging drop, and those which show actively motile bacilli
are tested with anti-typhoid serum. A fairly powerful serum
should be available, and a dilution of not less than one per cent,
should be employed.

A quicker and sometimes preferable method is to directly
test with the antityphoid serum each of the selected colonies by
rubbing up a little of the colony in a drop of one per cent, serum
on a coverglass. Only those which contain reacting bacilli are
subcultivated.

All those which fail to show agglutination are rejected, while

44 WATER

the tubes or colonies reacting are each subcultivated into litmus-
milk, glucose litmus broth (in a double tube), and lactose-peptone
solution (in a double tube).

All the organisms giving cultural characters in these media
which accord with those of B. typkosus are then fully worked out.
The tests should include accurate and extended agglutination
tests with highly dilute sera.

Some such procedure as the above will rapidly decide whether
any typhoid bacilli have been isolated.

Examination for Spirillum cJiolerae.

Of the different methods suggested, the simplest and most
satisfactory is the enrichment method employed by Koch and
others. This is best done by converting the water itself into a
nutrient medium by the addition of peptone and salt.

About a litre of the water is placed in twelve large sterile
flasks, 90 c.c. in each. To each is added 10 c.c. of a sterile
solution, consisting of 10 per cent, peptone, and 5 per cent,
sodium chloride. The flasks are then incubated at 37° C. An
alternative plan is to use only one flask containing I litre of
the water and to add the peptone salt solution to this. Even
quantities of ten litres can be treated in this way or, if incu-
bators of sufficient size are not available, the water can be dealt
with in a series of smaller bottles (2 to 4 litre capacity). After
six, twelve and eighteen hours' incubation, microscopic prepara-
tions and examinations in hanging drop are made from the
surface of each flask. The medium is one in which the cholera
spirillum grows very rapidly, and if present, is found in the very
thin pellicle on the surface of the liquid.

According to Koch, after six hours' incubation is the most
favourable time to examine but sometimes it is necessary to
wait longer. The flasks which show the presence of vibrios are
used to inoculate plates of the media selected, a loopful of the
fluid being withdrawn from the surface for this purpose.

When large quantities of water are examined some of the
concentration methods described as suitable for the isolation of
the typhoid bacillus may be used.

WATER 45

Agar and gelatine media have been most used for the isola-
tion, but recently a large number of special media have been
advocated for this purpose. While a number of them are use-
ful, that of Dieudonne, published in 1909, would appear to be
most valuable. It is an alkaline blood agar medium the exact
preparation of which is described in the Appendix.

The plates ought not to be used immediately after their
preparation. Dieudonne recommends keeping them for several
days in the incubator at 37° C. uncovered and face down, or to
heat them for five minutes at 65° C. According to the Report of
the International Commission1, an equally good result is obtained
by keeping them for 48 hours at laboratory temperature. During
this time the surface of the agar becomes slightly dry and loses
a part of its alkalinity. Once in condition the plates ought to
be used within a period not exceeding five or six days. Cholera
vibrios grow abundantly on this medium while the bacilli of
typhoid fever and dysentery and in particular the B. colt group
organisms grow either very badly or not at all.

On the other hand, B. proteus and B. pyocyaneus, both en-
countered frequently in diarrhroeal stools, grow nearly as well as
the cholera vibrio and considerably complicate the search. Also
many of the non-choleraic vibrios found in water and excreta
appear to grow in this medium. According to Dieudonne, how-
ever, most of these vibrios are suppressed.

The colonies of the cholera vibrio on this medium are trans-
parent and greyish with a glistening appearance by reflected
light but are not definitely characteristic. When many other
bacteria are present the identification of the cholera colonies
may be very difficult.

Ordinary nutrient agar is superior to gelatine but the agar
must be alkaline. The agar colonies are not very distinctive,
being flat discs, transparent and of a grey-blue colour.

All suspicious colonies on the agar, Dieudonne agar or other
medium used, are subcultivated and their characters and proper-
ties studied in pure culture. An alternative plan and one which

1 " Report presented to the permanent committee of the International Office of
Public Hygiene in the name of a Commission presided over by Dr Rufier." Trans-
lation : U. S. A. Public health, Reports, 1912, vol. xxvn, p. 371.

46 WATER

saves time, but not labour, is to directly test every suspicious
and possible colony with powerful anti-cholera serum. All the
vibrios which are agglutinated are subcultivated and their cultural
characters and serological properties tested in detail.

Houston1 in his investigations upon the vitality of cholera
vibrios in water and their isolation therefrom used agar, bile-
salt agar and Drigalski-Conradi agar. The bile salt however
considerably inhibits growth and only very small red colonies
are produced by the cholera vibrio.

If a large number of possible colonies have to be sorted out,
it is essential to have a few cultural tests which rapidly eliminate
the majority of those which are not cholera vibrios. Houston's
procedure was to subcultivate all possible colonies into saccha-
rose peptone water in double tubes. Only those which showed
acid without gas formation after 24 hours at 37° C. were further
investigated. Other important eliminating tests used by him
were growth in peptone water (for cholera-red reaction) and
upon gelatine slope. Failure to give a cholera-red reaction in 24
and in 48 hours, and rapid liquefaction of gelatine were sufficient
to exclude the organisms as not cholera vibrios. The organisms
which passed these tests were subjected to further cultural
investigation.

Since a considerable number of vibrios which closely resemble
the true cholera organism have been found in water it is essential
that all available tests be employed before an organism is accepted
as Sp. cholerae.

Morphological, cultural and pathogenicity tests are valuable
in the diagnosis of cholera vibrios but they are not sufficient in
themselves, and vibrios have repeatedly been isolated from water
which, as regards these tests, cannot be distinguished from those
isolated from the intestines of cholera cases but which on other
grounds cannot be considered true cholera vibrios.

By the use of additional tests, particularly agglutination tests
and the immunity reactions of PfeifTer and of Bordet, accurate
differentiation can be arrived at.

The International Office Report above referred to lays down
the following in regard to agglutination tests, based upon the

1 Fourth and Fifth Research Reports, Metropolitan Water Board.

WATER 47

work of Kolle and Gotschlich and others. This Report states,
4' It appears, therefore, that a rule might with advantage be
adopted to regard as choleraic every vibrio which is agglutinated
jn x — IOoo at least by a serum the activity of which is I — 4000
or over. For vibrios agglutinable by a cholera serum only in
stronger dilutions (i : 500 to I : 1000) the results should be con-
sidered as doubtful."

Other tests which have been shown to be of value are the
haemolysis test and estimation of deviation of complement.

A description of these tests would be out of place here, and
general and special text-books of bacteriology must be consulted.

Significance and interpretation of results.

The detection of the cholera spirillum or the typhoid bacillus
In a water, in whatever amount, is sufficient to condemn the
supply. The other results obtained in the bacteriological exam-
ination of water supplies are, however, only data from which an
opinion upon the purity or contamination of the water can be
deduced with more or less confidence according to the facts
available.

Such deductions require much special experience and a
careful consideration of a number of facts, some of which will be
briefly discussed.

As already explained, the contamination of water with excreta
or sewage is mainly dangerous (although not entirely so) from
the possibility that sooner or later it may be associated with con-
tamination with excrementitious matters containing the typhoid
bacillus and in this way an outbreak of enteric fever be set up.
Such specific pollution with the typhoid bacillus is almost invari-
ably associated with the addition of B. coli and streptococci to
the water, and while the typhoid bacillus is not readily detected
the existence of danger may be shown by these indicator organ-
isms. The vitality of B. typhosus in water and particularly the
relative vitality of this bacillus and B. coli in water is therefore of
great practical importance. On its negative side it is necessary to
know how far the absence or relative absence of bacilli of the
B. coli group can be accepted as presumptive proof that typhoid

48 WATER

bacilli are absent and the supply in its present condition a safe
one.

Viability of B. typhosus in water.

The viability of the typhoid bacillus in water has been made
the subject of many investigations. Only a few of the more
recent and extended can be mentioned here.

The results of many of the investigations are conflicting. In
explanation of the discrepancies several points may be urged. In
the first place, many of the experiments have been carried out
under highly artificial conditions. For example, the bacilli have
been added to water small in bulk and kept confined in vessels,
such as in flasks or bottles. Typhoid bacilli in such circum-
stances are not under conditions of light, movement and compe-
tition with naturally occurring bacilli such as must take place
under natural conditions. Deductions from such experiments
furnish no real guide as to the viability of the typhoid bacillus
under actual practical conditions of pollution of a water supply
by material containing this bacillus. When water supplies are
contaminated with typhoid bacilli the vehicle of infection is
usually the urine or faeces of cases. Such material generally
contains vast numbers of other bacilli. It is also probable that
the viability of any added bacilli is considerably influenced by
the addition of the considerable amount of organic matter which
usually in this way gains access to the water.

A further factor of undoubted importance is the individual
resistance of the bacilli added to the water. Many of the earlier
.experiments were conducted with typhoid bacilli which had been
cultivated in the laboratory outside the animal body for long
periods. Houston1 found that typhoid bacilli direct from the
animal organism died much more rapidly in river water than strains
previously cultivated in the laboratory. In 13 experiments with
the "uncultivated" typhoid bacilli the bacilli could not be found
in the infected water after one week (9 experiments), after two
weeks (3 experiments) and after three weeks (i experiment):
yet the same microbe after " cultivation " usually lived over five
weeks.

1 Seventh Research Report, p. 3.

WATER 49

Of recent investigations upon this subject the following are
two of the most important.

Jordan, Russell and Zeit1 carried out three independent series
of experiments which later were re-investigated by Russell and
Fuller2. They used recently isolated B. typhosus strains and the
bacilli were suspended in the water in celloidin or parchment
sacs. The waters used for experiment were lake and river water
and the impure Chicago Drainage Canal water.

Jordan, Russell and Zeit concluded that " under conditions
that probably closely simulate those in nature, the vast majority
of typhoid bacilli introduced into the several waters studied,
perished within three to four days." They however suggested that
specially resistant cells may be able to live for longer periods.

Their experiments show the relative rather than the ultimate
disappearance of the bacilli, since the whole bulk of fluid was not
examined for the typhoid bacillus but only one c.c. or other small
amount.

Houston3 has recently carried out some extended investiga-
tions. In one series " cultivated " bacilli were added to raw
river water stored in the laboratory in partially filled stoppered
bottles. In these experiments the vast majority of the bacilli
perished within one week, but a few specially resistant strains
persisted for several weeks, the final extinction of the bacilli (as
judged by inability to isolate it from 100 c.c. of the water) only
taking place after nine weeks.

In a second series the typhoid bacilli were added to the water
direct from actual typhoid bacilli carrier cases, the bacilli being
added either directly centrifugalized from urine or in the urine
itself.

In the first experiment the first week effected a percentage
reduction of 99*99 to 100 in the number of typhoid bacilli, their
ultimate death (as judged by inability to isolate it from 100 c.c.)
taking place either within one week or by the second week.
Subsequent experiments gave similar results but in one experi-
ment, using an outdoor tank, the uncultivated bacilli were still
found in looc.c. after three weeks.

1 Journ. of Infectious Diseases, 1904, vol. I, p. 641.

2 Ibid. Supplement No. 2, Feb. 1906, p. 40. 8 Seventh Research Report.
S. \V. 4

50 WATER

Houston also carried out some experiments upon the com-
parative vitality of the typhoid bacillus in raw Thames river
water at different temperatures. He found that temperature
exerted a powerful influence, the bacillus living considerably
longer at low (32 — 42° F.) than at higher (50 — gS'6° F.) tempera-
tures.

It may be said, as a general result from the experimental
evidence, that typhoid bacilli do not live for long periods in
drinking water, their total elimination being a matter of only a
few weeks, while the vast majority are destroyed within a few
days. At the same time it is hardly justifiable to assume that
all typhoid bacilli gaining access to water will be eliminated
even within a month in all classes of water. Such a conclusion
is probably true as applied to the storage of raw river water, but
with local well water contaminated with the faeces of a typhoid
fever case or carrier the conditions — particularly as regards the
nutritive condition of the water — are considerably altered from
those made use of in the experimental work and deductions from
these experiments may not apply. It is quite possible that the
bacilli may persist for considerably longer periods.

Wilson and Dickson1 added a mixture of typhoid "carrier"
urine and faeces to 60 litres of water in an open reservoir and
recovered the typhoid bacillus after three weeks and two days.

As regards the comparative vitality of B. typhosus and B. coli
in water, comparative results show clearly that the former is
the more delicate organism and dies out first. The proved
absence of the second organism may therefore be taken — since
almost invariably contamination with B. typhosus is accompanied
by much heavier contamination with B. coli — as evidence of the
absence of the typhoid bacillus. The only exception would be
when the contamination was by the urine of a typhoid case or
carrier containing only typhoid bacilli and no B. coli.

Bacterial content of different classes of waters.

Before the significance of the results of the bacteriological
examination of water samples can be properly appreciated

1 Journ. Royal Sanitary Instit. 1911, vol. xxxu,.p. 472.

WATER 51

certain general and special facts in regard to the bacterial
content of different classes of waters must be considered.

For present purposes water supplies may be divided into the
following groups :

A. Upland surface waters.

B. Deep water supplies — deep wells and springs.

C. Shallow wells and subsoil waters.

D. River water.

Each class of water must be separately considered.

Upland surface waters.

Supplies of this class are derived from rain water which has
been in contact with soil, usually uncultivated soil, but which
have not undergone any filtration through soil. Rain water itself
when pure contains but few bacteria and no excretal indicator
organisms. The bacterial content of upland waters will vary
with the bacterial nature of the soil they wash, the degree to
which they are liable to pollution from sources other than soil
and to some extent with the degree of storage to which they
have been subjected.

Upland uncultivated soil may contain numerous bacilli but
is free from B. coli and other indicator organisms. The bacterial
content of water in contact with such soil may be considerable
but it will not contain B. coli.

Moorland soil of this character is however usually used for
grazing to some extent and is the habitat of wild fowl and
other birds. The excreta of these birds and animals contain
numerous B. coli and other indicator organisms and thus serve as
a source of such organisms to the water. As a matter of prac-
tical experience therefore it is found that many upland surface
waters contain these indicator bacilli in considerable and some-
times in large numbers, and this even when the uplands which
serve as the gathering area are remote from human habitations
and free from all risk of contamination from human sources.

As illustrations of results obtained with such waters in which

4-2

52 WATER

topographical examination showed no possibility of human con-
tamination the following may be quoted :

B. coli

Source1 organisms per c.c. group organisms in

0'5 2 10 40 C.C.

- - + +

- - - +

37° C.

20° C.

Pure upland

Main contributory river

4

226

surface supply on

218

old red sandstone.

» » J>

Sheep and other ^
animals grazed but
no other source of

„ |
mile above the previous
two samples

!

2

188

contamination Spring rising near

4

458

Large upland
supply with two

r

345 mil. gallon reservoir

8
42

218
850

main collecting

reservoirs. Liable

222

36

410

to sheep and bird ^

& it it y J 99 ,,

65

995

from cultivated

reservoir

I32

35°

land or other

Small stream entering

:

2OO

sources \ reservoir

4

- +

It is obvious that for this class of waters bacteriological
analyses must be interpreted with considerable caution and only
in the light of all the topographical and other data available.
While the addition of animal excreta is not of anything like the
same potential harmfulness as human excreta, yet in considerable
amount it is unsatisfactory and should be prevented. It is there-
fore justifiable to look adversely upon water supplies of this
character which contain a large number of B. coli and possibly
other indicator organisms. Each case requires careful considera-
tion but in general the presence of * excretal ' B. coli in I c.c. or
less of such waters points to heavy and undesirable pollution
while if less numerous special topographical investigation is
required,

Deep water supplies.

Waters of this class — deep well or spring — are derived from
rain which has filtered through a considerable depth of soil and
which has usually taken a considerable time over the process.
The deeper layers of soil are germ free and in its passage the
bacilli are filtered out of the water.

1 For a detailed consideration of the bacterial content of upland surface waters,
especially in relation to topographical findings, see Savage, 1902, Journ. of Hygieney
IT, p. 320.

WATER 53

Deep well waters when uncontaminated contain but few
bacilli and no B. coli or other indicator organisms. In waters of
this class therefore a considerable increase in the number of
organisms and in particular the presence of B. coli, with or with-
out streptococci, must indicate pollution with surface water or
other undesirable material and much higher standards of require-
ments are reasonable. It is justifiable to maintain an attitude
of great suspicion towards any water from such sources which
contains B. coli in 100 c.c. or less.

To ascertain the precise source of contamination and to
measure and assess its danger often requires a very careful study
of topographical and geological conditions and the making of a
series of subsidiary bacteriological examinations.

One important cause of bad bacteriological results from deep
water supplies, which it is important to eliminate or confirm at
the outset, is contamination of the supply at the surface outcrop.

Frequently a spring runs some distance on the surface before
it is collected and utilised, and samples taken at a point on the
distribution side may entirely owe their bad bacteriological results
to surface contamination of an otherwise pure deep water supply.

It is important to realize how frequently such surface con-
tamination occurs and is allowed to exist owing to the care-
lessness or want of knowledge on the part of those entrusted
with the care of the water supply. The following is a striking
illustration of such contamination and its influence upon the
bacteriological examination results.

Samples were received in November 1912 from three separate
springs in connection with a proposed new water supply. All
three springs were from the same formation and indeed from the
same hill. No. 2 was satisfactory while No. I and No. 3 showed
evidence of contamination. A careful examination of the local
conditions was recommended. At first it seemed difficult to
account for any difference in quality since the three springs were
obviously coming from the same strata and in each case were
piped for about 100 yards from the springs, glazed open-jointed
earthenware pipes being used. Further examination however
disclosed the fact that in springs I and 3 the pipes had been
covered with a layer of bracken by the workmen employed

54 WATER

(contrary to instructions)' and the trenches were not filled with rock
and earth as carried out for spring No. 2. When this bracken
was removed and the trench satisfactorily filled with rock and
earth these springs yielded bacteriologically pure water.
The analyses made were as follows :

Organisms per c.c.

at at B. coli Streptococci

Source Date 37° C. 2i°C. (per litre)

No. i Spring Nov. 1912 6 30 1000 — 10,000 1000 — 10,000

(atypical)

» 2 ,, ,, ,, 13 35 absent in 50 c.c. absent in 50 c.c.

,, 3 ,, ,, ,, 8 25 1000 — 10,000 100 — 1000

(atypical)

,, i ,, March 1913 2 5 absent in 50 c.c. absent in 50 c.c.

" 3 »> 5> » 4 »> >» » »»

Admixture with surface water is particularly liable to occur
in samples of spring water taken in connection with proposed
new sources of water supply, since in such cases usually the only
samples which can be collected without expense are those ad-
mixed with surface water where the spring outcrops.

The following is an interesting illustration of the value of
careful investigation of deep water supplies and the use of
bacteriology to detect surface contamination.

The public water supply of a small town showed on bacterio-
logical examination distinct evidence of contamination (sample
No. i below). The supply was derived from three springs in the
old red sandstone, the water being conveyed to a brick-covered
reservoir (36,000 gallons capacity) by agricultural pipes. A
separate analysis was made of the water from each spring (Nos. 2>
3 and 4) while a further analysis of the water in the reservoir
(No. 5) was still not satisfactory. The springs were, as shown
from these and other analyses, mostly satisfactory and the un-
satisfactory condition of the supply was due to surface water
passing into the pipes and so into the main supply. Another
spring known to be liable to some contamination and excluded
from the supply was examined (No. 6) at the same time and
shown to be unsatisfactory. Steps were taken to put in proper
pipes and cut off all surface water and the latest analysis (No. 7)
was quite satisfactory.

WATER

55

Organisms

Date of growing

No.

Source

Exam. at 37°

at 21°

i

Service reservoir

Jan. 1912 no

250

2

Spring A

Feb.

4

30

3

B

,,

3

21

4

c

,,

9

80

5

Service reservoir

»>

4o

I2O

6

Spring D

„

41

120

7

Supply

Jan. 1913 18

90

B. coli Streptococci

(per litre)

1000 — 10,000 30 — 100

absent in 50 c.c. absent in 50 c.c.

30 — 100
absent in 50 c.c.

30 — 100
100 — 1000

A source of contamination sometimes met with as regards
deep wells is due to surface water gaining access either by the
side of the bore hole down to the deep supply or from the im-
pervious lining being faulty or not extending down far enough.
The following is an instructive illustration of this from the point of
view of bacteriological examinations. This water supply, which
supplies a considerable population, is derived from the mountain
limestone and is obtained from a well 3 1 feet deep with an
impervious lining and passing through a thin impervious layer
(lower Lias). Usually bacteriological analyses gave satisfactory
figures but after very heavy rain unsatisfactory bacteriological
results were obtained (e.g. Nos. 2, 3 and 4 below). Considerable
difficulty was experienced in tracing out the cause of the con-
tamination, but it was ultimately traced to surface water (from
the Fuller's earth formation) under certain conditions rising up
and passing down by the side of the bore hole to the deep supply.
Since steps have been taken to stop this source of contamination
the results have been very satisfactory (Nos. 5, 6 and 7) although
samples have been purposely taken after heavy rain.

B. coli Streptococci

(l er litre)

*absent in 10 c.c. absent in 10 c.c.
absent in 50 c.c.

Date of Organisms per c.c

Sample

examination 37° C. 21° C.

i

August 1911 20 .50

2

Jan. i, 1912 3 2500

3

„ H „ 57 170

4

March 1912 220 350

5

June „ 2 25

6

Sept. ,, 2 10

7

Dec. „ 5 70

IOO — IOOO
IOO — IOOO

absent in 50 c.c.

1000 — 10,000
absent in 50 c.c.

[* By inadvertence larger quantities were not examined.]

In addition to easily ascertained sources of surface contamina-
tion, deep water supplies are sometimes liable to contamination

56 WATER

which is very difficult to locate and for which bacteriological
examinations are of the utmost value. Such cases are usually
met with in supplies from the limestone or chalk and numerous
examples might be quoted.

The following is a good example of intermittent contamina-
tion of .a deep well in the chalk reported by Richards and
Brincker1. The well yielded from one to two million gallons per
day. The only source of contamination was from surface water
rather more than two miles away which was shown to gain access
to the deep well water. This was proved by adding a special
yellow bacillus to the surface water and tracing it into the well.
It was found that the test organism took in one experiment 78*5
hours and in the other 67*5 hours to pass from the s wallet hole
to the well. Its presence was associated with a marked increase
in the bacterial content of the water, together with the finding
of B. coli, in loc.c. or less of the water. The latter organism
was absent from the well when no pollution was taking place.

Shallow wells and Subsoil water.

Obviously bacteriological findings will vary greatly with this
class of waters and very widely varying results are in fact obtained.
Sources of bacterial contamination are mainly two, one due to
the entrance of bacteria through the mouth or round the sides
of the well and the other from bacterial contamination from sub-
soil water, due to imperfect filtration of bacteria through the soil.
In considering bacteriological analyses of surface wells it is always
most important to discriminate between these two sources of
pollution.

Contamination through the open mouth (in draw wells) or
from imperfect fittings and covers round pumps is extremely
common and is usually very heavy. Indeed it is very excep-
tional with open draw wells for their water to show other than
gross contamination, and this whether or no the subsoil water
itself is contaminated. When such wells are properly covered
and their sides are effectively rendered to a depth of 1 2 ft. or so,
no harmful bacterial contamination is sometimes subsequently
present, showing that the subsoil water is not itself contaminated.
1 Proceedings Roy. Soc. Med-t Efidem. Sec., 1908, I, p. 191.

WATER 57

The following analyses from a well treated in this way between
the two analyses illustrates this point.

Organisms "Excretal" B. coli

Sample per c.c. c.c. Streptococci
37° C. 21° C. o'i i 10 40 i 10 40
With defective pump-
covering etc. ... 2350 over 5000 + + + + + +
After impervious covering

and rendering ... 9 2020 _ _ _ _

The bacteriological results1 obtained with surface wells are
variable, but if properly protected from surface contamination
and from the entrance of water which has not filtered through
at least 12 ft. of soil they will usually show no excretal indi-
cators in 50 c.c. and certainly not in 10 c.c. If " excretal " B. coli
are present in loc.c. or less of a surface well water it is evidence
of undesirable and possibly dangerous contamination. Even
when only present in 50 c.c. the well and its surroundings should
be subjected to very careful examination. The influence of rain-
fall upon the bacterial content of surface wells is considerable
and must be kept in view.

The variable bacterial content of surface wells makes it
peculiarly dangerous to pass well waters as satisfactory from the
results of a single satisfactory bacteriological analysis. The
filtering action of the soil may be for the time satisfactory, but
at any moment it may break down and harmful bacteria be
washed through into the well water.

For example a recent bacteriological analysis of a well water
supplying five houses gave the following results : — Organisms
growing at 37° C. and 21° C. = 2 and 15 per c.c. respectively.
No B. coli or streptococci in 50 c.c. : larger amounts not examined.
This well was a shallow one sunk in porous ground and surrounded
up to a few yards by heavily manured gardens. It was on
inspection obviously liable to pollution, while a chemical analysis
disclosed a very high nitrate content (2*0 per 100,000) but low
ammonia figures. It is quite clear, as the very good bacterio-
logical analysis showed, that the soil was acting as an efficient
filter at the time the samples were collected. The bacteriological
data did not, however, disclose that the good quality of the

1 For numerous analyses and data see Savage, Journ. of Hygiene, 1907, vn, p. 477.

58 WATER

water depended upon an efficient soil filtration, and that this
might break down at any time and without warning.

River Water.

The terms river and stream cover so many conditions
which differ very widely that obviously no possible standards
can be forthcoming. Rivers, receiving as they do the washings
from the lands they flow through, usually show a high bacterio-
logical content, and as the land so draining is often cultivated,
a high content in excretal indicators such as B. coli and
streptococci. Such contaminated waters obviously cannot
be considered as satisfactory sources of supply for drinking
unless purification has taken place by sedimentation (storage)
or filtration. Their bacterial purity can be judged with great
accuracy by their content in the B. coli group of organisms.

Bacterial Standards.

From the above considerations it will be evident that no
very definite bacteriological standards can be framed, even when
each class of water is separately considered. At the same time
it is quite possible to frame working guides, as indicated above,
and if these are used with reasonable care to obtain most
valuable opinions upon any given water supply.

Of the different data available to form such an opinion the
B. coli group enumeration is by far the most valuable, the
other findings being more or less confirmatory. The above
considerations set out as a reasonable requirement that "excretal"
B. coli should be absent from 100 c.c. of deep water supplies and
from loc.c. of surface waters.

Sometimes the organisms isolated are not typical B. coli, but
differing in the absence of one or more of the characteristic
properties of this organism. In the writer's opinion the nearer
these lactose-fermenting coli-like bacilli approach typical B. coli
in their characters, the more nearly are our numerical standards
for that organism applicable to them, while if they lack essential
characters a proportionately greater number must be present to
justify an adverse opinion.

WATER 59

Determinations of the number of streptococci have been made
much less frequently than in the case of B. coli. As a provisional
guide, and without attaching an equal significance to the findings,
a standard similar to that for B. coli may be employed — i.e.y
their presence in 100 c.c. or less of deep- well or spring water, or
in 10 c.c. or less of surface and shallow-well waters, would justify
an adverse opinion as to the purity of the water in question.

On its negative side the streptococcus test is not of great
value, and the absence of streptococci, even in a considerable
bulk of water, cannot be taken as showing purity or freedom
from danger.

Opinion is not united as to the value of B. enteritidis
sporogenes as an indicator of pollution. It is fairly abundant
in sewage and excreta, but it is a spore-bearing organism with
prolonged powers of resistance, and therefore, even if it be
admitted that its presence indicates pollution, such pollution
may have taken place at some long antecedent period, a con-
tamination so old as to be of no significance. Its absence in a
large quantity of water is some evidence of purity.

The essential limitation to the value of bacteriological
examinations is that they only supply information as to the
existing conditions in the water at the time of sampling. They
cannot indicate liability to contamination which is not actually
taking place. This important limitation must never be lost
sight of, and it sets a decidedly restricted value upon the
bacteriological examination results of chance samples.

When the bacteriological data discloses distinct evidence of
undesirable and potentially harmful contamination it is safe to
report adversely upon such a supply, since — unless the sources
of contamination can be and are removed — once contaminated
always liable to contamination. Even in such cases a con-
demnatory report comes with much greater weight if more than
one sample has been found to be polluted.

It is on its negative side that caution is required. As stated
above contamination is often intermittent and chance samples
may not disclose it. Opinions on individual samples in such
cases should therefore be always statements of fact and not of
inference as to the purity of the particular supply.

6O WATER

Bacteriological Testing of Filter-Beds.

The testing of the efficient working of sand filter-beds is entirely
a bacteriological matter. All filter-beds should be systematically
tested to ascertain the percentage of bacteria removed.

The purification by filtration through sand is only to a small
extent mechanical ; it is mainly vital. This vital or biological
action is due to the formation, which takes place after a few
days of working, of a gelatinous layer on the surface of the
sand. This is composed partly of suspended matters and partly
of bacteria, algae, and other lowly forms of vegetable life, derived
from the water filtered. Such a filter-bed is capable under suit-
able conditions of removing the vast majority of bacteria from
the water. The percentage removed depends upon a number
of factors, of which the following are the most important : The
rate of filtration, the age of the filter, the depth and size of the
sand particles, the kind of filtration — i.e., whether intermittent or
continuous — and the nature of the water filtered.

Of these, the first two are the most important. If water is
passed rapidly through a filter, the percentage of bacilli removed
will diminish. Koch recommended that the rate should not
exceed 4 inches per hour. The rate of filtration must however
be largely governed by the quality of the water being filtered.
The age of the filter is in the main a question of the thickness
of the gelatinous layer. The percentage of organisms removed
should be 98 or more.

To estimate the bacterial efficiency, or percentage of organisms
removed by the filtration, gelatine and agar plates are made
from the water before and after filtration, and examined and
counted in the ordinary way. Each bed should be separately
tested. The rate of filtration and all particulars must be
recorded at the same time. The percentage removal of B. coli
should also be ascertained, and is on the whole a better test of
the efficiency of filtration than ordinary bacterial enumerations.

SOIL AND SEWAGE 6l

CHAPTER IV

SOIL AND SEWAGE
SOIL

Surface soil contains a vast number of organisms. With
increased depth the number rapidly diminishes, and below a
metre but few bacteria are to be found in undisturbed soil.

The rapid diminution in the number of bacteria in soil was
first clearly established by Fraenkel and has since been confirmed
by numerous observers. The following figures will give a good
idea of the number of bacteria in soils.

Houston1, working with soil in the grounds of Morningside
Asylum, Edinburgh, on a plot of land which was formerly a
vegetable garden, but which had lain untouched for some time,
found, as the result of a large number of experiments, that the
average number of germs in I gramme of soil was on the
surface about 1,688,000; at a depth of I foot, 1,100,000; 2 feet,
900,000; 3 feet, 174,000; 4 feet, 25,000; 5 feet, 920; and
6 feet, 410. These figures deal only with the numbers which
will develop on gelatine media, and do not give any true
idea of the total number of bacteria actually present, excluding,
as they do, for example, the vast number of nitrifying organisms.
They, however, illustrate the rapid decrease with increased soil
depth in the number of organisms which grow on gelatine media
under aerobic conditions. In Fraenkel's researches anaerobic
bacteria were also found to be absent, or relatively absent, in
the deeper layers.

The number also varies with the kind of soil, particularly
whether virgin or cultivated. Thus, Houston2, who examined
twenty-one samples of surface soil from different sources, found
that the virgin sandy soils gave less than 100,000 bacteria per
gramme, the other virgin soils about 1,000,000, the garden soils
from 1,000,000 to 2,000,000, and two grossly polluted soils,

1 Edinburgh Med. Journ. 1893, xxxviil, part ii, p. 1122.

2 Local Gov. Board Med. Officer's Report, 1897-8, p. 251.

62 SOIL AND SEWAGE

in one case 26,000,000, and in the other 115,000,000 bacteria
per gramme.

The number of spores relative to the total number of
bacteria in soil is large, the proportion being often as high
as from I to 10 to I to 3.

The presence of organic matter undoubtedly has some
influence upon the number of bacteria in soil.

Reimers1, examining graveyard soil, found an increased
number of bacteria in the vicinity of the coffin, the greatest
number being met with some distance above it.

Young2 examined bacteriologically 23 samples of soil all taken
from graveyards, some being from soil which had never been
disturbed, others being taken from the vicinity of old interments.
He found that the number of bacteria present in soil which
had been used for burial exceeded the number in undisturbed
soil at similar levels, and that this excess, though apparent at
all depths, was most marked in the lower reaches of the soil.
Roughly speaking, at the lower depths (8 — 9 feet) the soil
used for burials contained about twelve times as many bacteria
as the undisturbed soil and at moderate depths (4 — 6 feet) about
six times as many bacteria.

The bacteriology of soil has been very incompletely worked
out, but in addition to the nitrifying bacteria, which are only
isolated by special methods, the saprophytic bacteria which are
commonly met with are Cladothrix dichotoma, B. proteus vtdgaris
and other proteus strains, B. subtilis, B. mycoides, B. megatherium,
B.fluorescens liquefaciens, B.fluorescens non-liquefaciens, B. arbores-
cens, and micrococci of different kinds.

The bacteriological examination of soil from the public
health point of view is of but limited utility, its chief value
being in connection with the contamination of water from
surface washings.

Pathogenic bacteria in soil.

Pathogenic bacteria from various sources are constantly
gaining access to soil, and accurate knowledge as to their

1 Zeitschr.f. Hygiene, 1889, vol. vii, p. 307.

2 Trans. Royal Soc. of Edinburgh > 1895, vol. xxxvn, p. 759.

SOIL AND SEWAGE 63

vitality and their retention of virulence in such surroundings
is of great importance. This is particularly important in con-
nection with the contamination of water supplies. Although
numerous investigations have been made it cannot be said that
precise information on this subject is available. No doubt the
chemical composition and particularly the amount of organic
matter in the soil, as well as the number of other bacteria
present, play a very important part in the determination of the
length of life in soil of those pathogenic bacteria which are
not natural inhabitants of the soil.

B. typhosus. A number of investigations have been carried
out by Sidney Martin, Firth and Horrocks, Rullmann, Dempster,
Pfuhl, Savage, Mair, and others, upon the vitality of typhoid
bacilli in soil.

Of British investigators working with unsterilized soils
Robertson (1898) recovered the bacilli after 300 days, Martin
(1896-1901) could only recover the bacilli up to 12 days, Firth
and Horrocks (1902) found the bacilli to survive in some cases
up to 74 days, Lorrain Smith (1903) up to 21 days the average
being 15 days, Savage (1905), in polluted river mud treated
bi-weekly with fresh sea water, up to 5 weeks in one case and
fairly readily up to two weeks, and Mair (1908) in large numbers
for about 20 days and still present after 70 to 80 days.

These results show considerable discrepancies. They show
•that under favourable conditions the typhoid bacillus will survive
for a considerable period in soil, and that the factors influencing
its vitality are many and varied, the antagonism of other microbes
.and the physical conditions of moisture and temperature being
the most important.

B. diphtheriae. This organism does not appear to have any
important relationship to soil, and there is no evidence that
infection has resulted from diphtheria bacilli derived from soil.
The absence of diphtheria outbreaks spread by water is evidence
in the same direction. But little experimental work has been
done upon the viability of the diphtheria bacillus in soils,
although Germano and others have shown that it may retain
its vitality in dust for some time. Leighton has shown that

64 SOIL AND SEWAGE

diphtheria bacilli may remain alive for 18 days in moist warm
modelling clay.

B. pestts. The Advisory Committee on Plague in India
report1 some experiments on this bacillus in soil. They studied
how long floors grossly contaminated with plague bacilli would
remain infective for animals, the infectivity being tested by
rubbing scrapings into susceptible animals. They found that
cow-dung floors remained infective for 48 hours, while floors
composed of a mixture of sand and lime allowed to set did
not remain infective for over 24 hours.

Gladin claims to have recovered the plague bacillus after
two months from unsterilised earth, but Mackie and Winter
(quoted in the above report) found the bacillus (with difficulty)
up to 96 hours after its introduction, but not subsequently
either by culture or animal inoculation.

It would appear that, in general, B. pestis dies out rapidly
from soil.

B. antJiracis. As is well known, spores of the anthrax
bacillus may live and retain for years their virulence in soil
to which they have gained access, and such infected land if
used for grazing purposes may serve as a means of infection
for very prolonged periods.

A number of anaerobic, pathogenic bacteria are common soil
organisms. In particular the bacilli of tetanus and malignant
oedema are widely distributed in soil, especially when cultivated.
For their characters and general methods of isolation, general
text-books on Bacteriology should be consulted.

Excretal bacilli in soil.

The presence and viability of the common excretal bacilli
in soil is a matter of considerable public health importance,
particularly in relation to water supplies. Soils which have
been recently contaminated with organic matter in quantity —
for example, by sewage or manure — show evidence of this when
bacteriologically examined.

Houston2, in a prolonged series of experiments, watered

1 Journ. of Hygiene, 1906, vol. VI, No. 4 (extra No.). P- 509.

2 Local Govt. Board Med. Officer's Report, 1900-1, p. 405, and 1901-2, p. 355.

SOIL AND SEWAGE 65

soil with crude sewage and studied the fate of certain sewage
organisms — B. coli, streptococci, and B. enteritidis sporogenes-^-
in the treated soil. While all these organisms were abundantly
present in the soils immediately after inoculation, all diminished
with time, although their rate of disappearance varied greatly.
The majority of the streptococci very rapidly disappeared, the
B. coli less rapidly, and showed considerable variability, while the
spores of B. enteritidis sporogenes showed but slight diminution
during the period of observation.

Some experiments of the writer1 upon the self-purification of
tipped house and street refuse constituting "made soil," illustrate
the gradual replacement of the B. coli and other organisms,
originally present abundantly in the tipped refuse, by B. mycoides
and other soil organisms. In these experiments "made soil"
(the samples being collected at a depth of 2 feet) was examined,
taken from deposits of known ages. In the samples deposited
within two years B. coli group organisms were abundant, while
the numbers of B. mycoides and cladothrix were either very few,
or these organisms were absent altogether. Jn material deposited
several years previously, on the other hand, no B. coli were
found, while the soil organisms mentioned were very numerous.
The soil organisms had replaced those in the original material.
There was a general relationship between the number of B. coli
and the age since deposit.

The same facts are brought out from the results of the
examination of soils from different sources for these sewage and
excretal bacteria. Houston, Chick, Savage and others have
examined soils for these organisms. In virgin soils typical
B. coli and streptococci are absent, while spores of B. enteritidis
sporogenes are either absent, or present in only very small
numbers. On the other hand, in cultivated and other con-
taminated soils all these organisms are present in numbers
roughly comparable to the extent and age of the contamination.

The following results obtained by the writer2 will serve as
an illustration of findings likely to be obtained.

1 Journal of Sanitary Institute, 1903, vol. XXIV, p. 442.

2 The Bacteriological Examination of Water-Stipplies. H. K. Lewis, London,
1906.

S. W. c

66

SOIL AND SEWAGE

1

<u o <u

.SB 8-Ss

c/5

II III

.<•» "

S*

o

D O <U <U S? O O

B. enter
sporogenes

Gramme

!!"1l55iJjr s ' ttUlf"'

form
nisms

I

OJ

ll i 1|

v^ f^ '"^ O ^

*

II | ^J

h

£1 1|-|2 li!| § i

$«,

14: |.^4ls §§l* 1 §

I

J5
"3

•S

: | : : :^ g :^ :

| a .2 .. , .^^ §

ll 1 l«&
: ^ :-S : : :| s s =^|| : : : : : : :

"^3 tl/3 *^T*"P'(Y5fOiU-i-'^1

s "rH 11 ir ' 3sat|i. ? ,2=2S

^ .5 rO<u'O^- tCcS-TO

r3 -g"2igrt'o 13 -'O^H - ^-^^^

••-^P^ 3rt <Ufi- •• cj J2Crt

•§» "i = u^ 1-2 ' ' 1 ft||i

j> O ^ O PH O U O a:

cL

a> j_,

^ 8

G

Q

3 -i

S

0

<J . . .M , 8J .0 * , .0 ...

c» cf5 ^- uivo t^OO C> O e» f^ ^- 100 r- oo ON 0 ~ r»

fcj

SOIL AND SEWAGE 6/

The bacteriological examination of soil.

Soil may be required to be bacteriologically examined for
three purposes.

A. To isolate and study the organisms of nitrification and
allied processes.

B. To detect the presence of special pathogenic bacteria.

C. To study the degree of pollution with excretal organisms,
in relation to water supplies.

A. The organisms concerned with nitrification are present
in all soils, and their functions are of vital importance. They
are not readily isolated, while their study is chiefly of value for
research purposes. Methods of isolation are not therefore
included here.

B. The pathogenic bacteria which have to be looked for
are those enumerated above, and particularly the bacilli of
tetanus, malignant oedema, and occasionally B. typhosus and B.
anthracis. The methods for their examination are in no way
special and are conducted on ordinary lines. No detailed con-
sideration is therefore necessary here.

C. General Soil pollution. The data which give the most
valuable information with regard to recent or remote soil
pollution are the following : the total number of aerobic
organisms, number of spores present, number of B. coli,
B. enteritidis sporogenes, and streptococci.

In collecting soil for bacteriological examination the depth
from which it is obtained is of fundamental importance. If the
surface soil is to be examined, scrape up with a sterile spatula,
and transfer to a sterile receptacle. To obtain soil from a given
depth either a fresh cutting must be made and the soil collected
at the required depth, or, preferably, some form of borer may
be used. For this purpose Fraenkel's borer is convenient, its
chief drawback being that it holds only a small quantity of soil.

If Fraenkel's borer is used, it is advisable to collect at least
eight samples from spots about a foot apart, and to mix together

5-2

68

SOIL AND SEWAGE

to obtain a representative sample. Also in this way sufficient
soil will be obtained for a concurrent chemical examination.

By means of this borer the exact
depth of the soil taken can be
ascertained. Owing to its length it
cannot be sterilized in the hot-air
oven, but it can be conveniently
and sufficiently sterilized by pouring
in methylated spirit and igniting.
After sterilization wrap the lower
portion in a sterile cloth and secure
with string. This plan is very con-
venient when a number of samples
have to be taken in one day, and
at, perhaps, a long distance from
the laboratory, since the borer can
be resterilized at once before each
sample is taken, it being only
necessary to carry a bottle of spirit
and a number of sterile cloths in
a metal box. The soil is removed
by a sterile spatula from the in-
terior of the borer to the sterilized
tin or other receptacle used for the
soil.

The examination should be com-
menced as soon after collection as
possible.

To estimate the total number of
bacteria, and for some other steps
of the examination, very extensive
dilution must be practised. As an
example of a convenient method
of dilution the following procedure

Fig. 7. Fraenkel's Borer. is given : °ther methods of dilution

will readily suggest themselves. It

is important to remember that owing to a number of inherent
difficulties (such as the difference of coherence of different soils)

SOIL AND SEWAGE 69

numerical estimations are only relatively accurate, and in any
case the same method should be used throughout for each
investigation :

Accurately weigh a small sterile glass-stoppered bottle con-
taining 100 c.c. of sterile water. Quickly weigh in I gramme
of the soil (previously well mixed together) into the bottle,
using a sterile spatula to add the soil. With a little practice
I gramme can be quickly and sufficiently accurately added. Mix
very thoroughly by repeated shaking, if necessary breaking up
the soil by a pointed sterile glass rod. Call this solution
Dilution A. Allow the soil particles to settle, then add I c.c.
or more, according to the suspected contamination of the soil, to
a sterile flask containing 99 c.c. of sterile water. Mix thoroughly
and label Dilution B. Varying quantities of Dilutions A and B
are used for the examination.

To estimate the number of aerobic organisms make gelatine
plates from these dilutions. Thus, o!2, 0*5, I *o c.c. of Dilution B
are convenient amounts to add to the gelatine tubes. For the
number of organisms developing at 37° C. use, in the same way,
agar plates.

To estimate the number of spores, present as such, add
varying amounts of the dilutions to gelatine tubes. Heat to
So0 C. for ten minutes, then plate, incubate, and count in the
ordinary way.

For B. colt enumerations various fractions of the dilutions are
added to tubes of lactose bile salt media. These are incubated
at 37° C., and those in which acid and gas are produced are used
to inoculate solid media, and the organism isolated exactly in
the same way as for the isolation of this bacillus from water.

Streptococci and spores of B. enteritidis sporogenes are
examined for by methods identical with those used for water.

SEWAGE

The bacteriological examination of sewage is not, at the
present day, a procedure widely practised, although for special
purposes it is valuable. It is the only way the potential harm-
fulness of an effluent can be measured. Chemical standards

7O SOIL AND SEWAGE

and examinations, while of great value, are no guide in regard
to the extent to which pathogenic bacteria have been removed
by any process of sewage purification.

Bacterial content of crude sezvage.

As might be anticipated the actual bacterial content of crude
sewage varies enormously, since sewages vary so greatly in
strength and the extent to which they become mixed with
surface water. Speaking generally the number of bacteria
present, as shown by the number of colonies on gelatine plates,
usually ranges from I million to 100 million per c.c. This
number will vary from hour to hour according to the strength
of the sewage.

The organisms used as indicators of sewage and excretal
contamination are all very abundant in sewage, B. coli organisms
about 100,000, streptococci about 10,000, and spores of B. enteri-
tidis sporogenes about 100 to 1000 per c.c.

The number of different kinds of organisms in sewage is
very great, and it is probable that many of them occur in all
specimens of ordinary sewage, but except for the above organisms
their presence has not been ascertained with sufficient constancy,
nor has their numerical occurrence been sufficiently investigated
to enable them to be used as indicators of sewage pollution.
Further investigations in this direction are very desirable.

The organisms of typhoid fever and cholera have never been
isolated with certainty from sewage, although they must fre-
quently be added in large numbers when cases of these diseases
are present. Anthrax bacilli have been found by Houston in
Yeovil sewage, both in the septic tank and in the primary
and secondary coke beds, also in the mud of the banks of the
river Yeo.

Effects of sewage treatment upon the bacterial content of sewage*
The reports of the Royal Commission on Sewage Disposal
have made it abundantly clear that while the different processes
for treating sewage may effect their immediate purpose of
producing a non-putrefying effluent, yet they never yield one
which is sterile or anything approaching it. The results obtained

SOIL AND SEWAGE 71

as regards elimination of bacteria vary greatly with the method
of purification adopted.

Houston1 made a careful investigation for the Royal Com-
mission on Sewage Disposal of the biological qualities of the
effluents from sewage farms.

Compared with the original sewages, all the effluents
exhibited a high percentage degree of purification ; but, apart
from a reduction in number, they showed no very appreciable
biological modification. Houston remarks that sometimes the
relative number of spores of B. enteritidis sporogenes was
reduced, and that there was some evidence, especially in the
better class of effluents, of a greater proportionate reduction,
as compared with crude sewage, in the number of microbes
growing at blood heat, over those growing at 20° C. Also on
several occasions he failed to isolate streptococci from the
effluent.

In general, however, as Houston very definitely points out
(p. 169), "the results conclusively show that the treatment of
sewage on land cannot be relied on materially to modify the
potentially dangerous qualities of crude sewage. The actual
number of objectionable microbes persisting in the effluents is
too great to allow of much stress being laid on the great
percentage reduction effected in the total number of microbes
by the land treatment, or to insure any certainty that effluents
from land processes are * relatively safe.' "

Houston2 also extensively investigated the effluents from bio-
logical treatment processes. He sums up the general outcome of
the experiments as follows : " The effluents from septic tanks,
intermittent contact beds, continuous filtration beds, etc., contain
an enormous number of bacteria. In some cases the percentage
reduction of microbes in effluent as compared with raw sewage
is striking. But as an effluent must be judged by the actual
state it is in, and as the number of micro-organisms still
remaining is nearly always very large, percentage purification
would seem to be of minor importance. In not a few cases the
bacteria are practically as numerous in the effluent as in the

1 Fourfh Report Royal Sewage Commission, vol. IV, part ill.

2 Second Report Royal Sewage Commission^ 1902, p. 25.

72 SOIL AND SEWAGE

raw sewage. 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 reduction of
these objectionable bacteria been so marked as to be very
material from the point of view of the epidemiologist. No
definite proof has been furnished that the effluents from bacteria
beds are conspicuously more safe in this sense in their possible
relation to disease than is crude sewage. Indeed, all the avail-
able evidence tends to show that they must be regarded as
nearly, if not quite, as dangerous to health as raw sewage....
The inoculation of animals with the effluents from bacterial
beds seems to show that they are nearly as pathogenic as crude
sewage."

Judged from the bacteriological standpoint the only process
which yields a markedly purified effluent is filtration of the
sewage through sand. This method is very little used in
England, but is employed to some extent in the United States.
Extended experiments were carried out by the Massachusetts
State Board of Health with sand filters. These experiments
showed that 97 to nearly 100 per cent, of the B. coli could be
removed by these filters, and this whether they were treated with
raw or septic sewage.

Although the percentage purification is very great the actual
number of B. coli remaining is large. Thus the experimental
sand filters at Columbus, Ohio, are recorded by Johnson as
removing 98'$ per cent, of these bacilli, but 500 to 10,000
B. coli per c.c. remained in the effluent.

There does not appear to be much data available to judge
as to how far ordinary chemical processes of sewage purification
are capable of removing and reducing the bacteria in sewage,
but in general it may be said that they are not very effective.
Sterilization or partial sterilization of the sewage can be obtained
by chemical means, e.g. by the use of chlorine evolved in different
ways, but these are processes specially applied for this purpose,
and are not ordinary chemical methods of purification.

SOIL AND SEWAGE 73

The fate of pathogenic organisms in sewage.

The organisms of chief interest are B. typhosns and Sp. cholerae
and a number of investigations have been undertaken to study
their viability in sewage.

Klein1 studied the viability of both organisms in sterile
sewage kept at room temperatures. The typhoid bacillus after
some preliminary multiplication rapidly diminished in numbers,
some bacilli being alive, however, after eight weeks. The
cholera vibrio died out at periods varying from the eighth
to the twenty-fifth day.

Horrocks found typhoid bacilli alive after 60 days in sterile
sewage kept at 16 — 22° C.

When the sewage is not sterilized the life of the typhoid
bacillus is very much shorter, and MacConkey could only recover
the bacilli after thirteen days in one series and not after six
days in another, while Russell and Fuller found that when this
bacillus was exposed directly to the action of sewage bacteria
its longevity was greatly diminished, three to five days being
the longest time for which the organism could be recovered.
The difficulties of isolating typhoid bacilli from crude unsterilized
sewage are considerable, even with recent greatly improved
technique and bacilli actually present may easily be overlooked
and not isolated. After making due allowance for this it is
clear that the typhoid bacillus has a hard struggle to live in
raw sewage and dies out after a fairly short period.

The bacteriological examination of seivage and sewage effluents.

At the present day this is not often required for practical
administrative work. The bacteriological examination of sewage
effluents is occasionally valuable when the question of pollution
of a stream subsequently used for drinking purposes or the
possibility of shellfish pollution is under consideration. The
general methods are quite similar to those described under
water. The chief determinations which will be required are
the number of B. coli group organisms, the number of spores

1 Local Govt. Board Medical Officer's Report, 1894-5, p. 407.

74 SOIL AND SEWAGE

of B. entcritidis sporogenes and the number of streptococci.
Compared with water the B. ent. sporogenes determination is
more valuable and that for streptococci less valuable.

Great initial dilution of the sample is required before the
different fractions are removed for examination, the dilutions
to be used depending upon the degree of probable dilution of
the effluent Dilutions varying from O'l to O'oooooi c.c. should
be examined from unknown samples in routine cases.

It must be remembered that the bacterial content of both
sewage and sewage effluent samples varies from hour to hour. To
get representative results samples would have to be taken every
hour and in amount proportional to the flow, kept in ice until
the end of the 24 hours, pooled and a representative mixed
sample taken for bacteriological examination. It is only in
very rare cases that it is worth while taking all this trouble.
As a rule all that is necessary is to show that samples
of effluent taken at different times contain these indicator
organisms in large numbers, often not materially reduced from
their numerical presence in the crude sewage, and that there-
fore it is a reliable deduction that the sewage effluent in
question is not safe to discharge into any stream which itself
may ultimately be used for drinking purposes or which will
wash food material — watercress, shellfish, etc. — to be used for
human food.

CHAPTER V

SHELLFISH

It has now been conclusively proved that shellfish may
and do convey disease to man, typhoid fever in particular, but
probably also other diseases. This fact has directed considerable
attention to the bacteriology of shellfish and to their bacterio-
logical examination to detect and measure the extent of their
bacterial contamination.

SHELLFISH 75

This bacteriological investigation has been directed along
three lines:

A. To enable bacteriology to be used to judge to what
extent any given batch of shellfish is dangerous to health.

B. To enable judgment to be given as to the safety of any
given shellfish layings.

C. To study the conditions under which dangerously con-
taminated shellfish can be rendered free from their bacteriological
pollution and made safe for human consumption.

The shellfish chiefly concerned in the bacteriological trans-
mission of disease are oysters, mussels and cockles.

In connection with this subject the general bacteriology of
sea water and the sea water of estuaries is of considerable import-
ance.

The bacteriology of oysters in relation to sewage pollution.

The general bacterial content is of but little significance and
the only recorded data of importance refer to the extent to which
oysters contain the organisms used as bacterial indicators of
sewage. The following examples from Houston's work1 give an
excellent idea of this as regards oysters from different sources.

(a) Deep sea oysters. Samples were examined from a num-
ber of different sources. A very few organisms partially resem-
bling B. coli were isolated. Houston concludes : " Judging from
these experiments as a whole, the conclusion seems inevitable
that in deep sea oysters derived from deep sea water remote
from sewage pollution B. coli and coli-like microbes and also the
spores of B. enteritidis sporogenes are either absent or, at all
events, seldom detectable. The same is true of the surface water
over such oysters."

"B. coli and B. enteritidis sporogenes seemingly form no essential
part of the bacterial flora of pure sea water, and they have no
part in the economy of the oyster."

(d) Oysters from estuaries not exposed to serious sewage con-
tamination. The Helford River traverses a sparsely-populated

Royal Comtnission on Seiuage Disposal, Fourth Re for I, 1904, vol. in.

76 SHELLFISH

district, and ranks as one of the purest localities in England for
the growth and fattening of oysters. Houston says it is a river
which "on topographical grounds, has been, and would still be
considered eminently well suited for the breeding, growth and
fattening of oysters for market."

The following results were obtained by Houston with oysters
from this source, while the sea water results are also included for
comparative purposes.

Helford sea water.

B. coli or coli-like microbes in B. enteritidis sporogenes

100 C.C. 10 C.C. I C.C. 10 C.C.

Positive result Positive result Positive result Negative result

28% 40% 3'% ioo °/o

Helford oysters.

B. coli test.

i out of 25 (4°/0) contained i B. coli or coli-like microbes per oyster.
5 >t (20%) „ 10 „ „

16 ,, (64%) »> I0° » " » »

3 »> (I2°/o) >» 1000 » » » »»

B. enteritidis sporogenes test.

1 6 out of 25 (64 °/0) contained less than 10 spores per oyster.
8 j, 25 (32 °/0) contained 10 but less than 100 spores per oyster,
i „ 25 (4%) „ 100 „ 1000 „

(c) Oysters from estuaries seriously contaminated with sewage.
The Penryn River is polluted and its oyster layings lie under the
ban of suspicion. " The Penryn River, on topographical and
epidemiological grounds, would be regarded with great suspicion,
if not condemned."

The following results were obtained by Houston.

£ •_. .'."...<" ' -i

Penryn water.

B. coli or coli-like microbes in B. enteritidis spcrogenes

10 c.c. i c.c. o'r c.c. o-oi c.c. to c.c. 10 c.c.

Positive Positive Positive Positive Positive Negative

result result result result result result

-16% 36°/o 44°/o 4Q/o' 56C/0 44%

SHELLFISH $7

Penryn oysters.

B. coll test.

i out of 25 (4%) contained 100 B. coli or coli-like microbes per oyster.
13 »> (5*%) ». i°oo
ii „ (44%) ,» 10,000 „ „ „ „

B. enteritidis sporogenes test.

3 out of 25 (t2°/0) contained 10 spores of B. enteritidis sporogenes per oyster.
20 „ (So°/0) „ 100 „ „ „ „

2 „ (8%) „ 1000

These three typical groups of results show that B. coli and
B. enteritidis sporogenes are present in oysters in amounts roughly
proportionate to the extent of their pollution with sewage. In
the same way as for water supplies it is not the presence of these
organisms which has to be considered but their relative abun-
dance. Even oysters from reasonably pure sources will frequently
show some B. coli and it is only deep sea oysters which are quite
free. These general conclusions are in accord with the work of
Klein, McWeeney, Clark and Gage, etc.

As regards standards for comparative purposes Houston1 has
given the very useful classification contained in the table on p. 78.

The bacteriology of mussels in relation to sewage pollution.

Very few data seem to be available as to the bacteriological
content of mussels from perfectly pure surroundings. Some of
the earlier workers (e.g. Herdman and Boyce) found B. coli
organisms to be usually absent but these results were probably
due to incomplete examinations and are not in accord with later
experience.

Johnstone2 from extensive work states : " my own experience
has been that no sample of ten or more mussels can be
examined without finding Bacilhis coliy or at least some organ-
isms resembling this form." Even with mussels examined well
away from considerable sewage contamination (such as at some
beds in Morecambe Bay) B. coli were numerous. In the case of
ten mussels from this source the number of intestinal bacteria
isolated from about 0*2 c.c, of the stomach juices varied from
3 to 65, average 24*2.

1 Jo urn. oj Hygiene, 1904, vol. IV, p. 182.

2 Journ. of Hygiene, 1910, vol. ix, p. 412.

SHELLFISH

L-s ,

J 0 £ oT~ g ^

.1 ^If^l'l.sl^

' •'..:' •

•5'fl'a 0."2 rt u

O O'.H °Jj(u-~.^<u22

|rt§ «

fe "§•£ 1 J* a § ^

*

cL. o

llllfll

•a S j 1 1 g 2

p .. ~g ^S « aJ (3

.55.2^ g«= 2 >

Oi JH *r? C ^ fl O <4-*

Mn fir

ga

£ 3 gj<"»

HJ 4>

*«s £ 52 «> — • • <u

"sj -« i

• "<SJ2 g s^js "§ S1^ Il«« S^o"^

ll^lll

S c Mw "== "

•3 s^ §• H "3

Ji "^ A"0^^* ^^ oo "7- ""* 5 ^. ^^^ •» *o>^ 0 «

i-gl'8^11

X K^CCO •S'Soco ^'Gco Sro WS°>« ST &^ ^ "*

il§p

15 8 g|2

1^1 £t Ppi |lif 1

lN-S=3l

^

Li

CJ

S3

<u

_S

ard based on numer
lance of £. coli (or i
fying, gas-forming c
microbes) in the wh
ents of the oyster-sh
iquor, body and inte
uices of the oyster)

1

0

o"
|

1

1

I,

o

S,

1

1

b
B

O,

1

*r*

! j |.| § 7"

«

O

8

w

tf

to « 13""* >!i.

o§

-s,"

i •» flj
O <L> <u

tfi <U

cj c ^

CJ </>

O rt r^

I'i

o ^~-

'S>2 a

.$P O

B **•* S

S G °

"o C

-fi ^>

"S 2i 'd

03 *•* T3

^ ^3 2^

^3.2 -"2

J'fi

"n? 51

-Q ¥ 1»

2 c§ rt

rt O «*-

-!-• (-1. O

'> "^ Sj

1

<J

I

oyster showing no evidence (
gically) of objectionable conta

II

oyster showing appreciable, ;
light, evidence (bacteriologic
bjectionable contamination

III

oyster showing definite signs ('
Dgically) of pollution, etc., tl
ossibly to be viewed with som
f suspicion

IV (Stringent standard)
oyster showing such obvio
bacteriologically) of pollution
ondemned on a stringent stanc

V (Lenient standard)
oyster showing such unmis
vidence (bacteriologically) of ]
,s to be condemned on the basis <

VI
oyster showing such gross <
bacteriologically) of contamin
o be outside the pale of reco

G ^

G wa O

f_

13s— ' °

C5 « «

^

SHELLFISH 79

Buchan1 examined 25 samples of raw mussels from 15 known
sources coming into Birmingham. The bacteriological examina-
tions were not associated with any investigation into the actual
conditions at the layings nor are such important particulars as
the conditions and the time of storage set out. These omissions
make the results of no value as a basis of comparison with topo-
graphical conditions but they are useful as giving some idea of
the bacterial content of mussels as often put upon the market.
The method used enabled the results to be calculated as per
mussel.

The organisms growing upon gelatine plates averaged about
1000 million per mussel with variations of from 7 million to over
7000 million. On agar at blood heat the average was 182 million
with variations of from 2 million to 2400 million.

As regards B. coli group organisms the results per mussel
were:

I — 10 B. coli organisms in 2 batches,
lo—ioo „ „ „ 7

100— 1000 „ „ „ 10 „

1000—10,000 „ „ „ 3

! o,000— 100,000 „ „ „ 2

over 100,000 „ „ „ I batch.

B. enteritidis sporogenes (milk change).
Absent in I sample.
10 — 100 spores per mussel in 13 samples.

100— 1000 „ „ 10

1000 — 10,000 „ „ I sample.

Streptococci. Not found in five. In the others present,
10 — 100 per mussel in three, 100 — 1000 in five, 1000 — 10,000
in six, and 10,000 — 100,000 in six.

The bacteriology of cockles in relation to sewage pollution.

No definite and extensive comparative work appears to have
been carried out in regard to the bacteriology of cockles in rela-
tion to topographical conditions and degrees of sewage pollution.
From the different analytical results recorded by Klein and others,

1 Report to Birmingham City Council ', 1908,

8O SHELLFISH

it is evident that cockles are frequently extensively polluted
and show sewage organisms in very large numbers. A number
of isolated examinations are recorded by Klein1. With cockles
taken from extensively polluted layings in the river Orwell, the
B. coli in the stomachs of the cockles varied from 4000 — 5000 per
cockle. On the other hand, with a batch of cockles taken from
layings far away from sewage contamination, only one of the
cockles contained a B. cotz-\ike organism and all were free from
true B. coli, streptococci or spores of B. enteritidis sporogenes.

The writer has examined a large number of cockles taken
from various sources. Although some of them were personally
collected from sources well away from any considerable degree
of sewage contamination they all contained B. coli. The num-
bers present were, however, roughly in proportion to the extent
of the sewage contamination known to be present.

The purification and self-purification of shellfish.

The investigations of Klein, Herdman, Herdman and Boyce
have shown that when oysters are artificially inoculated by being
placed in water containing" very numerous typhoid bacilli, these
bacilli can be recovered from the oysters for from 10 to 18 days
subsequently, the actual period varying in the different experi-
ments. Herdman and Boyce2 found that the bacilli did not
increase in the body or tissues of the oyster. When infected
oysters were washed in a stream of clean sea water there was
always a great diminution or total disappearance of the typhoid
bacillus in from one to seven days.

Klein3 in a later series of experiments showed that oysters
infected with very large numbers of typhoid bacilli cleaned
themselves, if placed in clear sea water, within a short period.
Klein's experiments showed that after eleven days the number
of typhoid bacilli were reduced enormously since they could not
be found when one-eighth part of an oyster was examined.
These results are not conclusive in regard to the total disappear-
ance of the typhoid bacilli.

1 Oysters and Shell-fish, Report of the Fishmongers' Company, 1902-9.

2 Thompson-Yates Laboratory Reports, 1898-9, vol. II.

3 Experiments on Vitality of B. typhosus in Shell-fish, 1905.

SHELLFISH 8 1

When cockles were infected with typhoid bacilli, by infecting
the sea water in which they were placed with this bacillus, Klein,
in a single experiment, found that after ten days the cockles
were still heavily infected with typhoid bacilli. In the same
way with mussels the cleansing was only relative, in the few
experiments carried out.

Johnstone1 carried out some experiments on the self-purifica-
tion of mussels. Mussels were used taken from an undoubtedly
polluted source and containing, on an average, 1900 intestinal
bacteria per shellfish. They were then placed in wooden boxes
in sea water which while not unpolluted was reasonably clean.
After four complete days some of the mussels were examined
and the results showed that the intestinal bacteria had been
reduced to about 150 per shellfish (93 per cent, reduction). They
were left for a further three days but no, or very slight, reduction
in the number of bacilli occurred.

Carnworth2 also carried out two experiments with mussels.
The numbers of B. coli were reduced 100— 5<DO-fold after four
days' stay in pure sea water and a longer stay still further reduced
their numbers. A 99 per cent, reduction was met with within a
few days.

Houston3 took 250 oysters *from polluted layings and relaid
them in a locality free from sewage pollution. The experiment
lasted for only 26 days and during this time the oysters showed
no material diminution in the number of B. coli organisms.

Bacteriological examination of sea water and tidal mud.

In connection with shellfish examinations it is always of value
to bacteriologically examine the sea water over the layings and,
in particular, to examine the tidal mud of the estuary in which
they are contained and from their neighbourhood.

The writer4 in 1904 collected and examined a considerable
number of samples of mud from a tidal estuary, at the same time
making a careful topographical examination of the sources of

1 Journ. of Hygiene, 1910, vol. IX, p. 412.

2 Brit. Med. four. 1909, II, p. 695.

3 Royal Commission on Seivage Disposal, ^th Report, 1904, vol. ill.

4 Journ. of Hygiene, 1905, vol. V, p. 146.

S. W. 6

82 SHELLFISH

contamination. In the creek of the estuary was a very large
oyster fattening ground.

The polluted muds showed 10,000 to 100,000 B. coli per
gramme of mud where most polluted, decreasing very regularly
to 10 or less than 10 per gramme in the unpolluted creeks. The
number of streptococci showed similar and gradual diminutions
from 100 to 10,000 per gramme to less than 10 per gramme.

The writer concluded that mud samples yield more reliable
bacteriological evidence of the degree of contamination of a tidal
river than either water or oyster samples. The latter only indi-
cate immediate and actually present pollution while mud samples
show evidence of past contamination. If the muds show a rela-
tively high purity they point to a safe fattening area for oysters.

Data as to the intestinal bacteria in sea water over oysters is
given on pages 75 to 77.

The bacteriological examination of sliellfish.

A. Methods of collection and transmission. Care must be
taken to collect a representative sample and precise local par-
ticulars must be recorded. Shellfish should be collected from
different parts of the area as some parts may be much more
liable to contamination than others. The relationship of the
beds to the tide (if, as is usual, they are in tidal waters) must be
noted, particularly the extent to which the beds are covered or
uncovered by the tide. The neighbourhood and distances from
all sources of contamination, particularly sewer outfalls, must be
carefully recorded.

The samples must be transmitted for examination without
delay and under conditions which do not allow of contamination
in transit.

In the examination of consignments of shellfish collected
and brought into the market it is very desirable to obtain all
possible particulars in regard to the time since collection and the
intermediate conditions of storage. These particulars may not
be necessary for judgment as to the contamination or otherwise
of the particular samples, but they are necessary if the examina-
tion is to be used in any way as a guide to the condition of the
beds themselves.

SHELLFISH 83

B. Methods of sampling and dilution. These differ slightly
according to the shellfish.

Oysters. The most thorough method for the bacteriological
examination of oysters is that used by Houston for his work
for the Royal Sewage Disposal Commission. The procedure is
briefly as follows: .

1. The outside of the oyster-shells are well scrubbed with
soap and water, and cleaned as thoroughly as possible in running
tap-water, finally with sterile water.

2. The hands of the investigator are thoroughly cleaned,
washed in I in 1000 corrosive sublimate solution, and finally with
sterile water.

3. The oysters are opened by a sterile knife held in position
by a sterile cloth, and with the concave shell underneath. Great
care must be taken to avoid any loss of the liquor. The liquor in
the shell is poured into a sterile loooc.c. cylinder, and the oyster
and oyster liquor are added after the oyster has been cut into
small pieces by sterile scissors.

4. Ten oysters are to be treated as above in each experi-
ment.

5. The volume of oyster + oyster liquor is read off, and
usually varies between 80 and 1 20 c.c. For qualitative work
100 c.c. may therefore be taken as a fair average of the total
shell contents of ten oysters.

Sterile water is then poured into the cylinder up to the
loooc.c. mark, and the whole well stirred with a sterile rod.
Each 100 c.c. of this liquor may be considered to contain the
bacteria in one oyster.

10. Various amounts and fractions of this liquor are used
for the examination for B. coli, B. enteritidis sporogenes, and for
streptococci.

Mussels. The most satisfactory method is one similar to the
above, the whole mussel being used. The mussels are thoroughly
cleaned, also the hands of the investigator, as for oyster examina-
tions, while the procedure is quite similar after the shellfish are
opened.

6-2

84 SHELLFISH

Buchan found that a mussel of average size has a bulk of
about 15 c.c. but considerable variations occur with individual
mussels.

When the only opinion required is as to the relative purity of
mussels from different sources, and not a quantitative enumera-
tion, the following method of Johnstone (loc. cit.) may be found
useful.

The mussels are washed under the tap and opened so
that the adductor muscles of the shell, the pedal muscles, and
the muscles of the mantle, are alone cut through. The soft
parts of the mollusc are retained in the right-hand valve of the
shell. Sterilized knives are prepared, two for each shellfish. A
slit is made in the body of the animal immediately over the
stomach and through the dark green " digestive gland " — really
an extension of the lumen of the stomach — and a small quantity
of the stomach juices is withdrawn by a sterile pipette with
rubber teat (previously made in batches and drawn out to deliver
approximately the same volume of liquid) and distributed over
the surface of an L.B.A. plate. The volume of fluid taken
amounts to about O'l c.c. and usually a mussel contains enough
to make two or three separate inoculations. In this way a com-
parative estimate of the number of B. coli group organisms is
obtained.

Cockles. These small shellfish should be treated in the same
way as oysters. A sample of cockles consisting of ten animals
should be made up to IOOC.G. One c.c. of the cockle mixture
then contains O'l animal.

C. Bacteriological procedures. The number of organisms
developing upon gelatine and agar plates can be obtained in
the ordinary way but the enumeration is of no value and is
scarcely worth doing.

The determinations of value are those for the ordinary ex-
cretal indicators, B. coli group organisms, spores of B. enteritidis
sporogenes and, to a less extent, streptococci.

For B. coli Houston used cxoooi, O'ooi, 0*01, 0*1, I, 10 and
100 c.c. the amounts being added to bile salt broth, while for
B. enteritidis sporogenes the two extremes were omitted.

SHELLFISH 85

The actual isolation is carried out in the ordinary way. As
mentioned above for rough comparative purposes Johnstone used
direct plating of definite amount of shellfish juices upon L.B.A.
media.

CHAPTER VI

MILK

The bacteriological examination of milk even up to a few
years ago could not be said to be very satisfactory, and indeed,
apart from tuberculosis, very little was systematically done, while
in regard to the procedures that were carried out there was
neither uniformity of method nor consensus of opinion as to
what was desirable. Even at the present time it is by no
means on a uniform or satisfactory basis.

The bacteriological examination of milk may be utilised to
give information in the three following directions :

1. To measure the degree of contamination of the milk
from faecal and other sources and its general bacterial content.

2. To ascertain the presence or absence of definite disease-
producing organisms, e.g. B. diphtheriae, B. tuberculosis, B. typho-
sus.

3. To obtain evidence as to the healthiness of the milk-
producing apparatus of the cows which supply the milk.

The procedures to be used will obviously require to be varied
according to the purposes of the examination. The different
methods are considered in detail in the following sections.

Collection of Samples.

Care in collection is necessary to obtain samples which are
really representative of the milk to be examined. Ordinary
samples of mixed milk may be collected at the byres or in
course of transit or delivery. The cream and the sediment are

86

MILK

the richest in bacteria, so that it is necessary to well mix the
milk before a sample is withdrawn.

In the collection of samples from individual cows care must
be taken to avoid outside contamination. The cows' udders and
teats must be washed carefully, and the milker must wash and
disinfect his hands. The milk is milked direct into the bottle,
the stopper being held by its free end by a second person to
avoid contamination, and inserted immediately after the sample

Fig. 8. Collecting bottle and ice box.

has been collected. The bottle should not have too narrow a
mouth. In some cases it may be sufficient to collect a mixed
sample from the four quarters, and then care should be taken to
obtain as nearly as possible equal quantities of milk from each
teat. In other cases it may be necessary to collect a separate
sample from each quarter. In general the fore milk should be
rejected and middle milk sampled. In rare instances it may be
necessary to collect fore, middle, and end milk samples from one
special quarter or from each of them.

The milk should be collected in sterile bottles with accurately

MILK 87

fitting glass stoppers. For the ordinary examination of mixed
milk a sample of about a pint is a convenient amount to obtain,
but for most purposes a much smaller quantity is sufficient —
particularly for samples from individual cows.

To collect the sample the simple and efficient apparatus
described by Delepine (Fig. 8) may be used. It consists of a
metal case containing a 7 or 8 oz. bottle and a milk-scoop.
All the parts are thoroughly sterilized in the laboratory before
being sent out, and the sterilized case is opened only at the time
when the sample is taken. The sterilized scoop is used to
remove the milk from the cans or other vessels.

If the samples have to be transmitted any distance, or cannot
be examined within an hour or two, they must be packed in ice.

The ice-box used by Delepine (Fig. 8), which is similar to
the apparatus the writer uses for bacteriological samples of water,
is very convenient. The size the writer uses is made to just
hold four 2 oz. bottles and their tins, but larger bottles are more
convenient for milk samples. The glass-stoppered bottles are
sterilized in their tins, into which they just fit, being prevented
from moving by layers of asbestos cardboard above and below.

Some alteration in the bacterial content takes place at o° C.
or a few degrees above this temperature but it is not marked for
several days and there is no material alteration or error of judg-
ment in examining samples kept iced for not more than 12 to
20 hours.

Particulars to record with the Samples.

Certain particulars must be carefully recorded, and should
include :

(a) Date and time of sampling.

(b) Identification details as to farm or person from whom
the milk is obtained.

(c) If mixed milk or from individual cows.

(d) If mixed milk, whether collected from byre, in transit, in
shop, etc., or as delivered.

83

MILK

MILK 89

(e) If collected at the byre, the number of cows from which
obtained should be given, and whether the milk has been strained
or not, and if so the strainer used.

(/) If from individual cows, particulars of the quarter or
quarters from which obtained, and if fore, middle, or end milk.

It is useful to take and record the temperature of the milk.

Dilution of the Sample.

In the bacteriological examination of milk the proper dilution
of the sample is of the utmost importance if accuracy is to be
obtained. It is important that the milk should be thoroughly
well shaken before dilution, and that the dilutions should be
well mixed.

The following is the most convenient method of dilution.
A large number of glass-stoppered bottles of about 120 c.c.
capacity are used, each containing 90 c.c. of sterile tap-water
(Fig. 9). Sterile one-mark loc.c. pipettes are conveniently used
to add the milk, and these should be made short for convenience
of sterilization. After thorough shaking 10 c.c. of the milk is
removed and added to a 90 c.c. dilution bottle (Dilution A).
After well mixing 10 c.c. Dilution A is added to a second
90 c.c. bottle (Dilution B). In the same way Dilution C is
made from B and Dilution D from C. Each dilution, of course,
represents one-tenth dilution of the one immediately above it in
series.

Small i c.c. pipettes graduated in tenths of a c.c. are used to
add fractions of the different dilutions to the requisite media.

Dilution flasks or bottles containing 99 c.c. or 9 c.c. are often
used, and are recommended by the American Committee on
Standard Methods of Bacterial Milk Analysis1, but the writer is
of opinion that the addition of only I c.c. of milk for the primary
dilutions is unreliable and leads to error. By adding as much as
10 c.c. errors of measurement are reduced to a minimum, while
the use of glass-stoppered bottles enables the dilutions to be very
thoroughly mixed.

1 American Jonrn. of Public Hygiene, 1910, XX, p. 315.

90 MILK

Sources of bacteria in milk1.

Milk as secreted is a sterile fluid, but during every stage from
the udder to the consumer, contamination is possible, and under
many of the present-day conditions is invited. In consequence,,
milk as vended may, and frequently does, contain very many
thousands of bacteria. The sources by which bacteria gain access
to milk can be grouped as follows :

1. Intra-mammary.

2. Introduced during the milking operation,

3. From ordinary milk utensils.

4. From the use of special milk apparatus.

5. Contamination in transit.

6. Contamination upon the purveyors' or consumers' premises,

1. Intra-mammary Contamination. The view generally ac-
cepted up to a few years ago was that the interior of the udder
contained no bacteria, and that milk, if not quite sterile when it
left the teats, yet contained but a few bacteria derived from the
teat ducts themselves. This opinion was largely based upon the
fact that with a sterile tube it was sometimes possible to obtain
sterile milk from the udder. More recent investigations have,
however, shown that this is not in accordance with the facts
and that undoubtedly the teat duct, the milk cistern and the
larger ducts usually contain bacteria whose primary origin is out-
side the teat. Some selective action has however taken place,
the organisms found being almost invariably staphylococci,
streptococci and other forms of micrococci.

2. Introduced during the Milking Operation. The bacteria
introduced during milking are derived from three sources, (a) the
coat, udder and teats of the cows ; (b) dust from the milking
shed and the clothes of the milker, and (c) derived from the
hands of the milker.

1 For a more detailed account of the bacteria in milk see Savage, Milk and the
Public Health, Macmillan and Co. 1912.

MILK 91

All these sources add bacteria to the milk, the first probably
being responsible for the greatest contamination.

3. Bacteria introduced from Milk Vessels. When imperfectly
cleaned these are a source of great bacterial contamination to the
milk.

4. From Special Milk Apparatus. The commonest ap-
paratus used, not strictly appertaining to milking, is a milk-
strainer. It is an almost universal practice of the cow-keeper to
strain the milk, usually immediately after milking. If not kept
absolutely clean milk strainers may actually add bacteria to the
milk.

Milk-coolers are sometimes a source of bacteria in milk.
They may be placed in dusty places so that the milk passing
over them, exposing as it does a large surface to the air, takes
up a good many bacteria. Also they may not be kept perfectly
clean, traces of milk being left in which bacteria multiply enor-
mously and are subsequently added to the milk. A third but
much less common way by which coolers may contaminate the
milk, is by being leaky, the water being added in small quantity
to the milk.

In the same way milk-separators usually add bacteria to the
milk passing through them as estimated by numerical counts.
This increase is however largely apparent and due to the breaking
up of clumps of bacteria.

5. Contamination in Transit. It is an extremely difficult
thing to measure the amount of bacterial contamination of milk
which takes place in its transit from cowshed to milk purveyor
or consumer. The increase in the number of bacteria added
from outside sources is obscured by the increase due to the
multiplication of the bacteria already in the milk.

The addition of bacteria from outside in transit is probably
trivial compared with the bacterial additions at the cowshed and
in the consumer's house.

6. Bacteria added on the Purveyor s or Consumer's Premises.
Obviously the number of bacteria added must vary enormously
with the condition of the premises and the cleanliness practised
upon the premises.

92 MILK

Procedures to determine the Manurial and General

Bacterial Contamination of Milk.

The above considerations show that the sources whereby
bacteria gain access to milk are very numerous and diverse
while they are operative very unequally, according as cleanliness
conditions are or are not practised.

To gauge this general bacterial contamination, the following
estimations have to be considered.

A. Estimation of the Number of Bacteria.

The estimation of the total number of bacteria in milk is
impossible, and it is necessary to select an arbitrary basis for
enumeration. There are no known nutrient media, and no
known conditions of growth which will allow all the bacteria in
milk to develop. All we can say is that some media and some
conditions are more favourable to the growth of a larger number
of the bacteria in milk than are others, and by their employment
a higher count is obtained than by the use of less favourable
media and conditions. The bacterial count being relative, not
absolute, any enumeration method for the bacteria in milk should
aim not so much at obtaining the largest count, but one which
shows best the evidences of manurial pollution. Some investi-
gators use nutrient gelatine, others nutrient agar, others whey
agar, etc. In the writer's opinion nutrient agar (of + 1 reaction),
with the plates incubated for forty-two to forty-six hours at 37° C,
is the most convenient and satisfactory, but gelatine plates grown
for three days at 20° to 22° C. are used by many.

With unknown samples an extended series of dilutions will
have to be employed for plating in order not to have overcrowded
plates. A rough idea of the number of bacteria in a given milk
sample, and so a guide obtained as to the dilutions to plate,
can be ascertained from an examination of the stained centri-
fugalized deposit.

In milk examinations involving bacterial estimations, the
presence or absence of preservatives should be ascertained, as if
present they will have an inhibitory or retarding action on the
number of bacteria, and a stale milk so treated may show a low
bacterial figure.

MILK 93

B. Estimation of the Number of Bacillus colt and allied

Organisms.

The general principles and most suitable methods have been
discussed in Chapters I and II so that only any special modifica-
tions for milk samples require to be considered.

The methods to use differ in no way from those employed
for water samples except in regard to the dilutions to add to the
tubes of lactose bile salt broth.

Definite fractions of milk are added by pipette, the amounts
depending upon the suspected degree of pollution of the milk.
A usual procedure is to add ro, 0*1, croi c.c. if the sample is
byre milk, and if it is ordinary vended milk to add, in addition,
O'OOi,O'OOOi,O'OOOOi, O'oooooi c.c. and occasionally even smaller
fractions. The dilutions are obtained as described above.

The B. coli group organism is isolated from the smallest
amount giving gas and acid after incubation at 37° C. for
two days.

These dilutions are widely spaced, and consequently the
number of B. coli present can only be enumerated between
rather wide limits. The writer prefers therefore to add several
equal amounts to lactose bile salt broth tubes. Thus for samples
collected at the byre he adds I c.c. to each of five broth tubes
and O'l c.c. to two others, while for ordinary vended milk samples
greater dilutions are used, i.e. O'oooi c.c. four, O'OOi c.c. four,
O'Oi one, and O'l one.

By this procedure a much closer estimate of the number of
B. coli can be obtained than by the ordinary methods and with
but very little more expenditure of time and material.

C. Estimation of the Niimber of B. enteritidis sporogenes

Spores.

The milk itself is directly incubated. The usual quantities
of milk to examine are i, 10, and 20 c.c., the smallest amount
being added to a tube of freshly sterilized whole milk, while
the other quantities are placed in empty sterile test tubes. The
milk tubes are heated in a water-bath to 80° C., and kept at that

94

MILK

temperature for ten minutes, then cooled and incubated ana-
erobically as described in Chapter II.

The amounts given above are too wide apart to yield a
satisfactory estimation, and the following method is advocated

by the writer1 : quite small, nar-
row (4 by J inch), sterile empty
test tubes are used in batches of
ten for each estimation, 20 c.c.
of milk is employed for each
test, 2 c.c. being added by sterile
pipette to each tube. The ten
tubes are heated for 10 minutes
at 80° C, rapidly cooled and in-
cubated anaerobically in speci-
men jars with ground -glass
stoppers, just large enough to
take the ten tubes (Fig. 10),
the oxygen being absorbed by
the usual potash and pyrogallic
acid mixture. The tubes are
examined after two days' incuba-
tion at 37° C. for the characteristic
changes described in Chapter I.
In the slender test tubes advo-
cated the 2 c.c. of milk half fills the tube, and the condition of
the milk can readily be observed. The tubes being small
readily go into a small specimen jar which consequently requires
less chemicals to absorb all the oxygen, while space in the incu-
bator is economised. The essential value of the modification
is that the test is made much more delicate without additional
work.

Some arbitrary standard is convenient for recording. Each
positive result in a tube is counted as one B. enteritidis spore, an
assumption which is probably, but certainly not always, true.
Thus, if all the ten tubes show a positive " enteritidis change,"
the result is recorded as 10. All gradations between o and 10

Fig. 10.

1 Local Govt. Board Medical Officer's Report) 1909-10, p. 477.

MILK 95

may be met with, and a comparative sensitive test is in this way
available.

For milk work it is not usual to confirm the diagnosis by a
pathogenicity test. In certain cases this may be necessary and
may be carried out as described in Chapter I.

One great value of this test is that it is a non-multiplying
one, so that it is especially useful for vended milk. A draw-
back to its utility is the fact that the spores are not always uni-
formly distributed in the milk.

D. Estimation of the Sediment or Dirt in Milk.

Numerous investigators have attempted to judge the cleanli-
ness of milk samples from a determination of the amount of dirt
or sediment obtainable from them. The methods used are either
by slow sedimentation or by centrifugalisation.

Of slow sedimentation methods, Orr's modification of Hous-
ton's method is perhaps the best. The apparatus used is shown
in Fig. II. The top of the centrifugal tube fits outside the
lower part of the large glass cylinder, the two being connected
by rubber tubing. The centrifuge tube is graduated in tenths
and hundredths of a c.c. A litre of milk (after addition of I c.c.
formalin) is allowed to stand for twelve hours in the large
cylinder. At the end of twelve hours, a brass rod fitted with
a rubber stopper at the lower end is passed through the milk,
and fitting into the outlet of the cylinder prevents escape of the
fluid when the centrifuge tube is detached. The tube is centri-
fugalised (1500 to 2000 revolutions per minute) for three to
five minutes, and the milk poured off. Sodium carbonate solution
(i per cent.) is then added up to the 10 c.c. mark and the centri-
fugalising repeated. The deposit is then read off.

A litre of milk is a large amount to use for a single test and
for routine work a rough but sufficiently reliable estimation of
the volume of the sediment may be obtained by direct centri-
fugalisation. Tubes, which are of narrow calibre at the lower
closed ends (similar in shape to those used by Orr, Fig. n), are
used, preferably with a capacity of 50 c.c. The milk is filled in
to a definite volume (50 c.c.), centrifugalised for a definite period

MILK

at a known rate, and the volume of sediment directly read off in
the graduated narrow end. The results are doubled and re-
turned as volume of sediment per 100 c.c. of milk.

COM

15 C.

*/8 Natural Size

2/3 Natural Size
Fig. ir.

= Glass tube holding a litre of milk. C=Tube for measuring sediment.

2? = Rubber tubing connecting A and C.

At the side is represented a brass rod, with rubber stopper attached, for
plugging the lower end of tube A, when C is taken off.

Tube C is shown by the side, § natural size. It is graduated in
and ths of a c.c.

E. Determination of Acidity.

The acidity of milk is estimated by titrating 20, 50 c.c., or
other definite quantity, of the milk with ^ sodium hydrate,
phenolphthalein being used as the indicator. The titration

MILK 97

should be carried out in a narrow beaker with a control tube
by the side. The result is conveniently expressed in terms of
degrees of acidity, each c.c. of -^ alkali required for 100 c.c. of
milk being I degree of acidity. Thus, if 50 c.c. of milk are taken
and 9 c.c. of alkali are required, then 100 c.c. would require 18 c.c.,
and the milk has 18 degrees of acidity. Each c.c. of -^ alkali is
equivalent to O'OO9 gramme of lactic acid.

Comparison of methods.

The above five procedures are made use of by different
workers to estimate and measure the general bacterial contami-
nation of milk and their relative value for this purpose must be
considered.

Since prejudicial milk contamination is essentially a bac-
teriological question it might be supposed that the estimation of
the number of bacteria in milk would be a very reliable index of
the degree of pollution to which it had been subjected and a
measure of the undesirable condition to which it had attained.
There are, however, many objections to this estimation for this
purpose, the following being the most important.

(a) The number of bacteria enumerated varies greatly with
the exact laboratory procedures adopted. Unless these are
rigidly defined, widely different results will be obtained with the
same milk sample. The kinds of bacteria in milk are so varied
and so sensitive to slight variations of environment that quite
considerable differences are readily produced.

(b} This estimation makes no distinction between bacterial
contamination before and after the milk leaves the cow. Since,
as shown above, bacteria are added from the teats and milk cistern
ordinary bacterial counts cannot be taken as measuring purely
external contamination.

(c) While under rigidly defined conditions bacterial counts
are useful for samples collected at the byre this estimation is
useless to gauge the extent of outside pollution for ordinary
samples of market milk. The initial number of bacteria may be
masked by the great multiplication which has taken place sub-
sequently to milking. It is not in any way possible to gauge

s. w. 7

98 MILK

the amount of external pollution which has been added to milk
from an estimation of the number of bacteria in chance samples
of vended milk.

The determination of the number of B.coli and allied organisms

o

is a much more valuable estimation. These organisms as shown
in Chapter I are extremely abundant in cowdung, and although
there are other sources of milk contamination undoubtedly
manurial pollution is by far the most prevalent. These bacilli
are absent from milk samples collected directly from the teats
or under conditions of great cleanliness. Further they can be
easily isolated and numerically estimated by methods which are
likely to give the same results with different workers.

As regards samples collected at the byre there is no doubt
that this estimation is a direct and extremely valuable method
of measuring the degree and extent of manurial contamination.

The value of this enumeration for chance samples of ordinary
vended milk involves matters of great complexity, and it is a
problem of great difficulty to correctly gauge the significance of
these bacilli. If it were a practicable and universally adopted
procedure to thoroughly cool milk after collection, and then to
maintain it until sold at a temperature so low that B. coli will
not multiply in it, any standards framed for B. coli in byre milk
would be equally applicable to milk as sold. Under present
conditions this is by no means the case, and it becomes a very
complicated problem to say what number of B. coli may be
allowed in vended milk samples as sold under present-day con-
ditions.

In the writer's opinion the most reliable deductions can be
obtained from a study of the rate of multiplication of B. coli
in milk under different conditions of temperature, etc. From
his investigations in this direction1 he is of opinion that while
samples at the byre should not contain more than one B. coli
(including allied organisms) per c.c., samples of milk as vended
should not contain more than 100 to 500 in winter and 1000 to
2500 in summer as extreme permissible limits.

The estimation of the number of spores of B. enteritidis
sporogenes is also of considerable value more particularly because
1 Ideal Government Board ', Medical Officer's Report 1909-10, p. 474.

MILK 99

this organism does not multiply in milk and so is of equal value
for byre and vended milk samples. The spores are present in
considerable numbers in cowdung. If the method of enumera-
tion described in this Chapter is adopted, o or I (i.e. all ten tubes
negative or only I positive) might be considered good, 2, 3, or 4
positive as unsatisfactory, and 5 or over positive as bad.

The estimation of the sediment or dirt in milk is considered
by many workers to be a very valuable guide to the bacterial
contamination of milk. Since the bacteria which gain access to
milk are contained in particulate matter, generally in manure,
there is some justification for estimating the deposit in milk and
taking it as an index of the number of undesirable bacteria
\vhich have gained access to it. The test is also a non-multiply-
ing one, so is as valuable for vended as for byre milk samples.
The amount of sediment can also be fairly easily and quickly
estimated.

The objections on the other hand are considerable and
important and are briefly that there is no definite relationship
between the amount of sediment and the number of bacteria,
that the sediment is only partially manurial, part being harmless
inorganic substances or substances natural to milk, and that
the amount is directly proportional to the efficiency of straining
rather than to the cleanliness precautions adopted.

The last is the most important objection since straining
while it removes the larger dung-particles does not remove the
bacteria and makes the test of but little value.

Acidity determinations are of no real value to estimate
bacterial contamination since there is no relationship between
the number of bacteria or B. coli and the degree of acidity.

The detection of pathogenic bacteria in milk.
Apart from certain types of streptococci, the only pathogenic
organisms for which it is likely that milk will be required to be
examined are the bacilli of tuberculosis, diphtheria, and typhoid
fever.

Examination for B. tuberculosis. Tubercle bacilli, when
present in milk, are frequently only in small numbers. It is
therefore of importance to examine a considerable amount of

7-2

TOO MILK

the milk, and to thoroughly centrifugalise it in a powerful centri-
fuge. The sediment is examined microscopically for the tubercle
bacillus, and the rest is inoculated into guinea-pigs subcutaneously,
as described below. It is advisable to inoculate several guinea-
pigs from each sediment, as some of the animals may die from
the action of the other bacteria present in the milk1. Professor
Delepine has shown that this accidental mortality is much
diminished if only quite fresh milk is used, or, if delay is un-
avoidable, when the milk samples are ice-packed during transit.
This mortality can also be diminished by treating the sediment
with antiformin before inoculation.

Different authorities have advocated examining different
quantities of milk. The greater the amount examined the
greater the chance of finding tubercle bacilli, and the more
reliable a negative result. Delepine2, as a routine method, uses
two tubes of 40 c.c., and centrifugalises for a quarter of an hour
in a machine giving 3000 revolutions per minute. The cream
and all but 2 c.c. of the separated milk are then aspirated away,
and the separate residue from each tube, containing the sediment
and 2. c.c. of separated milk, after a microscopic examination of
the milk has been made, is inoculated into a guinea-pig.

Other authorities use much larger quantities of milk — e.g.
•J- litre — for each examination. Tubercle bacilli may also be
contained in the cream (being carried up with the fat), and to
complete the examination some of the cream may also be inocu-
lated, or, more simply, after a preliminary centrifugalisation
for half an hour, the cream may be thoroughly broken up by
a sterile glass rod, and the milk again centrifugalised. In this
way a large proportion of the organisms in the cream is
transferred to the sediment. In any case the sediment is

1 O'Brien (Journ. of Meat and Milk Hygiene, 1911, vol. I, p. 295) found that out
of over 9000 guinea-pigs subcutaneously inoculated with milk sediment 33 per cent,
died, from causes other than tuberculosis, before the 28th day. Careful investigation
in the cause of this non-tubercular mortality showed that while part was due to
infectious diseases which the animals were incubating at the time of inoculation and
part to bacilli inoculated with the milk, the greater part must be ascribed to a
combination of factors, including the disturbance caused to the animal's economy by
the proteins injected and the consequent lowering of resistance to auto-infeclion.

2 See Transactions of the British Congress on Tuberculosis, vol. II, pp. 286-7.

MILK 101

examined microscopically, and the remainder 'inoculated 'into
several guinea-pigs.

To examine the sediment microscopically for tubercle bacilli,
make cover-slip preparations, and either stain directly by the
Ziehl-Neelsen method in the ordinary way, or, preferably, stain
after a preliminary treatment with ether for several hours, to
get rid of the fat.

Little or no reliance is to be placed on the microscopic
examination of the sediment from mixed milk samples for the
detection of the tubercle bacillus. The microscopic examination
of the milk from individual cows is of more value since the
chances of finding tubercle bacilli are greater, but a negative
result cannot be accepted as reliable evidence that tubercle
bacilli are absent.

For inoculation purposes the method of subcutaneous injection
as advocated by Delepine1 gives very reliable results. In this
method the injection is made subcutaneously, with strict aseptic
precautions, on the inner side of the leg at the level of the
femoro-tibial articulation. If the inoculated material contains
tubercle bacilli, an infection of the glands on the inoculated side
takes place. The popliteal, superficial, and deep inguinal, and
usually the sub-lumbar glands on the inoculated side are enlarged
(see Fig. 12), and tubercle bacilli can be demonstrated in films
made from them. The time at which this can be done varies
with the number of tubercle bacilli injected. According to
Delepine, with milk containing numerous tubercle bacilli definite
evidence of tuberculosis can be seen in animals killed after ten
to fifteen days ; when not sufficiently numerous to be detected
microscopically in the milk (i.e. a moderate number of bacilli)
at the end of fifteen days, and when very few bacilli are present
in the milk, it may be difficult to obtain any clear evidence of
infection before the end of the fourth or fifth week. Delepine lays
stress upon the importance of keeping the inoculated animals
isolated and under favourable hygienic conditions. For ordinary
work a good plan is to inoculate two animals, one being killed at
the end of three weeks, and the other a week or two later.

While the pathological picture is practically conclusive of

1 British Medical Journal, 1893, vol. II, p. 664.

102

MILK

Fig. 12.

Experimental tuberculosis in a guinea-pig (about the third week after inoculation)
inoculated subcutaneously in neighbourhood of the left knee-joint. (From Curtis's

Bacteriology.}
The enlar

larged popliteal, superficial and deep inguinal and sub-lumbar glands on the
side of inoculation are very obvious ; also the retro-hepatic gland and diseased spleen.

MILK 103

tuberculosis, complete proof should always be obtained by
finding tubercle bacilli in the enlarged glands and other tuber-
culous lesions.

The animal test for tuberculosis takes at least three weeks
before any diagnosis can be made. To shorten this period
Bloch1 has suggested that the inguinal glands on the inoculated
side should be slightly damaged by squeezing them. He found
that in all positive cases the glands within 9 to 12 days were
markedly enlarged and tubercle bacilli present in large numbers
both in films and sections. The earlier development of tuber-
culosis is due to the greater growth in the slightly damaged
glands. This modified procedure has been found of value by
several other investigators.

The possible presence in milk of acid-fast bacilli other than
the tubercle bacillus, while it diminishes the value of simple
microscopic examination, does not to any considerable extent
interfere with the inoculation test. The death of the guinea-pig,
with lesions apparently those of tuberculosis, is almost certainly
due to tubercle bacilli. As a routine procedure, and certainly in
any cases of doubt, cultures should be made on glycerine agar
from the enlarged glands. The simulating acid-fast bacilli grow
readily and rapidly upon this and other nutrient media, unlike
the tubercle bacillus.

Since the acid-fast bacilli are of considerable importance in
relation to the diagnosis of tubercle bacilli in milk and milk
products a brief note of this group may be of assistance.

The acid-fast bacilli are a group of organisms fairly widely
scattered in nature and mostly of no particular significance in
themselves but which owe their importance to the fact that
morphologically they somewhat closely resemble the tubercle
bacillus while also, like that bacillus, they resist decolorisation
by acids.

Of recent years a number of such bacilli have been isolated
and studied of which the best known are the butter bacillus of
Rabinowitsch and Petri (see page 117), Moeller's timothy-grass
bacilli I and II, originally isolated saprophytically growing upon
this grass, Johne's bacillus the cause of Johne's disease (a form

1 Berlin, klin. Wochenschr. 1907, Vol. XL, p. 511.

104 MILK

of chronic enteritis of cattle), the smegma bacilli found in smegma
and the mist bazillns isolated from manure. In general all these
bacilli morphologically resemble the tubercle bacillus but are for
the most part shorter and thicker and stain more uniformly.
They resist decolorisation by acids unless submitted to pro-
longed treatment while their resistance to decolorisation by
alcohol is somewhat inferior to that of the tubercle bacillus.
Injected into animals most of them, like the butter bacillus,
exert pathological effects with the production of nodules re-
sembling tubercles. The characteristic which essentially differ-
entiates them from the tubercle bacillus is their comparatively
rapid growth upon and in culture media. Their growth in
days exceeds that of the tubercle bacillus in the same number
of weeks. They will also grow at room temperatures. A further
point of differentiation is that they do not produce tuberculin.

Examination for B. diphtheriae. This organism has been
found in milk, but only on a very few occasions.

The method usually employed is to take advantage of the
rapid growth of this bacillus on blood-serum.

The milk is centrifugalised in sterile tubes, and cultivations
are made from both the cream and the sediment.

Dilution is obtained by acting upon the same principle as
that used in brushing agar and gelatine plates. A little sediment
or cream is taken up by a sterile platinum loop, and is rubbed in
close vertical lines over the surface of three blood-serum tubes
without recharging the loop. A large number of blood-serum
tubes should be inoculated from both sediment and cream,
incubated at 37° C., and examined after twenty to twenty-four
hours. All possible B. diphtheriae colonies must be carefully
examined (to diminish the work, surface sweepings may be
taken) in cover-slip preparation. If bacilli morphologically
resembling this organism are found, they must be subcultivated,
obtained in pure culture, and the virulence determined.

Many observers have recorded the presence in milk of bacilli
which morphologically resemble the diphtheria bacillus, but
which from their other characters are certainly not that organism.
Their occurrence emphasizes the absolute necessity of carrying

MILK 105

out cultural and animal inoculation tests before a diagnosis of
diphtheria bacilli in milk is made.

Examination for B. typhosus. The detection of this organism
in milk is a matter of considerable difficulty. The milk sediment
after centrifugalisation is mixed with a little sterile water, and
fractions of the emulsion are distributed over the media in a
series of Petri-dishes. The media to use are one or other of
those which differentiate the typhoid bacilli to some extent, such
as lactose bile salt neutral red agar. All suspicious colonies
are investigated. The composition of the media and the method
of investigating the colonies are the same as for the isolation of
this bacillus from water.

Procedures to obtain evidence as to the condition of the milk-
producing apparatus of the cows from which the milk is
obtained.

The most important of these conditions is tuberculosis of the
udder. The detection of tubercle bacilli has been described,
and it is only necessary to remark that finding these bacilli in
milk, while generally pointing to tuberculosis of the udder of
one or more of the cows supplying the milk, does not invariably
bear that interpretation. Schroeder, and more recently the
English Royal Commission on Tuberculosis, have shown that
tubercle bacilli may be found in the milk of tuberculous cows
which show no evidence of udder tuberculosis. In many cases
they are derived from the faeces, which contain the bacilli and
are allowed to pollute the milk.

A. Estimation of the cellular content of milk. All milk
samples contain a certain number of cellular elements, and in
certain pathological conditions their number is enormously in-
creased. Deductions of great value can be made from accurate
determinations of their number in the milk of individual cows.
For mixed milk samples this determination is of much less utility.

The older method used was to centrifugalise the milk, spread
the sediment as evenly as possible over a cover-slip, dry, and
stain by methylene blue. The number of leucocytes in a number
of fields of vision was then counted, and an attempt made to

io6

MILK

estimate by returning the number of leucocytes as so many per
microscopic field. This method is very inaccurate, since the
number will vary with the size of the microscopic field, and
especially with the thickness of the film, the latter being im-
possible to accurately control. The writer worked out the
following method, by which the cellular elements can be readily
and accurately enumerated in milk.

Savage's Method1. The ordinary Thoma-Zefss blood-counting
chamber is employed. Direct counting of the cells is impossible
owing to the opacity caused by the large amount of
fat. One c.c. of the milk is accurately transferred
to a centrifugal tube (about 15 c.c. capacity) of
the pattern shown in Fig. 13, and freshly filtered
Toisson's solution2 is poured in to almost fill the
tube. The two fluids are well mixed and then
centrifugalised for 10 minutes. The cream is well
broken up by a clean glass rod, to disentangle
leucocytes carried to the surface, and the mixture
centrifugalised for an additional 5 minutes. All
the fluid is then removed down to the I c.c. mark,
great care being taken not to disturb the deposit.
This can be conveniently and readily done by
means of a fine glass tube connected to an exhaust
pump. Theoretically, all the cellular elements
present in the original I c.c. of milk are now present
in the I c.c. of fluid. The deposit is thoroughly
well mixed (with a wire), and distributed through
the I c.c. A sufficient quantity is placed on the
ruled squares of the Thoma-Zeiss apparatus, and
the cover-glass put on. The number of cells is counted in a
number of different fields of vision, moving regularly from one
field of vision to another. The diameter of the field of vision is
ascertained before counting by drawing out the microscope tube

1 Journal of Hygiene, 1906, Vol. VI, p. 123.

2 This is the well-known indifferent solution used in blood enumerations. It does
not injure the cells, but stains them sufficiently to render them clearly visible. Its
composition is methyl violet 0*025 grm-» sodium chloride i grin., sodium sulphate
8 grms., glycerine 30 c.c., distilled water 160 c.c.

\

1.C.C.

Fig-

MILK ID/

until an exact number of sides of the squares spans a diameter
of the field of vision.

The number of cellular elements per cubic mm. of

milk=— , where j/ = the average number per field of vision,

^=the number of squares which just spans the diameter, d is
determined once for all by marking the microscope draw tube so
that only 20 fields have to be counted, and the figures substituted
in the formula.

B. Examination of the stained centrifngalised deposit. To
obtain comparable results the sediment from a definite amount
of milk should be examined after centrifugalisation for a definite
period. Ten c.c. of milk centrifugalised for 10 minutes is con-
venient. Part of the deposit is spread thinly but uniformly
over a cover-slip, dried in air, fixed in the flame, or preferably by
soaking in a mixture of equal parts alcohol and ether for one
minute, stained by methylene blue and mounted in balsam.

The preparation may be utilised to gain an idea of the
general bacterial content, whether streptococci are present, and if
so in what numbers and whether intracellular, while, if considered
necessary, a differential count may be made of the cellular
elements present. For this purpose not less than 200 should
be enumerated. With care a rough but valuable estimate can
be obtained from this examination as to the probable number
of bacteria in the sample.

C. Determination of the number of streptococci. To estimate
the number of streptococci in milk the method recommended as
the simplest and most reliable is to add diluted fractions of the
milk, ro, O'l, O'OI, O'OOI c.c. etc., to tubes of glucose neutral
red broth. Ordinary broth will do, but the neutral red broth
is preferable and gives better results. The tubes are incubated
for two days at 37° C. and then examined, in hanging drop
preparation, for streptococcus chains. The deposit should be
selected for examination, and several hanging drop preparations
made. A positive result should only be recorded when quite
definite chains of cocci are detected, or, in doubtful cases, when
stained preparations show such definite chains.

IOS MILK

To isolate the streptococci, brush diluted loopfuls of the
positive tubes over plates containing nutrient agar. Incubate
for 24 hours, and if necessary for 2 days, at 37° C. Subcultivate
the colonies with the characters of streptococcus colonies into
broth or upon sloped agar in tubes containing condensation water.
In cases in which streptococci are likely to be scanty, part of the
centrifugalised deposit may be used to inoculate the agar plates.

The tests recommended to differentiate the streptococcus
strains isolated are the following : morphology, growth upon
sloped nutrient agar, growth in nutrient broth, growth upon
gelatine slope, action upon litmus milk, the production of acid in
lactose, saccharose, salicin, mannite, raffinose, and inulin.

The sugar-alcohol media for the differentiation of strepto-
cocci were introduced by Gordon. Their method of preparation
is given in the Appendix.

For some purposes it is of great value to ascertain the patho-
genicity of isolated strains of streptococci. This is conveniently
done by injecting mice subcutaneously or intraperitoneally.

The presence or absence of streptococci in milk may also be
studied by a careful examination of the centrifugalised deposit
stained by methylene blue. Failure to find streptococcus chains
does not mean they are absent, but only suggests they are not
present in considerable numbers. The stained deposits from
samples of vended milk, usually show numerous streptococci,
but in those made from fresh byre milk samples they are as a
rule not to be demonstrated.

Comparative value of above methods,

The estimation of the cellular content is of great value in the
examination of the milk of individual cows. Cellular elements
are invariably present in milk even from perfectly healthy cows
but as a rule not in large numbers. Numbers varying from
50 to over 1000 per cubic mm. have been recorded by the writer
from the milk of individual healthy cows. Any considerable
increase is due to physiological or pathological changes (old
or recent) in connection with the udder or teats.

Examples of physiological conditions are advanced pregnancy
or the first few days after calving. Pathological causes are any

MILK 109

local inflammatory or suppurative conditions. For these con-
ditions the number of cells present may be very high and in
cases of definite mastitis numbers as high as 300,000 per cubic mm.
have been found by the writer.

This estimation may be very useful to detect the commencing
stages of mastitis and it may be said that in general when
cellular elements are present in numbers more than 800 or so
per cubic mm. they indicate the need for careful inquiry as to
the local conditions of the cow's udder or teats.

For ordinary mixed milk samples this determination is of no
particular value.

The significance of streptococci in milk is a complex and
difficult subject. They are undoubtedly very prevalent not only
in ordinary samples of market milk but also in samples of mixed
milk collected immediately after milking when no multiplication
in the milk can have taken place. For example the writer found
streptococci present in every sample of mixed milk (although
many were collected at the farm) in a large series which he
examined, when I c.c. was sampled and in over 80 per cent,
when O'l c.c. was examined. Also streptococci were present in
over half the samples of milk examined drawn direct from
the teats of cows.

It would appear that there are four main sources of strepto-
cocci in milk, i.e. the teat passages and milk cisterns of healthy
cows, manurial contamination (streptococci are present in cow
manure O'l to 10 millions or more per gramme), stale milk from
unclean milk vessels and local disease of the cow (mastitis, etc.).

It is obvious from these varied sources of origin that the
presence of streptococci in moderate numbers in milk samples
cannot be accepted as of necessity of prejudicial significance or
as pointing, without further investigation, to local disease in one
of the cows supplying the milk.

Whether their presence in very large numbers in fresh milk
can be taken to indicate inflammatory disease of the milk-
secreting organs of one or more of the cows supplying is a
question which cannot be said to be at present settled, but it is
certainly an unsatisfactory condition and should lead to careful
examination of the cows.

110 MILK

In cow mastitis and some cases of teat ulceration the lesions
show streptococci in enormous numbers. In one instance the
writer found that when the cows of a cow-keeper whose fresh
milk contained enormous numbers of streptococci were examined
severe teat ulceration, associated with abundant streptococci,
was found upon several of the cows.

The differentiation of the streptococci found in mixed milk
samples is scarcely sufficiently advanced to furnish reliable
guidance as to which varieties are prejudicial but such differ-
entiation is more valuable when dealing with samples from
individual cows.

Mastitis, or garget, in cows may be due to a number of
different organisms, but in the writer's investigations 75 per cent,
of the cases were due to streptococci, while 80 per cent, of the
streptococci formed one common type, the Streptococcus mastitidis.
This is a long chain streptococcus, growing rapidly in broth, form-
ing a coherent deposit, but leaving the upper part clear. It grows
upon gelatine without liquefaction, produces acid in milk, and
clots it within three days, gives no neutral red reaction, and
produces acid in lactose and saccharose media, never in mannite,
and not usually in salicin, raffinose, or inulin. It is non-
pathogenic to mice. Inoculated into the teats of goats, it sets
up mastitis on the infected side1.

In investigating human outbreaks of sore throat, etc., spread
by milk and associated with the presence of a cow with mastitis
in the herd supplying the milk the best way to establish or
confute any causal relationship is to promptly ascertain if the
streptococci from the human disease can cause mastitis in goats
by infection of the teats. If they are of milk origin they should
produce, if not mastitis, at least a marked inflammatory reaction
in the inoculated goat2.

Summary of procedures recommended.

A number of different methods have been discussed. For
the convenience of those not versed in the examination of milk
samples the following procedures are recommended :

1 For detailed consideration of the bacteriology of mastitis see Local Govt. Board
Medical Officer's Report 1906-7, 1907-8.

2 Local Govt. Board Medical Officer's Report 1908-0,

MILK III

I. Ordinary mixed milk samples for general bacterial con-
tamination.

Make dilutions A, B, C, D as above.

(a) Examine the stained centrifugalised deposit from loc.c.
If streptococci are very numerous enumerate them using a series
of neutral red broth preparations.

(b) Estimate the number of B. coli and allied organisms.

(c) „ „ B. enteritidis sporogenes spores.

II. Samples to be specially examined for the tubercle bacillus.
Examine both microscopically and by animal inoculation.

III. Special examination of the milk of individual cows.

(a) Carefully examine the stained centrifugalised deposit,
including a differential cellular count.

(b) Estimate the number of B. coli and allied organisms.

(c) Estimate the cellular content.

(d) Examine for and estimate the number of streptococci.
Brush definite fractions over plates of suitable solid media and
isolate the chief varieties present.

(e) If necessary examine specially for B. tuberculosis.

CHAPTER VII

MODIFIED MILK AND MILK PRODUCTS

I. Condensed Milk.

Condensed milk is prepared by concentrating either whole
or separated milk, with or without the addition of sugar.
Theoretically, therefore, there are four kinds of condensed milk,
but the writer is not aware of any unsweetened separated
brands. The four kinds are :

(a) Sweetened condensed whole milk.

(b) Unsweetened condensed whole milk.

112 MODIFIED MILK AND MILK PRODUCTS

(c) Sweetened condensed separated milk.

(d) Unsweetened condensed separated milk.

Bacteria are present in all varieties of condensed milk, but
in relatively small numbers, the method of preparation being
sufficient to eliminate most of the bacteria present in the
original milk. The following figures have been recorded : —
230 to 70,940 per c.c. by Heggs, 130 to 20,000 per gramme by
Klein, 20,000 to 120,000 per c.c. by Dold and Garratt. Bold
and Garratt1 found B. colt and B. enteritidis sporogenes absent
from I c.c. in all the 19 samples examined, but streptococci
were found in I c.c. in 8 cases (32 per cent.) and in o-i c.c. in
3 cases. Pathogenic organisms could not be detected in any of
the samples, either by direct examination or animal inoculation.

Gordon and Elmslie2 examined 15 different brands of con-
densed milk, and found none of them sterile. Streptococci
with characters similar to those found in milk were present in
all the II sweetened samples. These investigators concluded
that during the process of condensing unsweetened milk, sterility
is secured, the organisms found being subsequently introduced
from the air, but that in the condensing of sweetened milk
sterility is not attained, some at least of the organisms in
the original milk surviving the process of condensation. No
organisms of the B. coli group were isolated from any of their
samples.

Delepine2 carefully investigated the resistance of tubercle
bacilli to the processes involved in the preparation of condensed
milk, and found that although the original milk had been made
artificially highly virulent, after condensation its power to set up
tuberculosis in guinea-pigs was lost, all the tubercle bacilli being
killed.

The analysis of condensed milk is mainly chemical, bac-
teriological examinations being seldom required.

The bacteriological examination of condensed milk should
be on the same lines as for ordinary milk, the sample being first
thoroughly mixed with a definite quantity of sterile water.

1 Journ. Roy. Inst. Public Health, 1910, XVIII, p. 794.

2 Report by Dr Coutts to Local Govt. Board on Condensed Milk, 1911.

MODIFIED MILK AND MILK PRODUCTS 113

Care must be taken to thoroughly cleanse and sterilize the
outside of the tin before opening, while the instrument to open
the tin, which may conveniently be an ordinary tin-opener, must
be sterilized before use.

The fact that condensed milk is a culture medium for
bacteria — although not a very suitable one — is of much greater
importance than the fact that it contains a few bacteria. The
latter fact is evidently of no public health significance.

II. Dried Milk.

Dried milk is now manufactured commercially to a con-
siderable extent. It is the powder obtained either by passing
milk rapidly between heated surfaces so that it is deprived of
its water, or by drying on a cylinder in a partial vacuum.
This dry powder, on being again mixed with water, is converted
into a fluid which looks like milk, and which, on ordinary
chemical analysis, shows the chemical constituents of that sub-
stance. In other respects considerable alterations have taken
place, for example the enzymes have been destroyed, the fat
globules physically altered, etc.

To examine bacteriologically the dried milk would have to
be dissolved in sterile water, and the mixture examined by
methods similar to those employed for milk.

The chief examinations which might be required would be
to determine if sterile, or if not the number and kinds of
bacteria present, and also to ascertain if tubercle bacilli were
present.

From the limited number of experiments which have been
made it appears that tubercle bacilli are killed in the process of
manufacture.

III. Cream.

The bacteriological examination of cream is of considerable
importance, since it is usually richer in bacteria than either
whole milk or separated milk. This is shown by direct bacterial
examinations, while in guinea-pig inoculation experiments it is
not uncommon to find that a considerably higher percentage

s. w. 8

114 MODIFIED MILK AND MILK PRODUCTS

succumb to acute infections when inoculated with cream than
when inoculated with the sediment of whole milk. In examina-
tions of milk samples for pathogenic bacteria, such as B. diph-
theriae, B. tuberculosis, etc., the examination of the cream should
never be neglected, while it follows that market-cream itself
should be more often examined than is the case.

Anderson1 gives the following figures as to the number of
bacteria in cream. In 26 samples of milk the average number
of bacteria in the sediment and in the cream layer obtained
from gravity and centrifugally raised cream was —

. J cream layer 68,690,000 bacteria.

Gravity raised j sediment layer ^^ ^

f cream layer 06,840,000

Centnfugahsed | ^.^ layer 18,840,000 \\

Whole milk ... 14,388,000 „

In Washington, U.S.A.2, the examination of 130 samples of
cream between February and July, 1907, showed that while all
showed over 10,000 organisms per c.c. only three contained less
than 50,000, while 10 per cent, contained between 10 and
25 million bacteria per c.c. The average for all the samples
was 12,130,000 bacteria per c.c.

The bacterial content of cream varies enormously with its age.
Quite fresh centrifugalised cream usually has a bacterial content
lower than that of the milk from which it is obtained. Ac-
cording to Swithinbank and Newman the result of centrifugalising
milk is that roughly 60 per cent, of the organisms will be found
in the sediment, 25 per cent, in the separated milk, and
15 per cent, in the separated cream. Subsequently, however,
on keeping the bacteria multiply enormously, cream being a
good nutrient medium. The Commission on Milk Standards
appointed by the New York Milk Committee recommended as
a standard for cream that it should not contain more than
300,000 bacteria per c.c.

Cream can be bacteriologically examined by methods similar
to those used for milk, the cream being diluted with sterile
water. For examination for tubercle bacilli the cream itself

1 Bulletin 56 Washington Treasury DcJ>arttttenty 1909, p. 739.
a Ibid. p. 259.

MODIFIED MILK AND MILK PRODUCTS 115

may be directly injected into guinea-pigs. A chemical examina-
tion for preservatives should always be carried out before the
bacteriological examinations are undertaken.

IV. Butter.

The bacteriology of butter comes under consideration from
two points of view. Part of the changes which occur in the
conversion of the cream into butter are bacterial in origin, and
much investigation has taken place as to the precise bacteria
involved, the changes they cause and the influence of the other
kinds of bacteria present. This economic bacteriology of butter
is of the utmost importance in relation to the production of
good butter, but can only be briefly mentioned here. The other
point of view is concerned with the part butter may play as a
vehicle for the transmission of pathogenic bacteria.

As pointed out in the preceding sub-section, bacteria are
very numerous in cream, and are of many different varieties.
Some of these are active participators in butter ripening, others
are prejudicial and may cause abnormal flavours, etc. Butter can
be made from either sweet or soured cream. Sweet cream butter
is prepared at once, while to make the latter the cream is
allowed to become sour, by the development and chemical
activities of the lactic acid group of bacilli, before it is converted
into butter.

If the cream is pasteurised first and the requisite bacteria
subsequently added in the form of the so-called " starters," the
process is better controlled, and therefore more uniform quality
butter is produced, abnormalities of taste, odour, etc., due to
foreign bacteria are prevented, while the keeping qualities of
the butter are improved.

In America to some extent, and in Denmark for the most
part, the butter is made from pasteurised cream to which starters
are added, while in England it would appear that the greater
part of the butter is made naturally, that is by trusting to the
suitable bacteria being naturally present.

The " starters " are more or less pure cultures of lactic acid
bacteria.

8-2

Il6 MODIFIED MILK AND MILK PRODUCTS

English butter usually contains when fresh a very large
number of bacteria, often one to ten million, and sometimes as
many as 40 million or more per gramme. The majority of these
are lactic acid bacilli such as B. acidi lactici, B. lactis aerogenesy
etc. The number of bacteria materially decrease with keeping.

Butter "faults" are not infrequently met with. The majority
of these are due to faulty management, resulting in the develop-
ment of unwelcome species of bacteria which produce rancidity,
abnormal flavours, bitter taste, etc. Of these unfavourable
bacteria may be mentioned B. subtilis, B. mesentericus, B. fluo-
rescens liquefaciens, etc., and probably some yeasts. Some of
these changes, for example rancidity, are due to the inter-action
of several kinds of bacteria. These unfavourable bacteria are
usually kept in check by the lactic acid bacteria.

If the cream is pasteurised and pure cultures added there is
obviously much less risk of these abnormal bacteria developing
and producing butter faults.

No data of special value are likely to be obtained from an
enumeration of the bacteria in ordinary butter samples. From
the public health point of view the bacteriological examination
of butter is at present limited to its examination for pathogenic
bacteria, particularly B. tuberculosis.

Pathogenic bacteria in butter.

B. tuberculosis. The investigations of numerous workers have
proved that tubercle bacilli are not infrequently present in butter.
Different investigators have, however, obtained very varying
results, their percentages of positive results varying from about 8
to 30. There are few available English records, but it is probably
safe to state that at least 10 per cent, of samples contain 'living
virulent tubercle bacilli. A recent report of the Local Govern-
ment Board1 states that a number of butter samples were
examined in the Board's Laboratory. Of 48 samples of foreign
butter four were " inconclusive," the others showed no tubercle
bacilli. Of 60 samples of British origin 22 were "inconclusive,"
while two showed the presence of tubercle bacilli.

J Local Govt. Board Medical Officer's Report, 1911-12, p. 184.

MODIFIED MILK AND MILK PRODUCTS 1 1/

Experiments have shown that the tubercle bacillus will remain
alive for considerable periods in butter, i.e. five months or longer
(Mohler, Washburn and Rogers). Teichert on the other hand
only found the bacilli alive up to 18 days.

To detect tubercle bacilli in butter the inoculation method
is the only satisfactory one. The butter is placed in centrifugal
tubes which are stood in warm water at 42° C. until the butter is
completely melted. The material is centrifugalised when liquid,
and the sediment inoculated into guinea-pigs as described under
milk. It is difficult to keep the butter liquid during the centri-
fugalisation.

It is important to remember that acid-fast bacilli may be
present in butter, and that not only may they interfere with
a simple microscopic examination, but that they may produce
pathogenic lesions closely resembling those caused by the
tubercle bacillus including the death of the animal. The Butter
bacillus isolated independently by Rabinowitsch and Petri from
butter is an acid-fast organism which morphologically resembles
the tubercle bacillus, and which will produce similar lesions in
guinea-pigs when injected intraperitoneally mixed with butter.
It can be readily distinguished from the tubercle bacillus by the
comparative rapidity of its growth on ordinary media, such as
nutrient-agar or glycerine-agar, a well-marked thick crinkled
growth being present after three to four days.

In every case in which butter is being examined for
B. tuberculosis it is very important not only to make microscopic
films from the affected organs and demonstrate acid-fast bacteria,
but also to make cultivations upon glycerine-agar, to ascertain if
the butter bacillus is present and the cause of the lesions. It is
also advisable to subcultivate upon blood serum and egg medium
to isolate any tubercle bacilli present.

Other pathogenic bacteria. The diphtheria bacillus appears to
have but a short life in butter, but the typhoid bacillus has been
shown by different observers to live for at least ten days. No
definite outbreaks appear to have been traced from their presence
in butter.

To examine, liquefy at 42° C. and centrifugalise as for tubercle

Il8 MODIFIED MILK AND MILK PRODUCTS

bacilli in butter. The sediment and liquid are pipetted off from
the fat and re-centrifugalised. The sediment is then examined
by methods similar to those described under water and milk.

V. Cheese.

The changes which occur in cheese ripening appear to be
partly chemical, partly bacteriological, the latter being more
particularly concerned with the processes which give the different
flavours. There appears to be great obscurity and difference of
opinion as to the bacteria concerned and the actual parts they
play.

The number of bacteria in cheese is always large, but varies
with the age of the cheese. According to Russell1 the ripening
process can be divided into a period of initial bacterial decline
soon followed by a period of great bacterial increase, and ending
with a period of bacterial decline.

A large number of different bacteria has been described in
cheese, the lactic acid bacteria forming the largest group. In
addition gas-producing bacteria, bacteria decomposing casein,
moulds, etc., are all numerous.

From the public health point of view the bacteriological
examination of cheese is seldom undertaken, and is of little
value. The only examination of any importance is for B. titber-
atlosis. Hormann and Morgenroth found tubercle bacilli in 3
out of 15 samples, Rabinowitsch in 3 out of 5 samples, and Eber
in 2 out of 50 samples.

Tubercle bacilli may also live for some time in cheese.
Using milk artificially inoculated with tubercle bacilli and then
made into cheese, Galtier found the bacilli alive in cheese
2 months and 10 days old. Harrison made cheese from milk
artificially inoculated with tubercle bacilli. In cheese made by
the Emmentaler method, they died between the 34th and the
4<Dth day; in Cheddar cheese, after 62 to 70 days.

Mohler, Washburn and Doane2 have shown that tubercle
bacilli may live for even longer periods. In their experiments

1 Agric. Exp. Station, Wisconsin, \^th Annual Report, 1896, p. 105.

2 Annual Report, Bureau of Animal Industry, U.S.A., 1909.

MODIFIED MILK AND MILK PRODUCTS lip

they found that infected cheese 220 days old set up, when
inoculated, marked generalized tuberculosis, and when 261 days
old slight tubercular lesions.

To examine cheese for tubercle bacilli the method of Mohler,
Washburn and Doane may be used. Portions of the central
part of the cheese are rubbed up in a mortar with sterile normal
saline solution. After being well ground the liquid is strained
through a layer of absorbent cotton and the equivalent of
2 grm. of cheese injected into each guinea-pig.

VI. Ice-cream.

The term ice-cream covers articles which are to some extent
prepared differently and which contain different constituents, but
these differences of composition are not extensive as regards
ice-creams made in this country.

In its simplest form ice-cream consists of milk and sugar,
with a thickening agent such as corn-flour and with or without
eggs. Flavouring agents and sometimes colouring matters are
added, while gelatine is sometimes used in the manufacture.

Cream, as such and apart from milk, does not appear to be
used in the preparation of the article vended in this country as
" ice-cream." While the mixture is always heated the duration
of the heating varies greatly, and frequently some constituents
are added afterwards. The mixture is cooled and subsequently
frozen.

From the bacteriological point of view the chief points of
importance are that the constituents form a food highly nutrient
for bacteria, that it is often made from materials already heavily
infected with bacteria, that this food is heated but not sterilized,
that the cooling is natural and therefore prolonged, and that the
subsequent freezing has no inhibitory action upon any of the
harmful bacteria which it may contain.

Bacterial content.

Buchan found that the average period of cooling before freez-
ing was about 2o| hours. From numerous reports by Medical
Officers of Health it is evident that both the preparation

120 MODIFIED MILK AND MILK PRODUCTS

of this food and the prolonged period of cooling take place in
very many cases under most insanitary conditions. It is not
therefore to be wondered at that ice-cream is frequently found
to be very heavily contaminated with bacteria, and that a
number of outbreaks of illness has been reported from its use.

The following data- from a valuable paper by Buclian1 will
give some idea of the average bacterial condition of ice-cream :

a b c

Organisms per c.c. (averages) ... 867,000 13 millions 372 millions
Percentage showing : — Percentages

Coli group organisms in i c.c. or less 16 58 94

" Enteritidis change" 10 c.c. or less 46 64 80

»» »» ioo c.c. ... 82 92 96

Streptococci in o'oi c.c. or less ... 6 36 68

i „ ... 38 52 82

a = fifty samples collected immediately after heating.
b= ,, ,, ,, at varying intervals after cooling had begun.

c— ,, „ ,, after the material had been frozen.

Pennington and Walter2 examined sixty samples of com-
mercial ice-cream in America, finding the number of bacteria to
vary from 50,000 to 151,000,000. Streptococci were found in
80 per cent, of the samples.

Bacteriological examination.

The ice-cream should be collected in a sterile vessel — e.g., a
wide-mouthed sterile bottle with glass stopper — and packed in
ice if it cannot be examined at once.

To examine, melt the ice-cream by placing for fifteen to
twenty minutes in the 22° C. incubator, then treat as a milk
sample.

The degree of dilution and the methods of examination are
similar to those used in the examination of milk.

The examinations usually carried out are : — estimation of
the number of bacteria, estimation of B. coli group organisms,
with identification if necessary of the chief kinds present,
enumeration of streptococci, estimation of spores of B. enteritidis

1 Journ. of Hygiene, 1910, X, p. 93.

8 New York Med. Journ. 1907, LXXXVI, p. 1013.

MODIFIED MILK AND MILK PRODUCTS 121

sporogenes. All these are carried out as for milk samples.
Pathogenic bacteria such as B. tuberculosis, B. diphtheriae and
B. typhosus occasionally have to be examined for, using the
same procedures as described under milk.

In addition B. enteritidis and other Gaertner group bacilli
will occasionally have to be examined for. A number of food
poisoning outbreaks have been ascribed to the consumption of
ice-cream, and some if not most of these are due to infection of
this food with Gaertner's group bacilli. The method of exami-
nation to employ is given in Chapter VIII.

Bacterial standards.

It is not at present practicable or expedient to lay down
definite standards of number of bacteria or of special kinds to
be permitted in ice-cream. Rough standards as a guide to the
sanitary conditions and cleanliness precautions observed would
however be of great utility. For these purposes standards based
upon the number of B. coli group organisms would probably be
of greatest service.

Buchan has suggested that ice-creams made under clean
conditions : —

(a) Should not contain more than 1,000,000 organisms
per c.c. capable of growing on nutrient gelatine (reaction + I per
cent.) at 20° to 22° C. in three days.

(b) Should not contain more than 1,000,000 organisms
per c.c. capable of growing on nutrient agar (reaction -f I per
cent.) at 35° to 37° C. in two days.

(c) Should not produce acid and gas in bile-salt glucose
broth with a less quantity than 0*1 c.c.

(d} Should not contain the Bacillus enteritidis sporogenes in
less than 10 c.c.

(e} Should not contain streptococci in less than O'ooi c.c. of
ice-cream.

122 BACTERIOLOGY OF MEAT AND MEAT PRODUCTS

CHAPTER VIII

THE BACTERIOLOGY OF MEAT AND MEAT PRODUCTS

Until the last few years but little was known of the normal
bacteriology of meat and meat products and much work is still
required before many important points related to meat con-
ditions can be regarded as settled. The text books contain
little or no reliable information and much of the following is
contained in reports and journals not readily accessible. For
this reason rather more extended information is supplied than
perhaps would otherwise be warranted by the importance of
the subject-matter.

I. The bacteriology of meat, fish, made-meats, etc., derived from

healthy animals.

Little or no work in this direction has been done in this
country, but under the better regulated meat inspection arrange-
ments which prevail in Germany numerous investigations have
been carried out in recent years. The following will illustrate
the findings.

Unprepared meat.

Gaertner in 1908 found that bacilli were present on the
surface of fresh meat but had not penetrated, if the examination
was made within three days of slaughter, while by the end of
ten days they had not penetrated more than 0*5 inch into the
interior. These results were confirmed by Forster (1908) as
regards the general sterility of the interior of fresh meat.

Conradi1 in 1909 used a special enrichment method (see page
128) to detect bacilli and obtained different results. He ex-
amined 162 organs from 150 healthy animals (bovines, calves
and pigs) slaughtered at Neunkirchen Slaughter House and

1 Zcit. f. Fldsch. u. Milchhyg. 1909, July, p. 341.

BACTERIOLOGY OF MEAT AND MEAT PRODUCTS 123

found bacteria in 72 cases. Both aerobic (42 strains) and
anaerobic bacilli (30 strains) were isolated. The aerobic bacilli
were B. coli communis, B. lactis aerogenes, Streptococcus acidi
lactici, B. mesentericus ', B. fluorescens non-liquefaciens, Diplococcus
pneumoniae and B. suipestifer. The anaerobic bacilli were chiefly
of the butyric acid group but the varieties were not accurately
defined.

Bierotti and Machida1 employed a modified Conradi method
and obtained very similar results. They examined 54 organs
from II quite healthy animals and found bacilli in 32.

Their results and those of Conradi are grouped together in
the following table.

Organ

Liver

Muscular tissue
Kidneys
Lungs ...
Lymphatic glands

Spleen

Testicles

Heart

216 104 48

Horn2 examined oxen slaughtered in the Leipzig Slaughter
House, using a modified Conradi method (enrichment method)
and also direct examination. He found, as might be expected,
that the number of bacteria in the muscles varied with the time
since slaughter. His results are briefly summarised in the
following table.

Musculature of healthy animals.

No. of specimens
examined

Bacteria
found

Percentage
positive

74

50

67-5

70

25

36

28

10

36

16

12

75

4

I

25

22

6

27

I

0

0

I

0

o

Time after

Number examined

Positive results

Percentage positive

slaughter

f * ^

f * >>

* — ,

when samples

with without

with without

with without

removed

enrichment

enrichment

enrichment

24 hours

36

36

5

2

T4

5*6

1—3 days
3 — 6 »>

7
13

7
13

o
4

0

2

o
307

0

r5'4

6—21 „

24

24

13

6

' 54

25

1 Mnench. Med. Woch. 1910, vol. LVII, p. 636.

2 Zeit. f. Infekt. Krank. der Haustiere> 1910, vol. vin, p. 424.

124 BACTERIOLOGY OF MEAT AND MEAT PRODUCTS

No proper differentiation of the isolated bacteria was carried
out.

These results may be compared with those obtained by the
same observer in regard to emergency slaughtered animals,
i.e. animals killed on account of some diseased condition.
Examining 49 specimens of musculature Horn obtained 31
per cent, of positive results (31 with and 21 without enrichment
methods).

Zwick and Weichel1 also investigated this matter using five
different methods. Their results agree with those of Conradi as
regards the frequency of the occurrence of bacteria in the liver,
but not as regards the other organs. For example, in only one
out of the 63 musculature specimens examined were bacilli
found (B. coli group).

The general results obtained by different investigators show
that bacteria may be present on the surface of meat, this being
largely a question of cleanliness and amount of contact with
bacteria-containing materials, and also in the depth of the meat.
As regards musculature the evidence is not concordant, but as
regards the liver, and to a lesser extent other organs, it is evident
that they may contain bacilli, some of which are possibly patho-
genic to man. The chief value of the above observations is
however rather to show that in examining organs from possibly
diseased animals too much stress must not be attached to the
mere presence of bacilli without corroborative evidence as to
their disease-producing role.

As regards fish Ulrich2 found that the number of bacteria in
or on raw fish is considerable even at ordinary temperatures ;
the bacilli present being largely B. coli and B. proteus group
organisms.

Prepared meat.

Chopped meat (Hackfleisch). German bacteriological
examinations do not usually mention the actual numbers and
kinds of bacteria present, but it is evident that very large
numbers of bacteria will be present. Gaertner group bacilli

1 Arbeit, a. d. Kais. Gestmd., 1911, vol. XXXVIII, p. 327.
8 Zeit.f. Hyg. 1906, LIII, p. 176.

BACTERIOLOGY OF MEAT AND MEAT PRODUCTS 1 25

have been found by some workers but not by others when
adequate differentiating tests are used. Zweifel1 at Leipzig
examined 248 specimens of " hackfleisch " obtained quite fresh,
and in 165 cases found B. proteus and other bacilli, all non-
pathogenic to mice by feeding.

Sausages. According to the writer's own examinations,
bacilli are present in large numbers and the kinds present
usually include bacilli of excretal type. The bacterial content
in B. coli group organisms in 27 sausage samples obtained from
different sources examined by the writer2 was as follows : —

B. coli organisms No. of specimens

Less than 10 per grm. of sausage-meat o

JO— 100

IOO — IOOO
1OOO IO,OOO

10,000 — 100,000
Over 100,000

Over loo but number not estimated ... ... 4

The number of organisms present was determined for a few
of the samples and varied from 360,000 to over 600,000 per
gramme.

Streptococci were also present in large numbers in a majority
of the samples examined.

These results were obtained with sausages examined quite
freshly prepared, all being made the same day.

Evidently in sausages as usually manufactured there is a
considerable access of material containing excretal bacilli.

Brawn. Brawn is or should be thoroughly boiled when
prepared so that whatever its original bacterial contamination it
should be free from B. coli and non-sporing bacteria when made.
As a rule this is the case, but brawn is a material very favourable
to bacterial multiplication so that if, after preparation, it is
placed in positions liable to bacterial contamination specimens
will soon show a high bacterial content. Of n samples
examined by the writer B. coli was absent in 0*1 grm. in seven,

1 Centralb. f. Bakt. 1911, vol. LVIII, p. 115.

2 Journ. Roy. San. Inst. 1908, vol. XXIX, p. 366, and Local Govt. Board Medical
Officer's Report, 1909-10, p. 446.

126 BACTERIOLOGY OF MEAT AND MEAT PRODUCTS

present I — 10 per grm. in one, present 100 — 500 in one, present
5000 — 10,000 in one, while in the remaining sample these
organisms were about 50,000 per grm. In several instances the
brawn samples were taken from open shelves, etc., in places
obviously markedly exposed to dust and possibly other forms
of contamination.

In regard to other made foods very little data seem to be
available as to their bacterial content under ordinary conditions.

In ordinary brines used for salting meat B. coli group
organisms are frequently present in large numbers.

II. The bacteriological examination of meat and meat products.

Such an examination may be required for the following
purposes :

(A) Examination for pathogenic bacteria.

(B) To study the general bacterial content, degree of
bacterial contamination, etc.

A. Examination for pathogenic bacteria.

The examination of meat for pathogenic bacteria can ob-
viously only be profitably undertaken in cases in which the
animals have shown symptoms of disease during life or which
post-mortem show definite pathological lesions. It is beyond
the scope of the present work to describe these lesions or even
enumerate the diseases in animals which give rise to them. For
this purpose and for a description of the bacteria which may
have to be looked for text books on veterinary bacteriology and
on meat inspection and examination must be consulted. For
practical convenience these bacteria may be classed together
into three groups :

(1) The bacteria found in septic conditions in animals.

(2) The bacteria associated with food poisoning.

(3) The bacteria found in special diseases such as tuber-
culosis, actinomycosis, glanders, anthrax, quarter-evil, etc., etc.

The bacteria associated with food poisoning are of great
importance from the Public Health point of view and their
examination and isolation are considered in detail below. As

BACTERIOLOGY OF MEAT AND MEAT PRODUCTS 12?

regards other bacteria no satisfactory procedure can be devised
or would be suitable for the general examination of diseased
meat, and the only satisfactory procedure is for bacteriological
examinations to be combined with careful clinical and patho-
logical investigations, the nature of the bacteriological examina-
tion required being largely indicated and determined by the
information furnished in this way.

B. To study the general bacterial content, degree of bacterial
contamination, etc.

Such an examination should be of much greater practical
value than is, at present, the case. For example, in cases of
incipient and early putrefaction meat may be seized and con-
demned by the meat inspector as unfit for food. Sometimes
his findings are disputed and it would be decidedly valuable if
the results could be corroborated by definite bacteriological
data from the results of the examination.

Unfortunately it cannot be said that any very complete and
reliable bacteriological studies are available to elucidate the
problem. Putrefaction is a complex process and the bacteria
concerned are numerous, and both aerobic and anaerobic.
Proteus strains are common in putrefying materials, but they
are a widely distributed group of organisms with most loosely
defined characters, and we cannot say with present knowledge
that their presence, or even their presence in large numbers, can
be accepted as evidence of incipient putrefaction. Much pioneer
work requires to be done in this direction.

Bacterial enumerations are sometimes of value to study the
cleanliness precautions adopted, etc., such as in the examination
of brawn or sausages. For this purpose an enumeration of the
number of B. coli group organisms is the most valuable. Samples
must be collected with care and exact particulars given of
source and particularly of the age of the specimens as regards
time since slaughter. Meat material provides an excellent culture
medium so examinations must be undertaken without delay or
the sample ice packed. Sterile wide-mouthed bottles are the
most convenient for transmitting samples to the laboratory.

If a quantitative enumeration of the number of B. coli or

128 BACTERIOLOGY OF MEAT AND MEAT PRODUCTS

bacteria generally is desired it will be necessary to make dilu-
tions. An accurate enumeration cannot be made since all the
bacteria cannot be brought into the watery emulsion obtained
from the solid meat, but comparable results can be obtained as
follows: — select the piece for examination, cut it up as finely
as possible with sterile instruments and introduce 2 grammes
(weighing in on a balance) into a bottle with glass or india-
rubber stopper containing 20 c.c. of sterile water. After thorough
mixing various quantities of the emulsion are used for examina-
tion, diluted fractions being obtained in the usual way.

For purposes of calculation it is assumed that the organisms
in the meat are all contained, after mixing, in the sterile water
and the results are recorded as per gramme of material.

To detect bacteria and study the varieties present the usual
plan is to make linear cultivations on tubes or plates of sterile
media with small fragments removed from the meat or organs
under examination. The bacteria which grow are then studied
in detail.

Conradi's enrichment method mentioned above is a more
certain method of detecting any bacilli present, but is somewhat
complicated for routine work. It is briefly as follows : — imme-
diately after the animal has been slaughtered a piece weighing
about 50 grammes of the organ to be examined is removed
with sterilized knife and forceps and then placed for four hours
either in 2 per cent, corrosive sublimate at 37° C. or, if it is to
be forwarded to the laboratory, in 0*2 per cent, sublimate. On
its reception at the laboratory the organ is placed in a large
sterilized conical glass with overlapping cover, which can be
hermetically sealed by resin-wax. It remains in this sterile
moist chamber at 37° C. for a further period of 12 — 16 hours.
The piece is then divided into two. The centre of one half is
plunged into fluid nutrient gelatine and kept at 37° C. to develop
anaerobic bacilli. The other half is smeared over the surface of
sterile plates, those selected by Conradi being, in order, a brilliant
green picric acid plate, a plate of Drigalski-Conradi medium
and a plate of nutrient agar. Finally a hanging-drop cultiva-
tion and a microscopic specimen stained by Gram's method are
made. This method is also used for Gaertner group bacilli.

BACTERIOLOGY OF MEAT AND MEAT PRODUCTS 129

III. Bacterial food infections and food poisoning.

Attacks of food poisoning due to bacteria vary greatly in
seventy and in extent, ranging from mild attacks limited to a
few individuals to extensive outbreaks involving a large pro-
portion of the consumers of the infected food and causing
amongst those affected severe symptoms and some deaths. The
food eaten is generally meat in some form but not invariably so,
and there are many outbreaks on record following the consump-
tion of ice-cream, potatoes, milk, etc. The meat has frequently
been eaten in the form of pies, brawn, sausages, tinned foods, or
other made up meat food1.

The bacteria concerned in these outbreaks may be con-
veniently classed into three groups :

(1) Putrefactive and intestinal bacteria such as B. proteus,
B. coli.

(2) B. botulinus.

(3) Organisms of the Gaertner group.

The respective parts played by these three groups in food
poisoning must be briefly considered.

Putrefactive and intestinal bacilli.

At one time most cases of food poisoning were ascribed to
the chemical activities of putrefactive and intestinal bacilli,
particularly one or other of the organisms described as Proteus
vnlgaris, Proteus mirabilis, etc.

This conception was largely based upon a series of interesting
investigations upon the chemical products of putrefaction. It
was ascertained that when meat was allowed to putrefy certain
basic bodies (called ptomaines by Selmi), which closely resemble
the vegetable alkaloids, could be isolated, and that these bodies
were possessed of highly poisonous properties as shown by their
injection into laboratory animals. The symptoms produced
were in some ways similar to those met with in cases of food

1 For a detailed consideration of the etiology, epidemiology and pathology of Food
Poisoning see "Report to the Local Government Board on Bacterial Food Poisoning
and Food Infections" by W. G. Savage, 1913. New Series, no. 77.
S. \7.

130 BACTERIOLOGY OF MEAT AND MEAT PRODUCTS

poisoning. In consequence it came to be believed that food
poisoning outbreaks were to be explained as due to the ingestion
of food in the early stages of putrefaction, the symptoms being
caused by the presence of ptomaines, hence the popular name of
ptomaine poisoning for these cases and outbreaks.

Later investigations have shown that whether or no putre-
factive bacilli and their toxins play any part, basic nitrogenous
bodies of the nature of ptomaines certainly do not, and the term
" ptomaine poisoning " should be abandoned as incorrect and
misleading.

Most writers, however, still maintain that incipient putre-
faction, due to the products of B. proteus and other putrefactive
bacilli, is a cause of food poisoning. While this may be true
for some individual cases, in the writer's opinion the role of the
putrefactive bacteria in food poisoning outbreaks is an extremely
small one and does not extend to the causation of extensive
outbreaks.

In this connection the greatest burden of suspicion has fallen
upon the Proteus group and to a lesser extent on the different
varieties of theJ?. colt group, but other bacilli have been suspected
in individual outbreaks.

B. botulinus.

This bacillus is the cause of a group of cases of /ood
poisoning, now fortunately very rare, included under the term
botulism and which in their symptomatology are quite different
from the ordinary outbreaks of food poisoning. In these cases
the symptoms are almost entirely referable to lesions of the
central nervous system. Most of the outbreaks have occurred
in Wiirtemberg and other parts of South Germany and have
been due to eating raw sausages, the condition being frequently
spoken of as sausage poisoning.

The symptoms in these cases are due to B. botulinus, a
bacillus isolated by Van Ermengen in 1895 from a poisonous
ham. A large (4 — 6 /u) bacillus, with terminal spores. An
obligate anaerobe. Feebly motile with 4 to 8 flagella. Stains
by Gram's method. Culturally it slowly liquefies gelatine, fer-
ments glucose with gas formation but not lactose or saccharose,

BACTERIOLOGY OF MEAT AND MEAT PRODUCTS 131

does not coagulate milk. All the cultures have a rancid butyric
acid odour. It produces a very powerful toxin, but, like the
tetanus bacillus, has but little power of development in the
tissues. Small doses of filtered cultures injected into rodents
produce symptoms of paralysis and the other symptoms met
with in outbreaks ; larger doses are rapidly fatal.

Organisms of the Gaertner group.

Modern investigation has shown that when outbreaks of
food poisoning are bacteriologically investigated, in the majority
of cases members of the Gaertner group of bacilli are isolated
and can be shown to be the cause of the outbreak. These
bacilli are now recognized as by far the most important cause of
bacterial food poisoning. The members of this group stand in
their cultural characters between the chemically active B. coli
group and the chemically rather inactive typhoid group. They
may be divided into two subgroups, (a) true Gaertner bacilli,
(b) para-Gaertner bacilli. The para-Gaertner bacilli are a number
of organisms, for the most part unnamed, which appear to be not
uncommon in the healthy animal and human intestine. They
are for the most part non-pathogenic and are also differentiated
from the true Gaertner bacilli by failing to be agglutinated by
the anti-sera of members of the true group, and by certain
fermentative properties. Many for example fail to ferment
dulcite or actively ferment salicin.

The true Gaertner group bacilli are culturally indistinguish-
able, but can be differentiated by means of agglutination and
other serological and immunity tests into at least three organ-
isms, i.e. B. enteritidis, B. suipestifer and B. paratyphosus /?.

There is still some controversy as to the question whether
the last two organisms are identical or not, the prevailing
German view being that they are identical, while English in-
vestigators, for the most part, have found differences.

In outbreaks of food poisoning B. enteritidis or B. suipestifer
is found, while B. paratyphosus /3 is found in cases of paratyphoid
fever.

The two former bacilli are also met with in connection with

9-2

132 BACTERIOLOGY OF MEAT AND MEAT PRODUCTS

a number of diseases in pigs (as secondary invaders), rats, mice,
calves, birds, etc.

The characters of B. enteritidis and the other true Gaertner
bacilli are briefly as follows : — short bacilli with rounded ends,
not staining by Gram's method, actively motile, no spores.
They grow readily in broth with or without scum formation,
do not liquefy gelatine, on potato form a white or yellow-brown
growth. Indol is not produced. The growth in litmus milk is
somewhat distinctive as acid is first produced, the medium
becoming subsequently distinctly and often very markedly alka-
line. The sugar-alcohol fermentations are important, glucose,
dulcite, mannite, maltose, galactose and laevulose being fermented
with the production of acid and gas, while lactose, saccharose,
salicin, inulin and raffinose are not fermented, neither acid nor
gas being produced. With media containing glycerine a little
acid is produced with some strains but never gas.

Of these varied cultural tests reliance is chiefly to be placed
upon the characteristic growth in litmus milk, the absence of
indol formation, the power to ferment dulcite and mannite, and
the failure to ferment lactose, saccharose and salicin.

When recently isolated most Gaertner group strains possess
high virulence to rodents, causing gastro-intestinal symptoms,
general infection and death. These symptoms follow intra-
peritoneal or subcutaneous infection and frequently can be
induced, but with far less certainty, by feeding. All the varieties
of the group produce toxins which are remarkably heat resistant.
The bacilli themselves are fairly easily killed (i.e. 30 minutes at
60° C), but their toxins are capable of resisting heating to
1 00° C. for so long as thirty minutes. Under artificial culti-
vation outside the animal body the power to produce heat-
resisting toxins is rapidly lost.

The above properties are shared by all the varieties included
under the true Gaertner group. The distinction between B.
enteritidis and the other two strains is readily made by aggluti-
nation tests. For example, the serum of an animal immunised by
the repeated injection of B. enteritidis in non-fatal doses aggluti-
nates highly and frequently to the maximum titre all strains of
B. enteritidis, whatever their origin, but with strains of B.

BACTERIOLOGY OF MEAT AND MEAT PRODUCTS 133

suipestifcr or B. paratypJwsus (3 has either no power of agglutina-
tion or will only agglutinate them in dilutions far lower than
those which cause rapid agglutination with B. enteritidis. In
the same way the sera of animals immunised against these two
organisms exert comparatively slight agglutination action upon
B. enteritidis.

B. suipestifer and B. paratyphosus (3 often, but not regularly,
show slight agglutination differences, but by the use of absorption
tests definite differences can be made out.

The bacteriological investigation of food poisoning outbreaks.

It is very desirable that all food poisoning outbreaks should
be bacteriologically investigated even if limited to one or two
families. Usually, at the present day, if any examination at all
is made the material is sent to a chemist and a chemical exami-
nation is made for poisonous metals or ptomaines. Naturally
such examinations are negative and the true source of the
outbreak is overlooked.

I. Collection and transmission of material.
The following, if available, should all be sent :

(a) The supposed incriminated food or foods. Samples
should be obtained not only at the source of supply (shop, place
of preparation, etc.), but also from the homes of the sufferers. The
latter is most important, as from these sources the bacteriologist
is much more likely to obtain portions actually infected.

(b) Material from autopsies on fatal cases. This is of
course extremely important, and such material is both easier to
work with and far more likely to show the bacillus concerned in
the infection than is the suspected food.

The materials most valuable to examine are pieces of liver,
the spleen and portions of bones (for bone marrow). A piece
of intestine, ligatured to retain its contents, is also useful.

(c) Blood serum from persons who have been attacked.
A good many different specimens should be collected. Some-
times, particularly if serum cannot be obtained, it is valuable to
examine the urine and excreta of such cases.

134 BACTERIOLOGY OF MEAT AND MEAT PRODUCTS

Care must be taken to accurately label everything and to
note any points in regard to selection, i.e. if samples taken from
outside or inside of a larger piece, etc.

The material must of course be transmitted at once, properly
sealed, and, if any delay is likely, sent in ice-boxes.

2. Bacteriological examination.

A. The suspected foodstuffs. The precise methods of exami-
nation will of course vary slightly with the nature of the food,,
but should be on the following lines :

The physical appearance and smell should be carefully
noted. Any deviations from the natural appearance of normal
food of the kind under examination should be particularly noted;,
for example slight liquefaction.

Aerobic and anaerobic cultivations, feeding and inoculation
tests, may all have to be made.

Mice should be fed with portions of the sample1, while other
animals — e.g. guinea-pigs — should be inoculated subcutaneously
and intraperitoneally from cultures and also from the broth
emulsions of the food. If any of the animals die, a complete
post-mortem examination should be made.

For cultural examination it is important to obtain a uniform
and characteristic sample. This can be conveniently done by
mincing up selected portions with sterile instruments and adding
to sterile water or broth in a flask, mixing thoroughly. If a
quantitative examination is to be made a definite quantity of
food must be added to a definite quantity of water. A fairly
complete examination would include examination for Gaertner
group bacilli, Proteus group bacilli, B. coli and allied organisms
and for anaerobic bacilli such as B. botulinus.

The most reliable method, in the writer's opinion, to isolate
Gaertner group bacilli is to brush some of the emulsion directly
over a series of lactose, neutral red bile salt agar plates (L.B.A.)
and to also add some of the emulsion to dulcite malachite

1 While it is a simple and useful procedure to examine mice the greatest caution
should be used in interpreting the results. Mice not infrequently are carriers of
Gaertner group bacilli and when fed, even with sterile food, may die and Gaertner
group bacilli be isolated post-mortem. Feeding mice is chiefly of value from the
negative aspect.

BACTERIOLOGY OF MEAT AND MEAT PRODUCTS 135

green broth tubes. After 12 — iShrs. incubation the latter, after
dilution, is brushed over a few additional L.B.A. plates. This
broth medium favours the growth of Gaertner group bacilli over
the other intestinal bacteria. The composition of these media
is given in the appendix.

All the white colonies (at least on the primary plates) must
be subcultivated and investigated. The labour is much dimin-
ished, without risk of overlooking true Gaertner bacilli, if to
the L.B.A. salicin and saccharose (J- p.c.) are added as well as
lactose. Many para-Gaertner bacilli are in this way eliminated,
as if they ferment salicin and saccharose they will form red
colonies.

All the white colonies are subcultivated into double tubes of
litmus broth containing salicin, saccharose and lactose. The
tubes, which after two days incubation show neither acid nor
gas, are further culturally examined The colonies which possess
the cultural characters of the Gaertner group must be fully in-
vestigated, including agglutination and virulence tests.

The agglutination tests must be carried out with anti-sera
from both of the food poisoning strains, i.e. B. enteritidis and
B. suipestifer. They should also include the immunization of a
rabbit and testing the agglutination properties of its serum upon
known members of the Gaertner group.

The presence of B. coli group organisms in the food can be
judged from the number of red colonies on the L.B.A. plates.

If it is wished to ascertain if bacilli of the Proteus group are
present this can be done by inoculating 5 per cent, nutrient
gelatine plates with the material and investigating Proteus-like
colonies. This is an unsatisfactory method and a useful medium
for the isolation of this group of organisms is greatly needed.

Anaerobic cultures can be made by the inoculation of glucose
broth and glucose agar plates, incubating all under anaerobic
conditions at suitable temperatures. The different kinds of
organisms present must be carefully examined and investigated.

It is also advisable to cut sections and otherwise micro-
scopically examine the meat to see if the bacilli are chiefly on
the surface and also if the meat fibres are from apparently
healthy animals.

136 BACTERIOLOGY OF MEAT AND MEAT PRODUCTS

B. Material from Jiuuian cases. The examination of materials
from autopsies is conducted in a similar manner but is usually
much easier, the bacilli being frequently present in pure or
nearly pure culture. Only a few L.B.A. and other plates need
to be inoculated from each organ. Frequently the presence of
Gaertner group bacilli can be demonstrated within 24 hours.

The testing of the agglutinative properties of the blood of
actual or suspected cases is a very important part of the investi-
gation, and even when no other material at all is available for
examination is sometimes sufficient to accurately determine the
cause of infection.

The sera must be tested against stock cultures, at least
B. enteritidis and B. suipestifer being used for each specimen.
Low dilutions, i.e. I : 30, should be used for the first test and, if
positive, then the serum tested to the limit of agglutination.

If possible a little of the serum should be retained to test
against any Gaertner group bacilli isolated from the food or
post-mortem materials.

CHAPTER IX

AIR

Bacteria are always present in air, but, unless the air con-
tains a large number of dust particles, in comparatively scanty
numbers.

It is not that they are not discharged into the air, but rather
that they do not thrive in it. Bacteria are constantly being
wafted from dry surfaces, and in this way considerable numbers
gain access to air and, owing to their lightness, are carried con-
siderable distances. Unlike soil, milk, water, etc., air does not
serve as a food for bacteria, and they do not multiply in it. On
the other hand, there are a large number of agencies at work
which diminish the number of organisms which have gained
access to air. Of these the most important are the germicidal
action of light, especially sunlight, and the action of gravity.

Pathogenic organisms are not readily detected in air, the

AIR 137

organisms usually found being moulds and saprophytic bacteria.
The determination of the number of organisms is of greatest use
as a means of comparing methods of ventilation.

Bacterial content of air.

While numerous bacteriological examinations of air have
been carried out from time to time the data given mostly only
consist of the number of bacteria present. Such figures to be of
value must be carefully considered in the light of the precise
local conditions prevailing when they were obtained. The figures
mentioned below will, however, give a good idea of the bacterial
content of air under different conditions and in different places.

Miquel found, as the average of six years' observations,
about 450 organisms per cubic metre in the air of the park at
Montsouris near Paris, while in a street of Paris he found the
average to be nearly 4000.

Carnelley, Haldane and Anderson1 carried out a series of
investigations in Scotland, chiefly in Dundee. They found that
the outside air in winter contained an average of 0*8 micro-
organisms (6 bacteria, 2 moulds) per litre. Some of their other
figures per litre of air were as follows, the results being the
averages of a large series of results. Elementary schools in
Dundee 152 bacteria, county board schools in Scotland 76, one-
roomed dwellings 60, two-roomed dwellings 46, better class
houses 9 bacteria.

Classified in relation to cleanliness, etc., some of their results
are shown in the following table :

Micro-organisms
per litre of air

Elementary schools in

Dundee with

natural ventilation

Elementary schools in

Dundee with
natural ventilation
One-roomed
dwellings in

Dundee

Two-roomed

dwellings in

Dundee

Cleaner than average ... 91

Average ... ... ... 125

Dirtier than average ... 198

Opened over 20 years ... 311

„ 2—20 „ ... 150

,, less than 20 years 38

Cleaner than average ... 18

Average 45

Dirtier than average ... 93

Cleaner than average ... 10

Average 22

Dirtier than average ... 69

1 Philos. Transactions, 1887 B.

AIR

Newman1 carried out some experiments upon the bacteria
in the air of bakehouses. The method used was to employ agar
plates and expose for 30 minutes, subsequently incubating for
22 hours at 37° C. While no accurate bacterial enumerations
could be made in this way valuable comparative data were
obtained. The average number of bacteria falling upon each
9*6 square inches (mean of three plates each) during the 30
minutes, was for underground bakehouses 600 — 800, for above-
ground bakehouses 200, while in the open air outside the
bakehouses it was 160.

Haldane2, using a slightly modified form of Frankland's
method, carried out a number of bacterial enumerations in
factories and workplaces. The following table shows the
average results obtained for most of them :

Average number of bacteria
per litre of air

Number of - " ^

Factory or Workshop observations Bacteria Moulds Total

Clothing Factories ... 7 7 4*5 n'5

Tailoring Workshops ..* 7 TO 2 12

Printing and bookbinding u 47 1*4 6'i

Ropemaking 3 317 10 327

The largest number of bacteria were present in the air of
the ropemaking premises. In one of the three factories 868-
bacteria per litre were found.

These and other available data show clearly that the number
of bacteria in air depends upon the amount of dust stirred up
into the air. Little significance can be attached to the presence
of even large numbers of bacteria in the air if these bacteria are
derived from material which is unlikely to contain disease germs.

The enormous number of bacteria in dust is illustrated from
the following two enumerations made by Gordon3 of the dust of
the Debating Chamber of the House of Commons :

(All per gramme) No. i No. 2

Bacteria 10,000,000 100,000 — 1,000,000

/?. coli loo— IODO 1000—10,000

B. enteritidis sporogenes 100— 1000 1000—10,000

Streptococci ... ... (none isolated) 10 — 1000

i Public Health, 1902, vol. xv, p. 152.

* Ventilation of Factories and Workshops, Departmental Committee, \st Report,
190 1.

3 House of Commons (Ventilation}, Report, by Dr Gordon, 1906.

AIR 139

Sewer and drain air.

The bacteriology of sewer and drain air is of considerable
interest and importance. The earlier investigators, Haldane and
Carnelley, Laws, Andrewes and Laws, all found that the air of
sewers contained but few bacteria, and those found appeared to
be of the same nature as those in the outside air and to be so
derived, and not from the sewage. In both, for example, a
large proportion consisted of moulds while the numbers present
increased or decreased with the degree of ventilation of the
sewers.

Delepine1 found that the average number of microbes in the
air of the Manchester sewers examined during the months of
May, June and July (mean of 33 observations) was about 855 per
1000 litres. Of these 552 were bacteria and 303 moulds.

More recent work has not invalidated the general truth
of these findings although it has demonstrated that under ex-
ceptional circumstances sewage bacteria may be present in
sewer air and, much more frequently, in drain air.

Horrocks2, in 1897, showed that specific bacteria present in
sewage may be ejected into the air under certain conditions,
such as the bursting of bubbles at the surface of the sewage, or
the separation of dried particles from the walls. The alternate
moistening and drying of the walls when the flow of sewage is
intermittent, in his opinion, would favour such detachment.

Andrewes3 has recently carried out three important series
of investigations. He showed that under certain circumstances
characteristic sewage bacteria are to be found in the air of
drains and sewers. The evidence for this is based upon a
careful study of the characters of the bacilli found in sewer air,
using newer and better methods than hitherto employed. For
example, the streptococci of drain air, when tested as to their
biological characters, were found to correspond with those of
sewage, and only to a slight extent with those which chiefly

1 Special Report to Manchester Committee, 1909.

2 Proceedings of Royal Society, February 7, 1907.

8 Local Government Board Medical Officers Report ', 1906-7, p. 183; 1907-8,
p. 266; 1910-11, p. 387.

I4O AIR

abound in fresh air. The bacilli of the B. coli group obtained
from drain air correspond essentially in their characters with
those of sewage, while such bacilli are absent, or nearly absent,
in fresh air. The presence of these bacilli was more readily
detected in the air of drains. When a test microbe (B. pro-
digiosus) was added in bulk to the contents of the drains, it
could be recovered, under suitable conditions of splashing, from
the drain air in great numbers and at considerable distances
from the points at which it was added to the sewage in the
drain.

In a second investigation, using for examination the drain
air of a large public institution and of a private dwelling,
Andrewes showed that sewage bacteria could readily be demon-
strated in the drain air in almost any situation, but that their
presence is of a highly intermittent character. He concluded
that, in the drainage system of a house or institution in ordinary
occupation, the determining cause of the presence of sewage
bacteria in the drain air is droplet contamination from splashing,
such droplets being of extremely minute size. As regards the
content of the drain air in faecal bacteria, the most important
determining factor is the faecal content of the sewage. The
relative number of lactose fermenting coliform bacilli in the
drain air bears a direct relation to the abundance of faecal
material in the sewage.

In a further report Andrewes showed that when the air of
the drain or sewer becomes infected from sewage splashing
natural air currents may transport sewage microbes for con-
siderable distances before they settle.

The scantiness of the bacteria characteristic of sewage in the
air of sewers is also shown by Delepine's1 systematic observa-
tions upon the air of a Manchester main sewer. He found
sewage bacteria to be very scanty, and out of several thousand
bacteria collected by him from sewer air only six proved to be
typical B. coli.

1 Loc. cit.

AIR 141

The bacteriological examination of air.

Of methods available for the bacteriological examination of
air a large number have been advocated and practised, but the
majority of them are but little used at the present day. Of those
now used the following methods may be mentioned :

I. Simple Plate Exposure. — Tubes of sterile agar and gelatine
are poured into ordinary sterile Petri dishes and allowed to
solidify. These plates are then exposed to the air under investi-
gation for definite periods, by simply removing their covers and
replacing them after the required time has elapsed, Fifteen to
thirty minutes is a convenient time of exposure. In removing

Fig. 14. Agar plates exposed for 30 minutes in a school

(a) under ordinary working conditions ;

(b) after the children had marched out of school.

the covers care must be taken that the air is -not unduly agitated,
and that dust — e.g. from the clothes of the investigator — is not
discharged in the neighbourhood of the plates. The plates are
then incubated at 20° to 22° C. for the gelatine, and at 37° C.
for the agar. By using both gelatine and agar plates and by
giving definite times of exposure valuable data may be obtained
as to the relative bacterial state of the atmosphere. The number
of organisms developing should be counted, and, if necessary,
the different kinds investigated by the usual microscopic and
cultural methods. For comparative work the area of the Petri-
dish should be calculated [area = (radius)2 x -^2-], and the results

142 AIR

expressed per square foot per minute. This simple method
gives results nearly as valuable as the more complicated methods,
and in the bacteriological examination of air should never be
neglected.

Additional results of value can be obtained by using Petri-
dishes containing media other than nutrient gelatine or agar.
For example, by using bile salt neutral red lactose agar or
fuchsin agar, valuable data as to the relative number of B. coli
and allied organisms may be obtained.

2. Trapping the Organisms by Means of Filters of Solid
Material — In the earlier work in this direction insoluble filters
were used. Thus Pasteur used an asbestos filter and Petri one
of sand.

Petri's sand-filter is convenient, but it has the objection that
the particles of sand may be mistaken for colonies. The sand,
after trapping the organisms in the same way as for the sugar-
filters described below, is mixed with liquefied gelatine, which is
distributed in Petri-dishes, solidified, incubated, and counted.

At the present day soluble filters are often used, cane-sugar
being the most satisfactory material.

Of apparatus of this kind the Sedgwick-Tncker tube is a con-
venient form. . The tube used consists of a glass cylinder about
12 inches long, part being a wide cylinder, bore about ij inches,
and the rest a fine glass tube, bore about J inch, as shown in
Fig. 15.

The free end of the narrow tube is plugged with cotton-wool;
above this is placed a coil of wire gauze, and above this and
supported by it is the sugar, which occupies the rest of the tube.
The wide free end of the cylinder is partially constricted and
plugged with cotton-wool, the cylinder itself being empty. The
whole is then sterilized in the hot-air sterilizer for three hours at
120° C. It is an advantage, however, to sterilize before the sugar
is put in, at a higher temperature — e.g., 15°° to i6o°C. ; then
put in the previously dried sugar, and sterilize again for three
hours at 120° C.

The apparatus is fixed horizontally in the air which is to
be examined. To draw the air through, the narrow tube is

AIR

143

connected by means of india-rubber tubing to an aspirator.
Just before use the cotton-wool plug in the large
•end is scorched in the flame of a spirit lamp and
removed. Air is then drawn through by means of
the aspirator. At least 40 litres of air should be
aspirated through. The amount of air examined is,
of course, known by the amount of water let out
from the aspirator. When the air has been drawn
through, the cotton-wool plug, well burned in the
flame, is replaced. The rest of the examination is
done in the laboratory. A tubeful of liquefied
gelatine (about 10 to 15 c.c.) is introduced into the
cylindrical part, the cotton-wool plug being momen-
tarily removed for the purpose, and by means of a
sterilized rod the sugar is pushed out into the gelatine
in which it dissolves. After thorough mixture a
roll-tube is made in the ordinary way. The whole
apparatus is then incubated in the cool incubator,
and the colonies which develop are counted and, if
necessary, subcultivated.

This apparatus has all the disadvantages of a
roll-tube — i.e.y it is difficult to examine the colonies
microscopically, and liquefaction of some colonies
may contaminate and spoil others, while only those
organisms which can develop at gelatine temperatures
will grow. Fig I5>

Sedgwick-

FranklancCs Method. — In this method special tubes, tube.
5 inches long and with a diameter of \ inch, are used.
Their form and the arrangement of plugs are sufficiently shown
in the figure (Fig. 16).

The tubes are fitted up and each placed in a separate outer
tube closed by an asbestos plug (as recommended by Haldane).
They are then sterilized by dry heat for three hours at 1.30° C.

When required for use, the tube is quickly removed from its
outer case, connected to the aspirator by the plugged (c) end
and placed in position. The plug a is then withdrawn, and the
aspirator set in action. When the requisite volume of air has

144 AIR

been passed through, the tube is replaced in its sterile box, and
conveyed to the laboratory. A file mark is made across the
centre, and the tube broken in half. The plugs of glass wool
and the sugar are pushed by a sterile wire into a flask containing
10 or 15 c.c. of liquefied nutrient gelatine. The contents are well
mixed, and a roll-tube preparation made by distributing the
gelatine over the wall of the flask.

The plan recommended by Haldane of using flat-bottomed
flasks for the gelatine, and so making an ordinary plate prepara-
tion, is preferable, or still better the nutrient gelatine, after the

-^

^^ te=-3g

r

a
1

b

7\p. 1 6. Frankland's tube.

c

a, plug of glass wool; b, plug of glass wool and finely powdered cane sugar;
cy plug of cotton wool.

admixture of the sugar, may be poured out into Petri-dishes and
solidified in the ordinary way.

In using sugar for air bacteriology, certain precautions must
be attended to. The sugar must be perfectly dry before it is
filled into the tubes, and its sterilization has to be very care-
fully done. If heated to too high a temperature it darkens,
coheres, and is useless. The particles of sugar must be of
suitable size. In practice, the sugar is somewhat coarsely
ground, and then passed through two wire sieves, one fine, the
other coarse mesh. In this way the fine powder and the larger
granules are rejected. The former would clog, and the latter
allow the passage of organisms between the particles.

One great difficulty which the writer has experienced with
sugar-filters is that if the examinations are made upon moist
air, such as sewer air, the water in the air passing through the
filter rapidly moistens the sugar, and makes the particles cohere,
so that after a short time filtration becomes exceedingly slow.

3. Examination of Air for the Presence of P articulate Con-
tamination.— Fliigge and his pupils have shown that minute
particles or droplets are expelled from the upper respiratory

AIR 145

passages into the air during coughing, sneezing and loud speak-
ing, and that the bacteria so expelled are wafted by air currents,
such as exist in ordinary rooms, to distances as far away as
40 feet. These facts were demonstrated by artificially infecting
the mouth with B. prodigiosus and then performing the above-
mentioned acts.

While the bacilli extruded into the air by these agencies are
numerous they settle rapidly. Koeniger, for example, found
that 60 per cent, of the bacteria sprayed out in this way dis-
appeared from the air in ten minutes, while after twenty minutes
less than 10 per cent, of the original number remained. Winslow
and Robinson1 have recently obtained very similar results, and
they remark, " Clearly the mouth spray is a fairly coarse rain
which settles out for the most part in 15 or 20 minutes."

Winslow and Robinson found that after inoculating the
mouth with a rich culture of B. prodigiosus and speaking loudly
and with vigorous enunciation for 15 minutes only seven colonies
of the specific germ could be obtained from 140 litres of air
collected at the close of the speaking from various points in
front of the speaker.

In a number of cases persons suffering from pulmonary
tuberculosis have been experimented with and B. tuberculosis
expelled by them in the act of coughing demonstrated to pene-
trate through the air to a distance of about one metre from the
mouth of the patient. Heymann (1899) succeeded in infecting
six out of 25 guinea-pigs exposed 20 to 45 cm. in front of
coughing consumptives.

Gordon2 has greatly extended our knowledge of such par-
ticulate pollution, and has shown that certain bacteria, which
can be detected and estimated, furnish means whereby these
different kinds of pollution can be recognized.

He has shown that certain streptococci are present in
enormous numbers in human saliva, and that their presence
serves as a means whereby the addition of saliva to air can be
detected.

Using salivary streptococci as the test he confirmed Fliigge's

1 fourn. of Inf. Diseases^ 1910, vol. VII, p. 17.

2 Local Government Board Medical Officers Report, 1902-3, p. 42?.

S. W. 10

146 AIR

results and showed the presence of particles of saliva in air at a
distance of 40 feet in front of the speaker and even at a distance
of 12 feet behind him.

To render this test reliable it must be shown that these
streptococci are not usually present in air. Gordon has care-
fully examined the open air of London, E.G., and at Black-
heath, S.E. He carried out ten observations upon London E.G.
air, 50 litres being examined for each experiment. Streptococci
were obtained from the air on eight of these ten occasions but
only one of these appeared to be the Streptococcus salivarius of
saliva. With Blackheath air streptococci were obtained on only
three occasions, while in each of these experiments as much as
100 litres of air were examined. All three were air streptococci.

His results showed that while streptococci were readily
found in rooms used for speaking and working they were very
rare in open air and that when present the varieties were almost
invariably strains not ordinarily met with in saliva.

According to Gordon, pollution of three separate kinds can
be recognized by bacterial tests.

a. Pollution from Material derived from the Upper Respira-
tory Passages. — The organism specially characteristic of such
pollution is the Streptococcus salivarius.

The method advocated to detect it is to expose Petri-dishes
containing broth to the air. The broth is then incubated ana-
erobically for forty-eight hours at 37° C. By brushing over agar
plates the streptococci are isolated in pure culture, and their
morphological and biological properties determined. Instead
of broth, ordinary agar plates may be employed.

b. Pollution from Material detached from the Skin. — Gordon
has shown1 that the Staphylococcus epidermidis albus is con-
stantly present on the human skin, and that by its detection in
air the presence of particles detached from the skin may be
deduced. It may be detected by the same methods as for the
Streptococcus salivarius.

c. Pollution by Material brought in from the street on
boots. — Such material consists largely of horse-dung, and may

1 ibid. 1904-5, p. 387.

AIR

147

be recognized by the presence of B. coli, spores of B. enteritidis
sporogenes and Streptococcus equinus.

These types of streptococci and staphylococci are mainly
differentiated by their behaviour towards certain sugars and
alcohols. In the following tables the -f sign indicates the pro-
duction of acid, with or without gas. Andrewes and Horder1
have summarized the main types as follows :

£

|

I

I

1

g

-

G

£

$

:s s

c g

No.

Name

^

0

I

c

rt

|

"S

'*

3
«

Cfl

H-5

ti

C/3

5J

o

IJ

,

Streptococcus

Pyogenes . . .

—

—

+

+

—

—

±

—

—

+

Longns

+

2

S. salivarius

+

ok

4-

4-

±

—

—

—

—

±

Brevis

—

3

S. an^inosus

+

±

+

+

-

-

-

-

_

dh

Longtis

+

4

S. faecalis . . .

+

f

+

+

-

-

+

+

+

+

Brevis

-

5

S. ecfttinus ...

—

—

4-

—

—

—

+

-f-

—

+

Brevis

—

6

Pneuniococciis

*

"

+

+

+

*

~

"

"

~

Brevis

+

Andrewes and Gordon2 have similarly summarized the chief
human staphylococci as follows :

B

T3

Name

Character
of Broth
Growth

Pigment
on Agar

A

JE>

D

Liquefactio
of Gelatine

V
"rt

's

V

1

"«

Lactose

Glycerine

'S
c

Patho-
genesis

Staphylococcus
pyogenes

Uniformly
turbid

Orange,
pale

-

+

-

+

+

+

*

t

Highly
patho-

yellow,

genic

or white

Staphylococcus
epiderniidis

Uniformly
turbid

White

+

+

+

+

+

+

+

-

Feebly
patho-

albus

genic

Staphylococcus

Clear with

White

-

-

-

+

+

-

+

-

Not

salivarius

deposit

patho- ;

genie i

Scurf

Inter-

White

-

_

-

-f

-

_

_

+

Not

Staphylococcus

mediate

patho-

clear to

genic

turbid

1 Lancet, 1906, September 15, 22, and 29.

- Local Government Board Medical Officer's Report ', 1905-6, p. 558.

10—2

148 ANTISEPTIC AND GERMICIDAL POWER

CHAPTER X

THE DETERMINATION OF ANTISEPTIC AND GERMICIDAL

POWER

The testing of the germicidal power of a given substance is
simple in principle, but in practice there are many possibilities
of error.

It is obvious that arithmetical statements of germicidal
power are useless unless both the duration of action and the
nature of the material acted upon are specified. In other words,
the three factors — strength of the solution, duration of action,
and nature of the material acted upon — cannot be separated
from one another. To give but one illustration : if we ascertain
that a given strength of mercuric chloride will kill typhoid
bacilli in broth culture in half an hour, it by no means follows —
and, indeed, would be distinctly untrue to state — that such a
strength in the same time would be sufficient to render typhoid
faeces harmless as a possible factor in the spread of enteric
fever.

Laboratory determinations, owing to the rigid control to
which they are susceptible, have great value, but unless their
limitations are recognized, deductions from them to practical
conditions, which perhaps differ considerably in their nature,
may be very erroneous.

It is very difficult to define standard conditions for testing
purposes, because in practice disinfectants are employed as
germicides under a variety of conditions.

Disinfection has been shown by Madsen and Nyman, and
also by H. Chick, to be a process exhibiting many analogies
to a chemical reaction, one reagent being represented by the
bacterium and the second by the disinfectant. When the dis-
infectant is present in considerable excess, the process proceeds
in accordance with a definite law, the number of living bacteria
per unit volume progressively and regularly decreasing with
increase of time in a logarithmic ratio.

ANTISEPTIC AND GERMICIDAL POWER 149

Investigations may be required to : —

(1) Determine the restraining and germicidal power of
different substances in solution.

(2) Determine the germicidal power of substances when
volatilized.

To determine Germicidal Power of Liquids.

Two separate determinations may have to be made, one for
the bacterium and one for the spore, if spores are produced.

In testing germicidal power it is essential that certain factors
be kept constant if results of any value are to be arrived at.
The following may be specially mentioned.

Temperature. Disinfectants act much more powerfully at
higher than at lower temperatures. The relationship is an
orderly one and capable of being mathematically expressed.
Even when phenol controls, under identical temperature con-
ditions, are being employed, as far as possible the same tem-
perature should always be used, or at least only variations
between narrow limits.

Number of bacteria present. Since disinfection is analogous
to a chemical reaction it follows that the greater the number
of bacteria the longer the time taken to disinfect, and this
applies even when the disinfectant is present in great excess.
When comparative experiments are undertaken as nearly as
possible the same amount of bacterial culture must be used for
the control and for the unknown disinfectant.

The resistance of the bacterium used. Not only do different
bacteria differ in their resistance but considerable differences are
met with in different strains of the same bacillus. If controls
are done at the same time with the same strain variations from
this factor can be eliminated.

Culture medium to be used to test viability. This is of some
importance since organisms damaged, but not killed, by the
action of the disinfectant may fail to grow if implanted into a
medium not perfectly suited to their requirements.

ISO ANTISEPTIC AND GERMICIDAL POWER

The vehicle in which the bacteria are suspended. This is of
extreme importance and very greatly influences the result. It
is discussed in greater detail below.

The nature of the bacillus to be employed as the test organism
will obviously depend upon the purposes for which the dis-
infectant is required. For general work B. typhosus is very
frequently employed and is a convenient organism for this
purpose, but care must be taken to use a strain which does not
show pseudo-dumping in ordinary broth.

A number of methods has been used to estimate the ger-
micidal power of liquids of which the thread, garnet and various
drop methods may be mentioned.

Thread method.

In Koch's thread method silk threads are sterilized and then
impregnated with the organism. They are then transferred to
the antiseptic solution under investigation, and left in contact
for a given time. After thorough washing in sterile water, to
remove the antiseptic, the threads are sown on agar or other
suitable nutrient medium, which is then incubated and examined
for growth.

In ascertaining germicidal power it is very important to be
certain that none of the germicide is carried over into the
cultivation solution, as a very small amount may be sufficient to
inhibit growth. In the thread method this danger is not fully
guarded against.

Absence of growth cannot be taken as a certain indication
of the lethal power of the antiseptic so employed, as in practice
it is extremely difficult to get rid of all traces of antiseptic,
and a quantity sufficient to inhibit growth may be left.

Garnet method.

In the garnet method of Kronig and Paul (1897) garnets
of similar size are selected, and, after careful cleaning, are
dipped into a filtered watery emulsion of sporing anthrax or
other bacillus selected. The emulsion is allowed to dry on
them in a thin film. The loaded garnets are then immersed in
the disinfectant solutions under investigation. After definite

ANTISEPTIC AND GERMICIDAL POWER 151

periods of time the garnets are taken out, the disinfectant
carried over removed by gentle washing, and, if necessary, by
washing in solvents (such as ammonium sulphide, if mercuric
chloride is used) to render inert any trace of disinfectant.
The bacteria or spores are then separated from the garnets by
shaking them in water. Definite amounts of the washings are
plated and the bacteria counted.

The garnet method is a valuable one, but on the whole
one or other of the drop methods is the most convenient for
determining the germicidal action of any given substance for
the ordinary bacteria. The following are all of this character.

Rideal- Walker method.

It is convenient to compare the germicidal power with that
of some standard disinfectant. Rideal and Walker in 1903 1
introduced a method by which comparisons with carbolic acid
are made, under definite conditions. Their procedure gives a
good idea of the principles of the drop method.

In the Rideal-Walker method a carefully standardized pure
carbolic acid solution is used as a control, accurate dilutions in
sterile distilled water being prepared. The carbolic acid solution
can be made up from the pure crystals, but they must be dry or
as dry as possible. The solution should be standardized by
means of titration with bromide. Dilutions must be freshly
made each day. A twenty-four hours' broth culture (Lemco),
grown at 37° C, of the B. typhosus or other organism tested, is
used. When B. typhosus is employed a reaction of + 1*5 for the
broth is recommended.

The dilutions of the disinfectant are, in general, made with
distilled water. In this method they are usually placed in test
tubes which for convenience are frequently fitted in special
racks. The dilutions must of course be very carefully prepared,
and to avoid errors in diluting at least 5 c.c. of each strength
should be taken to make the dilution required. All diluting
solutions, instruments, flasks and test tubes must be sterilized
before use.

1 Journal of Sanitary Institute, 1903, vol. xxiv, p. 424.

152

ANTISEPTIC AND GERMICIDAL POWER

To 5 c.c. of a particular dilution of the disinfectant in
sterilized water 5 drops of the broth culture are added. After
shaking, subcultures are taken every two and a half minutes up
to fifteen minutes. Care must be taken to use the same sized
platinum loop so as to remove, as far as possible, the same
amount of disinfectant and the same number of bacilli.

The subcultures are made into broth and incubated for at
least forty-eight hours at 37° C. Those with a growth are then
entered in the tables.

Care must, of course, be taken at every stage to avoid
extraneous bacterial contamination. The purity of the growth
in the positive broth tube from the greatest dilution of the
disinfectant after the longest time should be carefully tested.

A number of different dilutions of the disinfectant under
examination, and also several dilutions of the carbolic acid, are
tested at the same time, and under precisely similar conditions
of temperature, amount of disinfectant solution used, quantity of
typhoid broth culture added, etc. A dilution of the disinfectant
which possesses the same germicidal efficiency as the standard
carbolic acid dilution is obtained and the efficiency of the dis-
infectant is expressed in multiples of carbolic acid performing
the same work. The ratio obtained by dividing the former by
the latter is called the " carbolic acid coefficient." The results
are conveniently recorded in tables of which the following is an
illustration :

B. typhosus Twenty-Four Hours' Broth Culture at
37° C. Room Temperature 15° to 18° C.

Sample

Dilution

Tim
0

•i

e Culture exposed to Action
f Disinfectant in Minutes

Subculture

5

7*

IO

(4

15

Period of
Incubation

Tempera-
ture

Disinfectant w.
Disinfectant w.
Disinfectant w.
Carbolic acid

I : 70
I : 80
I : 90

I : 80

X
X
X
X

X

>;
X
X

Hours
48
48
48
48

Centigrade

£

37°

37°

X
X

X

X

i

Carbolic acid coefficient = £# = 0*87.

ANTISEPTIC AND GERMICIDAL POWER 153

With suitable modifications this method can be used to
obtain the carbolic acid coefficients for other organisms — for
example, B. pestis, Sp. cholerae.

It will be noticed that details are given in regard to
apparently trivial matters. Experience has shown that to
obtain identical and concordant results these small points must
be very carefully attended to and followed.

The Rideal-Walker method has been very extensively used.
It has the great merit of introducing a regular basis of com-
parison with a standard germicide, while, on the other hand, it
suggests a scientific accuracy which is scarcely warranted.

Lancet method.

Considerable modifications in the Rideal-Walker method have
been recommended by the Commissioners appointed by the
Lancet to investigate the question of the Standardization of
Disinfectants1. The modifications and differences they used are
briefly as follows :

B. colt communis was used as the test organism, a 24 hours'
growth at 37° C. in fresh meat broth being employed. All dis-
infectant dilutions were made up with distilled water. The
dilutions were placed in special glass pots instead of test tubes,
those actually used being 2\ inches high and f inch in diameter.
Instead of ordinary platinum loops being used for seeding,
specially constructed platinum spoons were used so that more
fluid could be conveyed. The spoons used took up 0*08 c.c. of
water. It is stated to be important that the spoons should be
withdrawn from the solution at the same speed. Lactose bile-
salt broth was used for the secondary culture tubes. Samples
were removed at intervals of 2\ minutes from each pot up to
15 minutes, and then after 15, 20, 25, and 30 minutes. All the
tests were done between 62° and 67° F.

The carbolic acid coefficient was deduced as follows : " The
figure representing the percentage strength of the weakest lethal
dilution of the carbolic acid control was divided by the figure
representing the percentage strength of the weakest lethal
dilution of the disinfectant being tested. This was done both

1 Lancet, 1909, November 13, 20, and 27.

154 ANTISEPTIC AND GERMICIDAL POWER

at the 2j minutes' line and at the 30 minutes' line, and a mean
of -the resulting figures was taken as the carbolic acid coefficient."

Method of Chick and Martin.

Chick and Martin1 have also modified the method using a
standard time (30 minutes). They describe their procedure as
follows :

"Everything used in the experiment, tubes, pipettes, etc.,
being previously sterilized, a series of tubes containing 5 c.c. of
the disinfectant in different concentrations are placed in a water-
bath at 20° C. When the tubes have taken the temperature of
the bath, they are one after another inoculated with five drops
of 24 hours' culture of B. typhosus from a standard pipette, the
time being registered by a chronograph. Exactly one minute
is allowed to pass between each inoculation. When 30 minutes
have elapsed since the first tube was inoculated, samples in
duplicate are taken from it with a platinum loop (of standard
size) and sown in 10 c.c. glucose broth containing litmus. One
minute later the second tube is sampled and so on. These test
cultures are incubated at 37° C. and always kept four days
under observation."

Preliminary observations are necessary to narrow down the
dilutions likely to be lethal. At the last trial a series of tubes,
containing various strengths of pure phenol are simultaneously
tested.

Modifying influences.

The Rideal-Walker procedure and its modifications furnish
valuable information, but the results obtained cannot be applied
without extensive modification to the practical use of disinfec-
tants. The utility of any disinfectant depends upon a number of
different factors. The most important of these is how far it is
uninfluenced by the presence of organic matter. The efficiency
of some disinfectants is greatly altered by the presence of
organic matter, while for others only a comparatively slight
diminution of power is so caused.

1 Journal of Hygiene , 1908, vol. VIII, p. 654.

ANTISEPTIC AND GERMICIDAL POWER 155

Organic matter may influence the potency of a disinfectant
in a number of ways. In some cases the disinfectant is used up
by acting as an oxidiser to the organic matter, in others the two
form an inert compound, while in a further group of cases the
action of the organic matter would appear to be largely or at
least in part due to its effect upon the emulsification of the
disinfectant, accelerating or retarding it.

Martin and Chick1 showed that the presence of particulate
organic matter, such as animal charcoal or faeces, affected the
germicidal value of emulsified tar acid disinfectants to a much
greater extent than it did phenol solution. By the suitable
addition of such particulate organic matter the whole of the
emulsified tar acid could be removed. When a 3 per cent,
suspension of dried finely-divided faeces is used the efficiency of
phenol is only reduced by about 10 per cent., while that of the
emulsified tar acids is reduced from one-third to one-eleventh of
the primary value. The soluble commercial cresols occupy an
intermediate position, the reduction depending upon the solu-
bility. The reduction in the case of the emulsified tar acids
they found to be higher the finer the emulsion. Some dis-
infectants are more efficient against one species of bacterium,
others against another. In the case of spores metallic salts are
most efficient. The removal of an emulsion of higher phenols
by bacteria is in the first instance a process of adsorption ; dis-
infectants which form fine emulsions possess superior efficiency,
because, owing to this adsorption, the bacteria rapidly become
surrounded by the disinfectant in much greater concentration
than exists throughout the liquid.

The Rideal- Walker method does not take into account the
influence of the presence of organic matter, and various attempts
have been made to obviate this difficulty. For this purpose
the addition of gelatine, serum, urine, milk, faeces, etc., has been
suggested by different workers, so that the germicidal power of
the disinfectant may be tested in the presence of organic matter.
None of these additions are altogether satisfactory, and it cannot
be said that a suitable method has yet been evolved. The effect

1 f our nal of Hygiene, 1908, vol. vin, p. 654^

156 ANTISEPTIC AND GERMICIDAL POWER

of the addition of these organic substances is in every case to
considerably lower the coefficient obtained with what may be
styled the naked germs, but the coefficient of some disinfectants
(for example, potassium permanganate) is lowered to a much
greater degree than others.

Organic matter method of Chick and Martin.

The method adopted by Chick and Martin, in which dried
faeces is used, is a useful one. It is carried out as follows :

The faeces used is dried, first in a water-bath then at 105° C,
ground to a fine powder, and passed through a fine sieve with a
mesh of 130 to the inch. Quantities of 0*15 gramme are weighed
out and placed in test-tubes, to which also are added 2' 5 c.c.
of distilled water. The tubes are sterilized in the autoclave,
covered with indiarubber caps, and stored in covered jars.

At the time of the experiment different amounts of a suitable
dilution of the disinfectant are added to each tube, together with
enough distilled water to make the total volume 5 c.c. The
tubes then contain different concentrations of the disinfectant in
the presence of 3 per cent, faeces.

The tubes are then placed in a water-bath at 20° C. When
the tubes have taken the temperature of the bath, they are one
after another inoculated with five drops of twenty-four hours'
culture of B. typhosus from a standard pipette, the time being
registered by a chronograph. Exactly one minute is allowed
to pass between each inoculation. When thirty minutes have
elapsed since the first tube was inoculated, samples in duplicate
are taken from it with a platinum loop (of standard size), and
sown in 10 c.c. litmus glucose broth. One minute later the
second tube is sampled, and so on. The test cultures are incu-
bated at 37° C., and always kept four days under observation.
For the first testing a wide series of dilutions must be employed.
The second series may be narrowed down, and at the last trial,
which may be the second or third series, a series of tubes
containing various strengths of pure phenol are simultaneously
treated.

ANTISEPTIC AND GERMICIDAL POWER 157

Modifications for special organisms.

The action of antiseptics upon certain special organisms
cannot be tested by the above methods. As a good illustration
of this the determination of the germicidal action upon tubercle
bacilli may be mentioned. The distinction between tubercle
bacilli in a moist state and when dried must always be kept in
mind. It is well known that tubercle bacilli can be dried without
causing their death, and in such a condition are highly resistant,
particularly when enveloped in dried expectoration.

The fresh sputum should be spread upon slips of wood or
other substance, arid dried in a desiccator over sulphuric acid.
Only completely dried slips should be used. The slips are
soaked in different strengths of the germicidal solution under
examination for a definite time (e.g., three hours). The slips are
then washed in sterile water, and the dried expectoration scraped
off, made into an emulsion with sterile water, and injected into
a series of guinea-pigs by the method described in Chapter VI.
If tuberculosis develops it is obvious that all the tubercle bacilli
were not killed. By using an appropriate series of dilutions the
correct lethal strength for the tubercle bacillus, under the con-
ditions of the experiment, can be ascertained.

If the test organism produces spores, it must be incubated
first under optimum conditions for spore formation, and cultures
used which contain large numbers of spores.

Chemical analysis.

It is very desirable that chemical analyses of disinfectants
should be made as well as bacteriological determinations. From
an estimation of the tar acids and other constituents data of
value as regards the probable efficiency of the disinfectant, and
more particularly as to its probable efficiency in the presence of
organic matter, can be obtained.

To Test the Action of Volatile Disinfectants.

Broth cultures of different organisms may be used. For
hygienic purposes B. typhosus, B. diphtheriae, and B. anthracis
are convenient.

158 ANTISEPTIC AND GERMICIDAL POWER

Sterile strips of linen are soaked in these solutions, then
removed and dried at 37° C. in a vacuum over sulphuric acid.
Such inoculated strips are exposed to the action of the gaseous
disinfectant, present in known percentage, for definite but
varying periods.

Some of the strips should be exposed freely to the dis-
infectant, others should be placed in the centre of rolled blankets,
mattresses, etc.

After the required time, the strips are inoculated into sterile
broth tubes, which are incubated and examined for growth.

The dried strips are conveniently carried in sterile Petri-
dishes.

All the different factors, such as the duration of action, the
percentage of disinfectants, the temperature, and the degree of
moisture present, should be carefully noted.

Splinters of wood, paper, wool, and other fabrics may of
course be used with or instead of linen strips.

APPENDIX

It is convenient to collect together in a form handy for
reference the composition of the media required for the exami-
nations described in this work.

While innumerable media have been described by different
investigators those of proved usefulness are not very many. In
the following descriptions only those mentioned in the text are
given.

Reaction Standardization of Media.

In all bacterial enumeration work it is of great importance
to work with media of definite standard reaction. The chemical
reaction of media greatly influences the number of bacteria
which will develop, and unless reasonably uniform conditions are
maintained bacterial counts are valueless. For the standard-
ization of media litmus is much less suitable than phenolphthalein.
Results are expressed in terms of normal acid, or alkali, per

APPENDIX 159

cent: using 4- and - to indicate acid and alkalinity respectively.
Thus a + ro medium indicates that the medium is acid to the
extent of ro c.c. of normal acid per 100 c.c. of the medium.

The reaction of + ro was adopted by the English Committee
on Standard Methods and should be employed as the standard
reaction for all media, the addition of alkali being stopped at
the first appearance of a pink colour.

The actual standardization may be conveniently done as
follows :

Ten c.c. of the medium (gelatine, agar, broth) are pipetted
into an evaporating basin or small beaker containing about
50 c.c. of hot distilled water. Half a c.c. of phenolphthalein
solution (o'5 per cent, solution in 50 per cent, alcohol) is added

N
and the mixture is boiled for several minutes. — sodium hydrate

solution is cautiously run into the beaker from a burette until
the first tinge of pink permanently remains. The amount of
alkali added is then accurately read off.

This is repeated with a fresh 10 c.c. of the medium. Results
not differing by more than O'l c.c. should be obtained. The

N
mean is taken, and from this the amount of — alkali required to

neutralise the whole litre of the medium is calculated.

The neutral point to phenolphthalein gives a medium too
alkaline and a + I per cent, reaction is required. From the
calculated amount of normal alkali required, ro c.c. is deducted
for each 100 c.c. of medium. The calculated amount of alkali,
less this deduction, is then added, and a+ ro per cent, reaction
is obtained. In other words alkali is added insufficient to quite
neutralize the medium to phenolphthalein, and the addition of

N

ro per cent, of — alkali would still be required to make it com-
pletely neutral.

As an example of an actual standardization, the following is
given :

Two separate 10 c.c. of a litre of nutrient agar each required

N
the addition of 1*2 c.c. — sodium hydrate solution to make them

I6O APPENDIX

neutral to phenolphthalein. Therefore 980 c.c. (i.e. 1000—20 c.c.)

will require I — x 980) = 117-6 c.c. — sodium hydrate solution

N
equal to 1 176 c.c. — alkali.

N

One c.c. of — acid is deducted for each 100 c.c. or TO c.c. for
o

N
the litre. Therefore 176 c.c. — alkali is actually added to the

medium and a + ro reaction is obtained.

If the medium is already alkaline to phenolphthalein, the
sample removed for titration should be made acid by the
addition of a definite and accurately pipetted amount of

N N

— H2SO4. The mixture is then boiled and — alkali run in in

the ordinary way. A simple calculation gives the required

N
result. This is preferable to the direct use of — H2SO4, titrating

to the neutral point.

Preparation of Nutrient Media.

Nutrient Broth (Lemco). Lemco broth is slightly less nutrient
than fresh beef broth but is more uniform in quality, cheaper
and more readily made. It is suitable for most public health
purposes.

Weigh out 5 grammes Liebig's extract of beef (Lemco)r
5 grammes of sodium chloride, and 10 grammes of Witte's
peptone, mix with I litre of sterile rain or distilled water in an
enamelled saucepan, and boil for a few minutes. Transfer to a
flask and make up to one litre again. Steam in current steam
for 45 minutes. Estimate the reaction, and bring to a + ro re-
action. Filter into a clean flask, and again steam for 30 minutes.
If then not perfectly clear filter again. Distribute into clean
and preferably sterile test-tubes, about 10 c.c. for each tube,
plug with cotton wool, and sterilize in current steam for three
successive days for half an hour each day.

Nutrient Gelatine. Boil up the constituents of broth as above
in an enamelled saucepan and transfer to a flask, making up to

APPENDIX l6l

I litre. Add 120 grammes of best "gold label" gelatine (i.e. 12
per cent.). Place the flask in the steam sterilizer and steam for
one hour to completely dissolve the gelatine. Estimate the
reaction and add the calculated amount of alkali to bring to
a + ro per cent, reaction. The increased bulk of the gelatine
brings the contents up to 1096 c.c. (Eyre) so the amount of

X

- alkali which has to be added must be estimated on this basis,
o

Cool down to about 50° — 60° C. and add the white of one
egg previously mixed with a little water. Mix well and keep
in current steam for about 30 — 40 minutes. The egg albumin
is coagulated and acts as a mechanical clarifying agent. The
gelatine above the clots should be clear. Filter into a clean
flask, funnel and flask being placed in the steam sterilizer. Put
into tubes (which should have been previously sterilized), and
sterilize for thirty minutes on three successive days.

Sometimes considerable difficulty is experienced in obtaining
gelatine media sterile. If this occurs great care should be taken
to have all flasks, test-tubes, cotton-wool, and other articles
used in the manufacture previously sterilized. It must be
remembered that gelatine media must never be heated above
1 00° C. and should not be heated more than is necessary. For
this and all media it is important to test the sterility of the
medium by incubating the finished media for several days at
37° C.

Nutrient Agar. Weigh out 5 grammes Lemco, 5 grammes
sodium chloride, 10 grammes peptone (Witte's) and heat to
boiling with a litre of distilled or rain water in an enamelled
saucepan. The solution must be distinctly alkaline. Transfer
to a flask and make up to one litre. Add 15 grammes (i.e.
I '5 per cent.) of thread or powdered agar. Digest in the auto-
clave at 115° C. for forty-five minutes. Estimate the reaction
and bring to a + I per cent, reaction. Cool to about 50 — 60° C.
and add the white of an egg previously mixed with a little
distilled water.

Heat in autoclave at H5°C. for forty-five minutes. Filter,
preferably through papier chardin, flask and funnel being placed

s. w. ii

162 APPENDIX

in the steam sterilizer. Distribute into tubes and sterilize once
in the autoclave for thirty minutes at 115° C.

Glycerine Agar. Made like ordinary nutrient agar except
that 6 per cent, of glycerine is added to the medium after
filtration.

Litmus milk. When machine-separated milk can be obtained
fresh it may be used. Frequently however it cannot be obtained
fresh; it is then so heavily bacterially contaminated as to be
unsatisfactory. In this case ordinary milk must be used and the
cream separated in the laboratory centrifuge. It should always
be tested to be certain that it is free from preservatives. Steam
for one hour in the steamer. Remove any coagulum or scum.
Estimate the reaction and bring to a + 1 per cent, reaction.
Add pure litmus to a suitable standard tint. Distribute into
tubes and sterilize in current steam for one hour on three suc-
cessive days.

Peptone water. Boil together 10 grammes peptone (Witte's),
5 grammes sodium chloride and one litre of rain or distilled
water. Filter tube, and sterilize in the autoclave for thirty
minutes at 1 15° C.

Blood Serum. Collect the blood from a slaughter house in
a clean glass or enamelled metal cylinder. Cleanliness must be
exercised, but it is not necessary that special precautions should
be taken to collect the blood under sterile conditions. Remove
to the laboratory with as little agitation of the blood as possible,
and place in the ice chest for 24 hours. Pipette off the separated
serum into a clean flask. Transfer to previously sterilized and
plugged test tubes, adding about 5 to 6 c.c. to each tube. Place
the tubes in the steam sterilizer in a slanting position. With
the square form of sterilizer, which is the most useful shape, this
can be readily done by making a little frame to support them in
this position. Generate a moderate amount of steam, and leave
the cover only loosely fitting, until the serum has quite solidified.
This takes from one to three hours and must not be hurried.
Then fit on the cover tightly and subject to current steam at
iooc C. for one hour. Sterilize again for the two following days
at 1 00° C. for thirty minutes each day.

APPENDIX 163

It is essential that the solidification takes place below 100° C.
If steam is generated in large amount with air exclusion so that
the medium is heated to 100° C. (or only just below) before com-
plete solidification takes place, bubbles and cavities in the serum
will be formed and the medium spoilt.

Luffler's Blood serum. This consists of three parts of blood
serum mixed with one part of I per cent, glucose broth. Other-
wise it is prepared as above. It is preferred by the writer to
ordinary serum for most purposes.

Egg medium (for B. tuberculosis). Fresh eggs are washed
and then partially sterilized by dipping them, held in forceps, in
boiling water for about half a minute. They are opened with
aseptic precautions, and the contents poured into a sterile flask,
to which normal saline is added in the proportion of one part
to two parts of egg. The eggs and saline solution are then
thoroughly mixed in the flask, which should be of large size.
The medium is strained through muslin to remove air-bubbles,
and poured as quietly as possible into a flask with a side tube
near the bottom. The medium is added from this flask to
sterile test-tubes, when the latter are in a nearly horizontal
position, sufficient being added to make a good slope. Care
must be taken to avoid soiling the other parts of the test-tube.
The tubed medium is inspissated at So0 C. on two successive
days.

Sugar media for fermentation testing. The power of bacteria
to ferment sugars, alcohols, etc. is best tested in double tubes
(Durham's tubes). The media in these tubes, after three days'
sterilization, should completely fill the inner tube. The stock
medium to which the different sugars are added is either ordinary
nutrient broth without the sodium chloride or without, in addition,
the Lemco (i.e. simply peptone water). The solution should be
faintly alkaline to litmus.

This medium is made and sterilized in bulk and to appro-
priate quantities in a flask, are added, as required, the particular
sugar or alcohol, in amount to make 0*5 per cent in the finished
medium, and sufficient pure litmus solution to give a blue tint.

II— 2

164 APPENDIX

The mixture is tubed and sterilized in current steam, twenty
minutes each day.

To avoid decomposition of the sugar-alcohols such media
must never be heated above 100° C. and should be heated as
little as possible.

Fresh-meat infusions must not be used in the preparation of
sugar media, since they usually contain inosite, etc. Great care
must be taken to ensure the purity of the substances to be tested
for fermentation.

Sugar-alcohol media for the differentiation of streptococci. A
stock solution is made up containing Lemco 10 grammes, peptone
10 grammes, sodium bicarbonate I gramme, 10 per cent, aqueous
litmus-solution 100 c.c., distilled water to I litre. This is boiled
and filtered in the ordinary way. One per cent, of the sugar,
alcohol or glucoside is added to portions of this stock solution
to make the different media. The tubes are sterilized in current
steam for half an hour for three successive days.

The sugar-alcohol substances recommended by Gordon are
lactose, saccharose, salicin, mannite, raffinose and inulin.

No double tubes are required since the presence of acid only,
not gas and acid, is recorded. To enable this to be accurately
done it is important that the same colour tint with litmus should
be produced for each batch. It is also convenient to use a sterile
control tube when comparing colour production.

Lactose Bile Salt Broth.

Sodium taurocholate ... ... 5 grammes.

Lactose ... ... ... ... 5 „

Peptone ... ... ... ... 20 „

Water 1000 c.c.

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

The sodium taurocholate prevents the growth of many
saprophytic bacteria.

APPENDIX 165

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

Aescidin Agar (Harrison and Leek1): — 10 grammes Witte's
peptone, 5 grammes sodium taurocholate (commercial), I gramme
aesculin, 0*5 gramme ferric citrate, 15 grammes agar, tap-water
I litre. The agar and other ingredients are dissolved in the
ordinary way, boiled, filtered, tubed, and sterilized.

Colonies of B. coli in this medium are black with a black halo,
and can be readily counted against a suitable background. The
aesculin (a glucoside) combines with the iron citrate and forms
a dark-brown salt, the reaction only taking place in sugar-free
media.

The colonies of some other organisms give the reaction,
notably B. lactis aerogenes.

Neutral Red Glucose Broth. — To ordinary broth made from
Liebig's Extract, peptone and sodium chloride, and made faintly
alkaline to litmus, 0*5 per cent, of glucose and 12 c.c. per litre
of a O'5 per cent, freshly-made watery solution of neutral red
(Grubler's) are added. The solution, after preliminary steaming,
is filtered and tubed — ordinary or double tubes — and sterilized
for three successive days for thirty minutes each day.

Malachite Green Agar (as prepared by Lentz and Tietz).
Three pounds of fat-free ox flesh are finely cut up and macerated
with 2 litres of water for sixteen hours. The extract is expressed
boiled for half an hour, filtered, 3 per cent, agar added, and the
mixture boiled for three hours. Then are added I per cent,
peptone, 0*5 per cent, sodium chloride, and I per cent, nutrose in
240 c.c. of cold water (the nutrose may be omitted). This
mixture is brought to the litmus neutral point by soda solution
with duplitest paper, then boiled for one hour and filtered through
linen. The reaction of the finished agar is sometimes distinctly
acid. It is filtered into small flasks of 100 to 200 c.c., and
sterilized three times before use in the usual way. Before the
addition of the malachite green, the hot agar is tested by

1 Centralbl. f. Bakt. n. Abt. 1909, xxn, p. 55.

166 APPENDIX

duplitest paper, and made alkaline with sterile soda solution
until the red slip is red-violet. To 100 c.c. of the hot agar
I c.c. of a I : 60 solution of malachite green in distilled water
(the solution keeps good for ten days) is added — i.e., the agar
contains I : 6000.

The finished agar is poured at once into Petri-dishes in
layers 2 millimetres thick. The dishes are well dried, and
can be kept in the ice-chest.

By this strength of malachite green the growth of most
kinds of B. coli, as well as of many alkali-forming organisms,
is greatly diminished. The B. typhosus colonies are also retarded,
but can be recognised, the size of a grain of sand, with the
naked eye after twenty-four hours ; after a longer period (two to
four days) larger, better developed colonies appear, which colour
the agar yellow.

Dulcite Malachite Green Broth. — Liebig's Extract, 10 grammes;
peptone, 10 grammes ; sodium chloride, 5 grammes, are boiled
up with a litre of distilled water. The mixture, after filtration,
is made up accurately to a 4- I per cent, reaction, and 5 grammes
of dulcite are added ; O'5 gramme of powdered malachite green
is very accurately weighed out, and also added. The mixture,
usually slightly turbid, is steamed for thirty minutes, and again
filtered. It is tubed, 10 c.c. into each tube, and sterilized for
thirty minutes on two successive days.

Brilliant Green Agar. Fawcus' modification. — Conradi's
brilliant green agar has been modified by Fawcus1 by the
addition of lactose and bile-salt and by increasing the percentage
of brilliant green. Prepared as follows : —

To 900 c.c. of tap-water add 5 grammes sodium taurocholate,
30 grammes powdered agar, 20 grammes Witte's peptone and
5 grammes sodium chloride. Dissolve in the steam sterilizer
for three hours. Clear with white of egg, filter through wadding
and bring to a reaction of -H 1*5 per cent.

Dissolve 10 grammes of lactose in 100 c.c. of distilled water
and add to the melted agar. Mix well and filter through
Chardin paper. To each 100 c.c. of the clear bile-salt lactose

1 Journ. Royal Army Med. Corps, 1909, XII, p. 147.

APPENDIX 167

agar add 2 c.c. of a O'l per cent, watery solution of brilliant
green (extra pure) and 2 c.c. of a ro per cent, watery solution of
picric acid. The resulting clear bright green agar is poured
without further heating into Petri-dishes. After solidification
the plates are dried uncovered and upside down in the
incubator and after two to three hours are ready for use. The
medium should not be kept in flasks ready made. The most
convenient plan is to distribute the bile-salt lactose agar into
flasks about 150 c.c. into each. When required for use one
of these is melted and 3 c.c. of each of the solutions of the
brilliant green and the picric acid added and well mixed. The
finished agar contains O'$ per cent, bile-salt, I in 50,000 brilliant
green and I in 5000 picric acid.

Colonies of B. typhosus and Gaertner group bacilli are round
and quite transparent. B. coli colonies have a dark green
opaque spot in the centre.

Neutral Red Lactose Bile Salt Agar (L.B.A.}.— Sodium
taurocholate 5 grammes, Witte's peptone 20 grammes, and
distilled water I litre, are boiled up together, 20 grammes of agar
are added and dissolved in the solution in the autoclave in the
ordinary way. The medium is cleared with white of egg and
filtered. After filtration, 10 grammes of lactose and 5 c.c. of
recently prepared one per cent, neutral red solution are added.
The medium is then tubed and sterilized for 15 minutes on three
successive days.

L.B.A. Crystal Violet. — Some workers use L.B.A. to which is
added crystal violet (i in 100,000). The crystal violet acts also
as an inhibiting agent.

Fuchsin Agar. — Introduced by Endo. The following modified
method of preparation has been found by the writer more satis-
factory and uniform than the original :

Peptone, 10 grammes ; Liebig's extract of beef, 10 grammes ;
sodium chloride, 5 grammes, are boiled up in an enamelled
dish with i litre of distilled water. The mixture is then poured
into a flask, 30 grammes of powdered agar added, and the whole
heated in the autoclave at 115° C. for one hour. The flask
is removed, and, after cooling to about 60° C., the white of one

168 APPENDIX

egg mixed with a little distilled water is added. The contents
are coagulated by heating in current steam in the usual way,
filtered, and the filtrate made up to I litre. The mixture is
made neutral, litmus paper being used as the indicator. Then
19 c.c. of normal sodium carbonate solution and 10 grammes
of chemically pure lactose are added. The flask is replaced for
thirty minutes in the steam sterilizer. Almost invariably there
is a considerable precipitate, and the mixture has to be again
filtered.

Seven c.c. of the fuchsin solution (see below) are added,
followed by 25 c.c. of a quite freshly prepared 10 per cent, sodium
sulphite solution. The mixture becomes much less red, but
is not immediately decolorized. It is then tubed, conveniently
into small flasks, each containing 50 to 60 c.c. of media, and
sterilized in current steam for two days, thirty minutes each day.

Thefuchsm solution is made as follows :

Three grammes of powdered crystalline fuchsin are placed
in a dry flask, and 60 c.c. of absolute alcohol are added. The
contents are thoroughly well mixed, and the flask, tightly
stoppered, allowed to stand for exactly twenty-four hours at
20° to 22° C. The alcoholic extract is then decanted and
preserved in a clean glass-stoppered bottle. Made in this way a
uniform fuchsin extract is obtained which keeps well, and the
same quantity of fuchsin is added each time a fresh batch of
medium is prepared, a matter of much importance.

The medium must be stored in the dark, since light gradually
turns it red. When solidified it is almost free from colour.

B. coli colonies are bright red, round, and have prominent
margins ; B. typhosus colonies are round, colourless, very trans-
parent, and have thin margins.

Drigalski-Conradi Agar. — To 3 pounds of finely-cut-up beef
or horseflesh add 2 litres of water. Allow the mixture to stand
until next day. Boil the expressed meat-juice for one hour and
filter ; add 20 grammes peptone sicca (Witte), 20 grammes
nutrose, 10 grammes sodium chloride ; boil the whole again for
one hour and then filter. Add 70 grammes bar agar, boil for
three hours (or one hour in the autoclave), render slightly

APPENDIX 169

alkaline (using as indicator litmus paper), filter, boil for half an
hour. Add 260 c.c. litmus solution (Kubel and Tiemann), and
boil for ten minutes ; add 30 grammes of chemically pure milk-
sugar, and boil for fifteen minutes. Add the hot litmus milk-
sugar solution to the liquid agar solution (cooled to 60° C.) ;
shake well, and make faintly alkaline ; then add 4 c.c. of a
hot sterile solution of 10 per cent, water- free soda and 20 c.c. of a
freshly prepared solution of 0*1 gramme crystal violet (B. Hb'chsf)
in 100 c.c. warm sterile distilled water. The result is a meat-
water peptone nutrose agar containing 13 per cent, litmus and
O'Oi per 1000 crystal violet. The medium can be kept in tubes
or in small flasks containing enough for three or four plates. It
is sufficient to sterilize once in current steam for thirty minutes.

After the plates are inoculated they should be thoroughly
dried uncovered, either in the laboratory or preferably in the
incubator. They are then covered, inverted, and incubated.

This medium is chiefly used in the isolation of B. typhosus
and in particular to differentiate this bacillus from B. coli and
other organisms. After sixteen to twenty-four hours at 37° C.
the colonies can be distinguished from one another. The B. coli
colonies are red, not transparent, and have a diameter of 2 to 6
millimetres, but considerable variation in size and degree of
colour are met with. The B. typhosus colonies are blue, with a
violet tinge ; they are transparent and resemble dewdrops, and
have a diameter of I to 3 millimetres, seldom larger.

Drigalski-Conradi medium is rather a trouble to prepare, and
is not always satisfactory in use.

Dieudonnes Alkaline Blood agar. Prepared as follows :

Equal parts of normal caustic potash solution and defibrinated
ox blood are mixed and sterilized in the autoclave (Solution A).
Nutrient agar of ordinary composition but exactly neutral to
litmus is prepared (Solution B). Seven parts of B are mixed
with three parts of A and poured into Petri-dishes.

When the mixture of blood and alkali is heated a part of the
latter is absorbed, but the final agar still preserves a very strong
alkalinity corresponding to about 0*6 per cent, of potassium
hydrate. The free alkali and the alkaline combinations formed

I/O APPENDIX

during the heating of the blood together give to the medium
some special qualities.

The plates ought not to be used immediately after their
preparation. They should be kept either for several days in the
incubator at 37° C. uncovered and face down or for 48 hours at
laboratory temperature.

This medium is useful for the isolation of the cholera
spirillum (see Chapter III).

ADDENDUM

The first edition was only issued some two years ago, and
no material advances have been made since that date in the
branches of Public Health Bacteriology treated in this volume.
It is unnecessary therefore to completely revise the text, more
particularly as none of the recent work invalidates any of
the methods and conclusions given.

A number of different investigations have been published
which extend and confirm the conclusions arrived at. In this
addendum only those advances and new methods which
facilitate the examination of the substances under considera-
tion have been selected and summarised. A number of other
procedures have been advocated, but they have not as yet been
sufficiently controlled to include them as methods of approved
value and utility.

i. The Differentiation of the Streptococci.

The organisms included under the term streptococci com-
prise such a large class, are responsible for so great an
amount of human disease, and are so widely diffused in nature
under saprophytic conditions that it is to be anticipated that
repeated attempts will be made to arrive at some trustworthy
methods for their differentiation. Such differentiation is as
essential from the Public Health standpoint as it is from the
more limited pathological aspect, since accurate means of
separating streptococci of human from those of animal origin,
and the saprophytic from the pathological types, would be of
the greatest service.

1/2 ADDENDUM

A number of investigations with these objects have recently
been carried out, particularly in America, mostly since the first
edition was written, and it may be of assistance to discuss,
rather more in detail than in the text, how far these investiga-
tions have yielded results capable of practical application.

From this point of view these tests may be classified into
three groups — haemolytic and virulence tests ; fermentation
reactions ; other tests.

I. Haemolytic and virulence tests. The haemolytic test for
differentiating streptococci was first introduced by Schotmiiller.
He used a blood-agar medium and separated the streptococci
by this means into three groups, i.e. :

(1) Streptococcus pyogenes vel erysipelatos. A long chain
form with greyish colonies and with a zone of haemolysis.

(2) Streptococcus mitior. A short chain form with greenish
colonies producing very slight haemolysis.

(3) Streptococcus mucosus. Capsulated organisms with
colonies of a mucous, slimy consistency. No haemolysis.

Two distinct methods have been used to carry out this
test.

(a) Blood-agar plates. Schotmiiller used a medium con-
sisting of two parts sterile human blood mixed with five parts
of ordinary agar. Other observers have used animal blood.
Stowell, Hilliard and Schlesinger \ for example, added a few
drops of fresh sterile rabbit's blood to cooled melted agar tubes,
the contents after mixing being poured into Petri dishes in
the ordinary way. Haemolysis is tested by streaking the plates
from a young culture.

Other workers have used ordinary blood-agar medium, the
blood being smeared over the surface of the agar. This method
was employed by Hopkins and Lang, but has the drawback
that the degree of haemolysis in part depends upon the
thickness of the blood layer.

1 To save repetition references are given at the end of this section.

ADDENDUM 1/3

(b) Use of liquid blood media. A more severe test of
haemolysis is to test with red corpuscles in solution. The
method employed by Lyall and also by North, White and
A very was to add a definite amount (0-5 c.c.) of an 1 8 hours'
ascitic broth culture (ascitic fluid I part, peptone broth 5 parts)
to i c.c. of a 5 per cent, solution of washed sheep's red blood
cells and incubate in a waterbath at 37° C. for I hour (Lyall) or
2 hours (North, White and Avery).

The blood-agar plate method is the simpler and the one
usually employed, but with its use there is some danger of
obtaining indefinite reactions. As Davis points out many
streptococci may produce a narrow, greenish, greyish or
brownish zone on which, especially after 2 or more days, at
times some clearing of the media may occur. Such strains
do not produce true haemolysis.

The haemolytic test is of value since it is closely related to
virulence. Most of the pathogenic streptococcus strains isolated
from human cases of disease have been haemolytic. Apart
from this property it does not serve to distinguish human from
streptococci of animal origin. Ruediger strongly advances the
view that the ordinary milk streptococci can be distinguished
from the pathogenic types by the haemolytic tests, but the facts
in his 1912 paper in which he advances this view, while not
contradicting it do not establish any such connection.

M'Leod in a later paper concludes that for streptococci
virulence and the possession of haemolytic power are closely
allied, if the organism is growing in the body or under
cultural conditions closely resembling those met with in the
body.

Davis advances the view that the haemolytic test is of great
value in the differentiation of streptococci causing outbreaks of
sore throat, and points out that all the recent epidemics of sore
throat spread by milk in which this point has been tested have
been caused by streptococci of the haemolytic variety. His
suggestion that the types of bovine mastitis due to haemolytic
streptococci are those pathogenic to man, while the non-
haemolytic strains do not cause human disease, is very in-
teresting but unproved and confirmation is desirable.

174 ADDENDUM

In using the haemolysis test fresh cultures must be used.
Jupille, for example, found that this property is a transient one
which depends upon the age of the cultures and is usually
lost in cultures five days old ; while M'Leod found that the
haemolytic property reached a maximum of activity within 17
hours and remained constant up to 48 hours, after which it
gradually disappeared.

The extent of the haemolytic action cannot be used as a
basis of classification as it depends (as M'Leod shows) upon the
extent to which the medium favours the active and prolonged
multiplication of the streptococcus.

Broadhurst (1915) found that haemolysis does not seem to
be correlated with the results in litmus milk, gelatine or with
any of the Gordon reactions. Like other workers she found that
samples from pathological sources yielded the greatest propor-
tion of haemolyzing strains. Such strains may occur in throat
samples and in those from the alimentary canal. A markedly
high percentage (32 per cent.) was found in 101 strains isolated
from the stomachs and intestines of eight dogs.

2. Fermentation reactions.- Two procedures have been
used to determine the capacity of streptococci to produce acid
from sugars, alcohols and allied substances. The original
method introduced by Gordon and employed by most workers
in this country is to qualitatively record the production of acid
as shown by the action on litmus added to the culture medium.
This procedure is described on page 8.

The alternative method as used by Winslow and Palmer is
quantitative, the amount of acid produced being estimated by
direct titration, using phenol-phthalein as the indicator. Some
of the results obtained by Winslow and Palmer with this method
are given on pages 8 and 9.

A considerable number of investigations have been recently
carried out either by this new modification or by the original
procedure, and the results obtained are interesting and valuable
as a basis for further investigation, but it cannot be said that any
very decided and constant differences have been established. In
other words the fermentation reactions of any given streptococcus

ADDENDUM

cannot establish its origin, although they do enable deductions
as to its origin to be made with some degree of probability. The
varying methods of recording make it impracticable to tabulate
all the results together and draw general deductions as to their
relative grouping. Some special points and conclusions may
however be mentioned.

The results of Stowell, Hilliard and Schlesinger are of con-
siderable interest. They studied 240 strains from milk and
from the normal human throat using the titration method.
They draw attention to the value of testing the fermentation
abilities of streptococci at low as well as high temperatures.
They suggest that six different tubes of media will differentiate
any streptococcus between these two sources, i.e. glucose,
lactose, raffinose and salicin at 37° C, and lactose and sac-
charose at 20° C. Four of these tests would be considered
diagnostic.

They believe that the following features are sufficient to
separate the milk from the throat streptococci. The former yield
over 2'5 per cent, acid in lactose and saccharose at 37° C., seldom
ferment a substance higher than saccharose in the metabolic
series, readily grow at 20° C. in glucose, lactose and saccharose
media, while on the other hand the throat streptococci seldom
yield over 2*5 per cent, acid in any substance at any tempera-
ture, over 40 per cent, yield more than I *2 per cent, of acid in
either salicin or raffinose or in both at 37° C. ; at 20° C. they
almost never attack any of the test substances.

Fuller and Armstrong, using the titration method, investi-
gated the biological characters of streptococci from animal and
human excreta. They found that the streptococci of human
faeces were characterised by high acidity in glucose, lactose
and mannite, a low acidity in raffinose and a relatively low
acidity in glucose, lactose and especially mannite. Many
individual variations were however met with.

Houston has recently re-investigated this question, using the
American titration method. He found it very difficult to isolate
lactose fermenting streptococci from the lower animals except
the dog, whereas the great majority of human faecal streptococci
yield streptococci which ferment lactose. He adds " so few

I ? ADDENDUM

streptococci derived from the horse, sheep, rabbit (wild) and
gull were lactose positive that these animals might almost
be excluded from further consideration." A gross total of 296
subcultures yielded only four positives. As regards any satis-
factory differentiation Houston remarks that the results were
disappointing.

Broadhurst studied 100 milk streptococci all obtained from
mixed milk samples. These strains fell into 20 groups classified
on the basis of their ferment activities. As regards the extent
to which the individual substances were fermented she obtained
the following results : salicin 82, lactose 76, saccharose 66,
inulin 38, raffinose 13, mannite 27 per cent. She concluded
that the milk streptococci are characterised by unusually high
fermentative powers, and that they resemble human rather than
bovine strains, and show practically no resemblance to the
equine strains.

Broadhurst in a later paper studied the characters of 767
strains, all but 23 being freshly isolated for the purpose. Her
careful and prolonged investigations must be studied in the
original and cannot be conveniently summarised, but in general
her work affords confirmation of the view that the differentiating
characters employed are insufficient to indicate the origin of a
particular strain and that great caution must be used in using them
for direct sanitary application. Interesting points amongst her
results were that by all methods human throat strains practically
failed to ferment mannite while in human faeces mannite strains
were common. Rafifinose fermenters were more prominent in
bovine faeces than in the faeces of other animals, while they
were strikingly lacking in milk. Many of the strains from
pathological sources failed to ferment raffinose and mannite.

The separate papers must be studied to see the actual
fermentation properties of the individual streptococci. Lactose,
saccharose, raffinose, mannite, inulin and salicin would all appear
to have some differentiating value, but some observers would
omit saccharose as fermented by nearly all strains, others
mannite as fermented by too few, while Houston in his last
investigation found salicin of no value in his attempt to
differentiate human and animal excretal streptococci.

ADDENDUM 177

The titration method — the estimation of the amount of acid
produced — takes a good deal more time than the simpler pro-
cedure of recording the qualitative production of acid. A
careful study of the results obtained does not establish that any
additional advantages are gained by its employment. The fact
that acid is produced shows that the sugar, alcohol or glucoside,
has been split up, but the amount of acid produced largely
depends upon the suitability of the medium as a nidus for the
multiplication of the streptococcus strain tested. Hopkins and
Lang using the titration method found that the fermentation by
a given streptococcus ceases where a certain acidity is reached,
irrespective of how much acidity might be formed by the
carbohydrate decomposition. They found that the acidity pro-
duced by the same strains in the same medium varied, but were
unable to discover the reasons for this variation.

Hopkins and Lang also tested 24 of their strains at
intervals of 8 to 34 weeks and found the reactions qualitatively
unchanged.

The constancy of these fermentation and the other cultural
characters have been investigated by a number of workers.
With reasonably comparable technique the amount of constancy
has usually been high although considerable variations may be
induced by special methods. Broadhurst (1915) has investigated
this question very extensively and found that, for the most part,
inconstancy of characters was not marked. In one series of
tests with 134 cultures of various types and from a varfety of
sources and retested after considerable intervals upon the usual
Gordon media, the three and four months retests gave 63 and
64 per cent, of constancy results respectively. As might be
anticipated newly acquired fermentation powers were more
variable than those originally possessed. Under ordinary
laboratory conditions there was a greater tendency to gain
rather than to lose fermentation powers. Twenty-nine per
cent, gained the power of fermenting one or more substances
while 19 per cent, lost powers of fermentation. A strain that
lost a power usually regained that power later while one
which gained usually retained it. Broadhurst found that
morphological characters varied considerably with the medium

s. w. 12

178 ADDENDUM

employed but that such induced variations were largely tem-
porary.

3. Other tests. Other tests which have been employed
to differentiate streptococci include morphological characters,
particularly the length of the chains, characters of the growth in
litmus milk and the characteristic methods of growth upon solid
media.

All recent work is in the direction of showing that little
value can be attached to these tests. It should however be
mentioned that Crowe has recently demonstrated that charac-
teristic colonies of value for classification purposes are obtained
by the use of a neutral red egg medium.

Agglutination reactions have been employed by some
workers but so far have not yielded results of special value.
Kligler has recently re-investigated this property for strepto-
cocci using sera obtained from four types as differentiated by
their haemolytic and fermentation characters. He found con-
siderable correlation between agglutination and the fermentation
characters employed since in general the serum of one fermenta-
tive type was capable of agglutinating only strains of that
particular group and not the others. The correlation of
agglutination with haemolysis was less in evidence.

References.

J. BROADHURST, Journ. of Infect. Diseases, 1912, x, p. 272

J. BROADHURST, Journ. of Infect. Diseases, 1915, xvn, p. 277.

H. W. CROWE, Proc. Royal Soc. of Medicine, 1913, VI, p. 119.

D. J. DAVIS, Journ. of Infect. Diseases, 1914, XV, p. 378.

C. A. FULLER and V. A. ARMSTRONG, Journ. of Infect. Diseases, 1913, xm,

p. 442.

J. G. HOPKTNS and A. LANG, Journ. of Infect. Diseases, 1914, xv, p. 63.
A. C. HOUSTON, Metropolitan Water Board : Tenth Research Report, June,

1914, p. 22.

F. JUPILLE, Ann. de Flnstitut Pasteur, 1911, xxv, p. 918.
I. J. KLIGLER, Journ. of Infect. Diseases, 1915, xvi, p. 327.
H. W. LYALL, Journ. of Med. Research, 1914, xxx, p. 487.

ADDENDUM 179

J. \V. M'LEOD,yiwr». of Path, and Bact., 1912, XVI, p. 321.
J. W. M'L,EOD,/0wr;2. of Path, and Bact., 1915, XIX, p. 392.
C. E. NORTH, B. WHITE and O. T. AvERY,/0«r». of Infect. Diseases, 1914

xiv, p. 124.

E. C. ROSENOW, Journ. of Infect. Diseases, 1912, XI, p. 338.
C. F. RUEDIGER, Amer. Journ. of Public Health, 1912, II, p. 107.
STOWELL, MILLIARD and SCHLESINGER, Journ. of Infect. Diseases, 1913,

xn, p. 144.

2. Water. Report of Second Committee on tJie Standardization
of Methods for the Bacterioscopic Examination of Water.

Both Committees were appointed by the Royal Institute
of Public Health. The first reported in 1904 and the second
in 1914. Dr Houston was the Chairman of the second Com-
mittee.

The procedures recommended do not differ essentially from
those advocated in Chapter III except in minor details. Certain
procedures in regard to which there is considerable variation
of procedure amongst bacteriologists at the present time may
be quoted in extenso.

The Committee recommended that in all cases the number
of bacteria present in the water should be estimated, and the
number of Bacillus colt. They recommended that as many as
possible of the following tests should be employed :

"(i) Number of microbes per cubic centimetre in :

(a) Meat broth gelatine (or, when the climatic con-
ditions render it necessary, meat broth agar),
incubated at 20° to 22° C.

(b) Meat broth agar incubated at 37° C.

Brand's Essence of Beef or Lemco may, if desired,
be substituted for meat broth in the preparation of
these media (a) and (b).

(c) Lactose bile salt agar (MacConkey). Incubated at
37° C.

The medium may be tinted either with litmus or
with neutral red, according to taste.

12—2

l8o ADDENDUM

" (2) Bacillus coli test.

" In certain cases it is advisable to apply also :

"(3) Streptococcus test.

" (4) Anaerobic spore test (commonly known in England as
the B. enteritidis sporogenes test).

" (5) Tests for pathogenic microbes.

" The composition of the various media recommended by the
Committee, in connection with these tests, is given in an accom-
panying appendix.

"Tests (i) (c\ (3), (4) and (5) need be employed in special
cases only."

As regards the time of counting for the enumeration plates
the recommendation of the Committee was :

" As a matter of convenience it is a common practice to
count gelatine plates after seventy-two hours. The Committee
recognize that a considerable number of bacteria are thus missed,
and that four days is a better period. In any case the day on
which the count is made should be stated.

" It is convenient to count the agar and bile-salt agar plates
at the end of 24 hours, as this enables the observer to arrive
rapidly at certain conclusions, and, if the results are decidedly
unsatisfactory, to issue a tentative warning. This is a matter of
considerable importance in all cases, and in some may be
vitally important. In the agar plates the sporing bacteria are
apt, in forty-eight hours, to spread over the surface of the plate and
render counting difficult or impossible. This does not occur in
the case of the bile-salt agar plates, but it is convenient to count
both sets of plates on the same day."

Bacillus coli. The Committee recommended the use of bile-
salt peptone for the primary test, while as regards the amount
of water to examine "the aim should be to have such a range of
amounts as will always include a positive and negative result."

Isolation of Bacillus coli.

"If the primary bile-salt cultures show gas formation the
differential medium with which the worker has most experience
can be used for plating purposes.

ADDENDUM l8l

"Identification Tests for B. coli. Several suspicious colonies
should be picked off the plates and tested for gas formation in a
lactose peptone medium, and for indol in a peptone medium.
The paradimethylamidobenzaldehyde test is the most delicate
test for indol production.

"The Committee suggest that all reports should contain a
simple statement as to the presence (or absence) of ' lactose 4-
indol +' microbes and the smallest amounts of water in
which they were present, and the largest amount of water from
which they were absent.

" This is common ground for practically all bacteriologists.

" Thus, for example, the simple statement that B. coli
(lactose + indol + ) was present in 100 but not in 10 c.c. of water
avoids raising the many points concerning which there is so
much difference of opinion.

"There are many other tests which it is of advantage to use ;
for example, gelatine for absence of liquefaction, tests for gas
formation in glucose, saccharose, dulcite (dulcitol) and adonite
(adonitol) peptone media, and for acid clotting in milk cultures.
The results thus obtained may be used for classification purposes
or merely registered in the report.

"On the. whole, the Committee consider that judgment as to
the purity of a water should be based primarily on the quantita-
tive estimation of ' lactose + indol + ' microbes, although this
need not prevent the utilization of the results of other tests as
additional factors bearing on the position.

"As regards 'glucose fermenting coli-like microbes' which
yield negative results with one or other, or both the lactose and
indol tests, it is difficult to speak with any degree of certainty.
Bacteriologists are unfortunately frequently called upon to
decide upon individual samples, and it is not necessarily safe to
pronounce favourably on a supply of water because a particular
sample contains no ' lactose 4- indol + ' microbes if it contains
* glucose + ' microbes. Experience has shown that if a supply
yields ' glucose + ' microbes it will almost inevitably be found
sooner or later to contain * lactose -f indol -f ' microbes as
well.

1 82 ADDENDUM

Streptococci^,

" For direct culture the Committee favour the Drigalski and
Conradi medium, but only a decidedly impure water would
contain streptococci in the maximum amount (i c.c.) which can
be spread over a plate. Hence for direct cultures concentration
methods appear to be necessary.

"As regards indirect liquid cultures, glucose formate peptone
broth is a suitable medium incubated anaerobically. If it be
merely desired to determine the presence or absence of strepto-
cocci, microscopic examination of the sediment will suffice, but
the Committee recommend the isolation and study of the
streptococci by secondary cultures on Drigalski's and Conradi's
medium.

"The Committee consider that it is convenient to exclude all
streptococci which do not yield acid in a lactose medium,
because the great majority of human faecal streptococci attack
lactose, and streptococci of little or no * water significance ' are
excluded in this manner.

" The Committee are of opinion that a water should not be
condemned on the streptococcus test alone, if the B. coli and
other results are quite satisfactory."

Standards. The Committee did not go into this difficult
question in detail, but reported as follows :

" The Committee do not think it is practicable to lay down
any fixed standards to govern all cases. Speaking generally, the
Committee consider that too much stress should not be laid on
the number of microbes present in a water, unless the B. coli
tests yield confirmatory results.

" A good water should not contain any B. coli in 100 c.c.,
but a water containing B. coli in 100 c.c. should not necessarily
be objected to without the examination of further samples.

" Experience has shown that even initially impure waters

1 Although the writer signed the report as a member of the Committee it should
be added, to prevent misunderstanding, that the method for the enumeration of
streptococci in water described in the text (page 37) which is not included by the
Committee in their recommendations is still considered by him as by far the most
suitable and valuable for routine work.

ADDENDUM 183

may be purified at a reasonable cost, so as to yield no B. coli
in 100 c.c. in the majority (about 75 per cent.) of samples
examined.

" It is much more difficult to suggest a standard by which a
water should be condemned. All that the Committee feel
justified in stating is that the further a water departs from the
above standard of purity (no ' lactose -f indol 4- ' B. coli in
100 c.c.) the greater is the suspicion attaching to it, unless the
local conditions and circumstances are such as to exclude
undesirable pollution."

3. Water in Swimming Baths.

The bacteriological content of the water in swimming baths
has been in recent years made the subject of a number of
investigations. It is a matter of some importance since a not
inconsiderable quantity of water is swallowed by bathers,
while in this and other ways it is obvious that infection may be
spread.

In a number of cases outbreaks of infectious disease, such as
typhoid fever, dysentery, Weil's disease, trachoma and infectious
conjunctivitis have been definitely ascribed to bathing in con-
taminated rivers or in public swimming baths.

The methods of examination are similar to those employed
for drinking water except that since the contamination may be
considerable, dilution methods must be practised and amounts
as small as O'Oi, cxooi c.c., etc., may have to be examined for
B. colt, number of bacilli, etc.

It is obvious that the bacteriological results will vary very
greatly with the local conditions, and no standards of permissible
numbers can be as yet satisfactorily laid down. The chief local
conditions affecting the bacterial content are the volume of water
in relation to the number of users, the frequency with which the
water is changed, the character of the users as regards personal
cleanliness, the initial purity of the water, the temperature of the
water and the local conditions of the examination (i.e. time of
examination in relation to freshness of the water).

A few particulars of investigations will serve to illustrate the
kind of findings met with.

1 84 ADDENDUM

Pearce and Sutherland1 studied the bacterial content of the
water of the swimming bath at Batley (Yorkshire). The first-
class swimming bath examined after being used for three days
contained on one occasion 2400 — 2850 organisms per c.c. and on
another 219,000 per c.c. The second class bath water after three
days gave 15,000 to 17,000 organisms per c.c. on one occasion
(with 767 users) and on another 300,000 per c.c. (with 974 users,
mostly elementary school children).

Atkins2 carried out a long series of studies of the water in
the swimming tank of the University of Chicago. The tank
holds about 91,000 gallons The examinations were extended
over a period of about two months (November and December),
during which time the bath was in daily use (Sundays excepted)
by about 100 people. Twice a week the water was completely
drawn off and the bath refilled with city water passed first
through a Jewel filter, using alum as a coagulant. The water
was raised to 76° — 80° F. by steam.

The entering water contained 250 — 700 per c.c. on gelatine
plates and about 100 — 125 organisms per c.c. on agar plates. By
the end of the first day the counts were in one series 192,000 (gela-
tine), 1 64,000 (agar) and in another series 47,000 (gelatine), 45,000
(agar). By the end of the second day in the first series they had
risen to 625,000 (gelatine), 500,000 (agar), and in the second to
510,000 (gelatine), 300,000 (agar). In the second series counts
were also made at the end of the third day and showed 500,000
(gelatine) and 200,000 (agar). Atkins remarks that this reduction
in numbers on the third day is in accord with other observers.

The B. coli results showed that the water before use contained
about 100 B. coli per litre, at the end of the first day of use about
200, and at the end of the second day nearly 1000 per litre
and about the same at the end of the third day.

Atkins found no parallel between the rise in the total
number of bacteria and the increase in the number of B. coli.

A number of experiments have been carried out to test
the efficiency of disinfectants to purify the water. Of these

1 Lancet, Aug. 1910, n, p. 542.

2 Proceedings of the yd meeting of the Illinois Water Supply Association. Feb.
1911.

ADDENDUM 185

hypochlorite of lime has been extensively tested and found to be
very satisfactory in action. Alexander at Poplar has also re-
ported upon the satisfactory use of an electrolytic disinfecting
fluid which liberates chlorine.

All these disinfection procedures should be carefully tested
and controlled by bacteriological methods.

4. The isolation of B. typhosus from Water, Sewage, etc,

The different methods and procedures are described at con-
siderable length on pages 39 — 44. While infrequently required
it is important that the most reliable methods should be used.
The following additional particulars will be of interest.

Houston1 has recently further studied this question and
compared several methods. He tested sewage artificially
inoculated with a small previously ascertained number of
typhoid bacilli, examining the first series of these experiments
in duplicate, using both direct platings on malachite green agar
plates and the enrichment method with ox-gall medium, and
the second series also in duplicate, using the same direct method,
but instead of ox-gall employing the brilliant green enrichment
method described by Browning, Gilmour and Mackie2.

In all six experiments the typhoid bacilli were recovered in
considerable numbers (varying from 9 to 1 1 8), but in every case
by the direct method. In no case was a typhoid bacillus
recovered by either of the indirect methods. As carried out the
experiments were strictly comparable, and the tests were applied
to a sewage medium and not to typhoid excreta. This is
important since it allows direct deductions to be made from
them as to the reliability of the methods tested for the isolation
of typhoid bacilli from sewage or water.

These experiments bear out the conclusion given on page 40
that "enrichment and selective enrichment methods are less
satisfactory than sedimentation or direct concentration."

Stated briefly Houston's method for the examination of
water or sewage for the typhoid bacillus is as follows :

500 c.c. of the sample is centrifugalised and the deposit

1 Metropolitan Water Board : Tenth Research Report, 1914.
^ Jo urn. of Path, and Bacteriology, 1913, xvm, p. 146.

1 86 ADDENDUM

spread over 16 malachite green bile salt agar plates. After
incubation at 37° C. for 24 hours a large number of the colourless
colonies are picked off (Houston in his experiments examined
250, if so many were obtainable), subcultivated and investigated.

The malachite green agar used by Houston differs from that
given in the appendix, and has the following composition : agar
2 per cent, peptone 2 per cent., bile salt 0*5 per cent., lactose,
saccharose, adonite, raffinose and salicin 0*2 per cent. each.
Neutral red is added in amount equal to 4 c.c. of a I per cent,
solution to each litre of medium. The medium is made of
this composition and stored. Just before it is used for pouring
into the Petri dishes malachite green is added in the proportion
of O'l gramme per litre of medium.

This medium can also be used for the isolation of bacilli of
the Gaertner group. These bacilli as well as the typhoid bacillus
produce colourless colonies.

In the utilization of methods for the isolation of the typhoid
bacillus from materials containing variable amounts of organic
matter it is important to realize that methods found satisfactory
for the isolation of this organism from excreta may be less suit-
able when water or sewage is the substance under investigation.

Recent investigations by Krumwiede and Pratt1 upon the
growth of bacteria in and on media containing various aniline
dyes throw light upon the relationship of the method to the
vehicle examined.

They ascertained that in general the paratyphoid-enteritidis
types are highly resistant to the green dyes, the typhoid bacillus
less so but somewhat more resistant than the coli types, but they
also found that slight changes affected the action towards
B. typhosus. For example small additions of proteid substances
change the action of the dye.

They directly tested the influence of faeces upon brilliant
green and the typhoid bacillus, and found that faeces reduced
the activity of the dye by about one-third. They add " when
we add faeces we are introducing a variable factor, and the
typhoid strains also vary somewhat in their resistance to the
dye. Success depends upon the right adjustment."

1 Journ. of Exp. Med. 1914, xix, p. 501.

ADDENDUM

.87

5. Antiformin in the Examination of Milk for Tubercle

Bacilli.

In the text (p. 100) it is explained that it is advisable to
inoculate several guinea-pigs from each centrifugalised milk
sediment to determine if the tubercle bacillus is present, since a
certain proportion of the animals always die from the action of
other bacteria present in the milk.

The chemical solution known as antiformin has been found
to readily destroy most bacteria, but the acid fast bacilli resist
destruction for a considerable time. This suggests the possibility
of treating the milk or the centrifugalised deposits before inocu-
lation with this substance so as to kill as many as possible of
the extraneous bacilli.

The essential points which have to be settled before this
method can be employed are, on the one hand, the extent to
which the extraneous bacilli are killed or sufficiently reduced in
numbers, and, on the other hand, as to how far any tubercle
bacilli present are destroyed or their vitality injured by the
treatment with antiformin. The experimental results obtained
by Eastwood and Griffith1 bear directly upon these points.
They found that neither 3 nor 5 per cent, solutions, with
exposures of 30 to 100 minutes, were sufficient to prevent over-
growth by extraneous organisms. Using a 10 per cent, solution
with times of action from 20 to 60 minutes the results were
variable, as in some cases pure cultures of the tubercle bacillus
were obtained, while in others this strength failed to eliminate
extraneous bacilli.

1 5 per cent, and stronger solutions were found to be more or
less destructive to the tubercle bacilli, and in one experiment
even a 10 per cent, solution exerted some harmful action since
only scanty tubercle bacilli were recovered in pure culture.

These experiments are in general agreement with those of
other workers with sputum, etc. While not conclusive they
suggest that treatment of milk sediments for 20 to 30 minutes
with a 10 per cent, solution of antiformin will very greatly

1 Report of Medical Officer, Local Government Board, 1914, XLII, p. 203.

188 ADDENDUM

diminish and in some cases eliminate extraneous bacteria, while
exerting little or no pathogenic action upon any tubercle bacilli
present.

6. Butter.

Insufficient attention is probably given to this food from the
bacteriological Public Health point of view so that the following
additional methods and particulars should be of interest.

To examine butter for the tubercle bacillus the technique
recommended by Eastwood and Griffith1 may be employed.

A quarter of a pound of the sample is cut up and transferred
to a sterile flask with a side-tube at the bottom. About 60 c.c.
of sterile normal saline solution, warmed to a temperature of
37° C., is added to the contents and the flask placed in the
incubator (37° C.) in a vessel filled with water at the temperature
of the incubator. The time required for thorough melting of
the butter is generally about two hours. It is important to get
the butter melted as soon as possible so as to avoid unnecessary
multiplication of extraneous organisms, but time must be
allowed for all the fat to rise to the surface. When the melting
is completed the lower milky portion is withdrawn through the
side-tube, made up to about 100 c.c. by the addition of more
saline, well shaken and centrifugalised for 20 minutes (centrifuge
running at about 3500 revolutions per minute used). The whole
of the deposit is then used to inoculate guinea-pigs. Eastwood
and Griffith inoculated two each with |ths and a third with -J-th
of the deposit. A fourth guinea-pig was generally inoculated
with 10 c.c. of the clear melted fat.

1 08 samples of butter were examined in this way.

40 samples were purchased in London and the suburbs with
the assurance that the material was of British origin. Of these
20 yielded inconclusive results, 1 8 were definitely negative, while
2 were found to be tuberculous.

20 samples were obtained from Shrewsbury and produced
locally. Of these 2 yielded inconclusive results and the
remaining 1 8 showed no tubercle bacilli.

1 Report of Medical Officer, Local Government Board, 1912—13, p. 295 : issued in
1914.

ADDENDUM 189

48 were foreign samples received from the Board's Inspectors
of Foods. Of these 4 yielded inconclusive results, while none of
the remaining 44 were found to be tuberculous.

The " inconclusive " results were the cases in which prema-
ture death of the guinea-pigs was caused owing to the high
infectivity of these butters for these animals. Excluding these
butters 2 out of 82 (2*4 per cent.) were found to contain tubercle"
bacilli.

In no cases were acid fast bacilli, other than the tubercle
bacillus, found in animals inoculated with the fat-free deposit,
but these organisms were found in a number of cases when
guinea-pigs were inoculated either with the clear melted fat or
with material containing an admixture of fat.

Rosenow, Frost and Bryant1 have recently published a paper
on a study of the Market Butter of Boston, U.S.A.

In their method of examination for bacteria, B. coli, etc.,
I grm. of butter was withdrawn by sterile spatula and mixed
with 100 c.c. of sterile tap water at 40 — 45° C. in a jar. While still
warm the jar was shaken for 1 5 minutes in a shaking machine.
To make the required dilutions I c.c. of this emulsion was
quickly transferred to a flask containing 100 c.c. of sterile water.
From this further dilutions were made in the ordinary way.

For testing for the tubercle bacillus various quantities were
used. Some of the guinea-pigs received the butter itself, others
were inoculated with washings or sediment obtained from
centrifugalising the washings, and some with both butter and
sediment. The majority received the sediment from 50 grms.
of butter. For this purpose the authors suggest that the best
method is to shake the bacteria out of the butter by violently
agitating 50 grms. in warm water in a shaking machine. The
mixture is rapidly cooled, the water and bacteria poured off and
centrifugalised and the sediment used for the animal inocula-
tion.

The authors investigated the bacteria washed out from the
butter and found that 83 to 99 per cent, (average 90*4 per cent.)
were removed in the first washing.

25 butters were examined, all in the months of June and July.

1 The Journal of Medical Research) 1914, XXX, p. 69.

ADDENDUM

All were purchased from retail traders, and for most the age was
unknown.

For the tubercle bacillus test 84 guinea-pigs were infected
intraperitoneally from 21 of the 25 samples. The remaining four
samples were not tested because they were known to come from
pasteurized cream. Two of the guinea-pigs developed tubercu-
losis, a percentage of 9'5. Antiformin was used to destroy
extraneous bacilli in a few cases, but the authors remark " it is
doubtful whether the use of antiformin is serviceable in this
connection."

The number of bacteria (agar plates at 37° C.) found per grm.
of butter averaged 2,700,000 for the salted butters and 30,000,000
for the unsalted butters. The lowest count was 8600 and
17,000,000 the highest for the salted and 41,000,000 for the
unsalted. The wide variations in the bacterial count were
apparently not associated with any other constituent deter-
mined, such as salt, acidity and moisture. Like earlier
investigators they found the bacterial content largely influenced
by the age of the butter. All the samples examined within a
few days of being made contained from one to several million
bacteria per grm., while all the low counts were from butters
which had been sent for some distance or which were of uncer-
tain age.

This question was directly tested in a few cases. In one
sample there was a reduction of 85*8 per cent, in two weeks, in
another 937 per cent, in four weeks and in another 95*6 per
cent, in six weeks.

B. coli was found in only 6 of the 25 samples when O'Oi grm.
or less was examined, and then only in small numbers. From
a few special experiments made it seems that this bacillus soon
dies out in butter.

Streptococci were found in 14 of the 25 samples in 0*01 grm.
No special relation between the presence of streptococci and
virulence to guinea-pigs could be made out.

B. welchii (B. enteritidis sporogenes) was absent from I grm
of all the 25 butters.

ADDENDUM 191

7. Condensed and Dried Milk.

Two investigations upon these varieties of milk have been
recently carried out in this country and some of the findings are
of considerable interest. Andrewes1 found that condensed milk
like fresh milk samples contained large numbers of cells, including
numerous polymorphs, but when quantitative estimations were
made the actual numbers per c.mm. in machine-skimmed con-
densed milks were hardly greater than in fresh uncondensed
milk, although in the latter the milk is reduced to a quarter or
less of its bulk. Andrewes considers that the reduction is due
to the mode of preparation, the centrifugalisation removing
many cells with the cream, debris and dirt.

Andrewes examined bacteriologically 43 samples of machine-
skimmed condensed milk and all contained bacteria but the
quantitative method used was only a rough one. The results
obtained fell into three main groups :

(a) Seven which yielded a moderate number of colonies,
amongst which staphylococci were relatively few. B. mesen-
tericus was commonly present.

(b) Five yielded a more abundant growth in which white
staphylococci and coliform bacilli were the prevailing organisms
in almost equal numbers. Staphylococcus aureus was only
present in one sample and in that in scanty numbers.

(c) Thirty-one samples yielded almost pure cultures of
staphylococci, amongst which Staphylococcus pyogenes aureus was
conspicuous and often predominant.

Andrewes discussed the significance of Staphylococcus aureus
and albus in the samples and showed that staphylococci can
multiply abundantly in condensed milk, the numbers found
depending upon the age of the milk. " It would appear that if
only a few Staphylococcus aureus were present and escaped
destruction in the process of condensation, there is no limit
to the number which later may be found on opening the tin."

1 Journ. of Pathology and Bacteriology, 1913, xvm, p. 169.

IQ2 ADDENDUM

While admitting that the question is undecided he gives his
opinion that the presence of Staphylococcus pyogenes aureus in
large numbers in a sample of condensed milk is objectionable
and a ground for its condemnation.

Delepine1 investigated the effects of certain condensing and
drying processes upon the bacterial content of milk subjected to
them. The three processes studied were : (A) The preparation
of sweetened condensed milk, condensed by evaporation in
vacuo at 4O°-45° C. ; (B) the preparation of dried milk by
passage between heated revolving cylinders, and (C) the pre-
paration of dried milk by spraying the milk into a current of
hot air.

In all three processes the total number of bacteria in the
original milk was markedly reduced, most by Method A, least
by Method C. In each of the three methods of treatment there
was a stage in which the reduction in the total number of
bacteria was much greater than that observed in the finished
article ready for sale. The increase in the number of bacteria
observed during the final stages was due to recontamination
subsequent to the partial or complete sterilization stages. The
reduction in the total number of bacteria was almost entirely
due to the death of streptococci, staphylococci, sarcinae, bacilli
of the B. coli type, streptothrichae, yeasts, etc. At none of the
stages of preparation was the milk ever found completely sterile.
The amount of heat to which the milk was subjected was
insufficient to bring about the death of several saprophytic and
of some pathogenic bacteria. Among the saprophytic bacteria,
which were invariably found to resist pasteurisation, those most
commonly detected were sporing bacilli of the types included
under the term B. mesentericus. Some streptothrichae appeared
to have survived in some cases, but the evidence on that point
was not conclusive.

Careful investigation was made as to the efficiency of the
methods to kill tubercle bacilli. Naturally infected tuberculous
milk was used, fortified by the addition of numerous tubercle

1 Report to the Local Government Board upon the effects of certain condensing
and drying processes used in the preservation of milk upon its bacterial contents :
Food Reports, No. 21, 1914*

ADDENDUM

193

bacilli from a virulent bovine culture. Some living tubercle
bacilli of bovine origin were found to have survived the treat-
ment given according to Method B, and Delepine adds " It may
be safely assumed that Method C, which yields a product giving
a higher total bacterial count than Method B, has even less effect
upon tubercle bacilli. The same bacilli resisted the process of
pasteurisation which forms part of Method A."

" The tubercle bacilli which had survived pasteurisation in
Method A and drying by heat in Method B, were still capable
of producing progressive tuberculosis in guinea-pigs inoculated
subcutaneously with milk containing these bacilli but the course
of the disease produced by these bacilli was very much slower
than that of the disease produced in guinea-pigs inoculated with
the same amount of untreated tuberculous milk. The tuber-
culosis produced by the heated milk was latent or occult for some
four weeks. Young rabbits fed with milk containing these
modified bacilli did not contract tuberculosis."

8. Bread.

Bread is a food-stuff which is subjected to a certain amount
of bacterial contamination under the conditions of distribution
which prevail and it may be of interest to give some particulars
as to the extent to which it has been found to be bacterially
contaminated.

From the Public Health point of view the only questions of
importance are :

1. The extent to which bread and its constituents are
infected before baking and if any of the bacteria so taken up
survive the baking.

2. The extent to which bread becomes infected with bacteria
on its surface and any evidence of the transfer in this way of
pathogenic bacteria.

Apart from these points there are certain so-called diseases
of bread which render the bread sticky, slimy or sour. Such
conditions are bacterial in causation and different bacteria have
been isolated and described such as B. panist B. liodermis and

s. w. 13

IQ4 ADDENDUM

various varieties of B. mesenteriais. Occasionally bread has
been coloured red by B. prodigiosus. Upon stale bread moulds
such as Mucor mucedo and A spergillu s glaucus may develop.

The bacteria in bread are for the most part killed in the
process of baking but several investigators (i.e. Roussel,Marchand)
have inoculated dough with B. tuberculosis and recovered this
organism in a pathogenic state after baking. In experiments
conducted by Auche on the other hand this organism was
destroyed, or was incapable of producing tuberculosis in inocu-
lated animals, after the bread was baked.

Of more importance is the question of the extent to which
bread is bacterially contaminated after baking either in the
bakery or during its distribution. The writer is unaware of
investigations in this country but the following undertaken in
America and Germany illustrate the extent of bacterial con-
tamination which may occur. K. Howell1 examined 100 loaves
from various districts of Chicago, some wrapped, some un-
wrapped, and from shops of all degress of cleanliness. The
method of examination used was to swab the entire loaf with
wet sterile cotton the cotton then being thoroughly rinsed in
loc.c. of sterile water. From this suitable dilutions were made.
The platings were done upon gelatine and lactose litmus agar,
the former being counted after 72 hours at 20° C. and the latter
after 24 hours at 37° C. All acid colonies were picked off and
examined for B. coli and streptococci. In addition I c.c. of
the original emulsion from each loaf was added to each of five
fermentation tubes containing lactose broth, for gas production
and B. coli.

B. coli was isolated from three loaves and streptococci from
30 samples. As regards the number of bacteria the following
results were obtained :

Percentage of samples with Unwrapped Bread Wrapped Bread
More than 10,000 bacteria* 17 o

Tooo-10,000 „ 45 15

Less than 1000 „ 38 85

*Agar at 37°C.
1 American Journ. of Public Health^ 1912, n, p. 321.

ADDENDUM 195

Examinations of bread as vended in New York have also
been carried out1. The method of examination employed was
to scrape the crust of the loaf on to sterile paper. The scrapings
were placed in a sterile test tube to which was added 20 c.c. of
salt-free broth. The mixture was shaken 25 times and plated
on agar and Conradi medium in amounts of O'l, O'5, ro and 2 c.c.
These amounts were also added to lactose neutral-red fermenta-
tion tubes for B. coli group enumerations. Three loaves which
were claimed to be free from human handling until wrapped up
yielded 600,200,280 bacteria (agar 2 days at 37° C.) respectively
while all were free from B. coli in 5 c.c. Three similar loaves
from the same batch were given to three different groups of men
to handle and were then examined. All three loaves still showed
absence of B. coli but the numbers of organisms (2 days agar at
37° C.) were respectively 15,140, 1080 and 1360.

Six other samples of unwrapped bread showed bacteria
varying from 2720 to 325,500 per loaf while B. coli was isolated
from four of them in I c.c. of the emulsion. A single sample of
wrapped bread yielded no B. coli in 5 c.c. and 800 bacteria per
loaf.

Furst2 heavily inoculated bread with emulsions of various
bacteria and studied their viability under different conditions.
He found that such bacteria lived much longer in the crumb
than on the crust, B. typliosus, B. paratyphosus (3 and two
dysentery bacilli (Flexner and <y strains) living as long as 10 to
25 days on the former and only I to 8 days on the latter. On
white bread these bacilli survived for considerably longer periods
than on German black bread.

While it is obvious that bread handled by an infectious
person may serve as a vehicle for the transmission of disease
there is little or no definite evidence showing that such trans-
mission has taken place in specific instances.

Furst (loc. cit.) studied an outbreak of typhoid fever, for the
most part of mild type, which occurred in an institution for
children in August 1913 and considered that it was spread by
contaminated bread. The bread was obtained from a bakery in

1 The Medical Officer, Jan. 9th, 1915.

2 Munch, med. Woch., 1914, LXI, p. 1442.

13—2

196 ADDENDUM

which several of the workers, including the wife (probably),
brother and daughter of the baker, suffered from enteric fever.
Further cases occurred in other families in the place in which
the bakery was situated, making 15 in all, of whom 7 died.

According to Lumsden1 several cases of enteric fever in the
Government Hospital for the insane at Washington were traced
to the fact that the attendant who cut the bread as his first duty
in the morning, had been taking care of a typhoid patient in his
home.

As regards the procedures to be used for the bacterio-
logical examination of bread those described above sufficiently
indicate the methods which may be employed. The bacterio-
logical count, estimation of number of B. coli and examinations
for any pathogenic bacteria suspected to be present are the
most important. The methods to be used, after the emulsion
containing the bacteria from the bread has been prepared, are
similar to those generally employed for these examinations.

1 Bulletin, No. 78, Hygienic Laboratory -, U.S. Public Health and Marine
Hospital Service, p. 165.

INDEX

Acid-fast bacilli in butter 117, 189

in milk 103
Acidity of milk 96, 99
Aesculin agar 165
Agar 161
Air 136

bacterial content 137

bacteriological examination 141

of sewers and drains 139

Streptococci in 146
Anthrax (see B. anthracis}
Antiformin methods 187, 190

Bacillus anthracis, in sewage 70

soil 64

arbor escens in soil 62
botulinus 130, 134
buiyricus 10, n
coli, group 3

characters of 3

in excreta 4, 5
distribution of 13
in air 141

brawn 125

bread 194

brines 126

butter 190

cockles 80

condensed milk 112, 191

" deep" waters 53

dust 138

fish 124

food poisoning outbreaks 135

ice-cream 120, 121

milk 93, 98

mud 82

mussels 77

oysters 75

putrefaction 130

river water 58

sausages i 25

sea water 75

sewer air 140

"shallow" well water 56

soil 65

Bacillus coli, in swimming baths water

183

in upland waters 51
isolation of 17

from water 34, 36, 180
pathogenicity of 6
diphtherias in butter 117
in milk 104
viability in soil 63
enteritidis 131, 132

sporogenes, characters of 9
distribution of 15
in butter 190
cockles 80
condensed milk 112
dust 138
ice-cream 120
milk 93, 98
mussels 79
oysters 75
sea water 75
soil 65
water 38
isolation of 22
fluorescens liquefaciens in soil 62

non-liquefaciens in soil 62
liodermis in bread 193
megatherium in soil 62
mesentericus 192, 194
mycoides in soil 62, 65
panis in bread 193
paratyphosus /3 131, 132
pestis in soil 64
proteus in fish 124

in food poisoning outbreaks 130,

134

hackfleisch 125
putrefaction 127, 129
soil 62
isolation 135
subtilis in soil 62
suipeslifer 131, 132
tuberculosis in bread 194
in butter 116, 188
cheese 118

198

INDEX

Bacillus tuberculosis in condensed milk
112, 192

in cream 114
dried milk 192
expired air 145
ice-cream 121
milk 99, 187
isolation of 101, 187
typhosus, identification tests for 43
in butter 117
cockles 8 1
milk 105
oysters 80

isolation from water 39, 185
viability in sewage 70, 73
in soil 63

water 48
on bread 195

Bacterial excreta indicators n, 15
sewage indicators n, 15
standards for water 58
Blood serum, preparation of 162
Botulism 130
Brawn 1 30
Bread 193

Brilliant green agar 166
Brines, bacterial content 126
Broth, preparation of 160
Butter, acid-fast bacilli in 117, 189
bacteriology of 115, 1 88
B. diphtheriae in 117
B. tuberculosis in 116, 189
B. typhosus in 117
B. welch ii in 190
starters 1 1 5
Butter-bacillus 103, 117

Cellular content of condensed milk 191
of milk 105, 108

Cheese 118

Cladothrix dichotoma in soil 62

Cockles, bacteriology of 79

bacteriological examination 82

Condensed milk in, 191
B. coli in 112

B. enteritidis sporogenes in 112
B. tuberculosis in 112
bacteriological examination 113
Staphylococci in 191

Cream 1 1 3

Deep well water 52

Dieudonne's alkaline blood agar 169

Dilution methods 25

Disinfectants 148

Drain air 139

Dried milk 113, 192

Drigalski-Conradi agar 168

Dulcite malachite green broth 166

Dust, bacteria in 138

Egg medium 163

Excreta, bacteria in 5, u, 12

Excretal B. coli 6

Filter beds, bacterial testing 60
Fish, bacteria in 124

B. coli in 12

Food poisoning outbreaks, bacterial in-
vestigation 133
Fraenkel's borer 68
Frankland's tube 143
Fuchsin agar 167

Gaertner group bacilli 131

isolation in food outbreaks

Germicidal power

Chick and Martin's method 154
determination of 148
Garnet method 150
Lancet method 153
modifying influences 154
organic matter method 156
Rideal-Walker method 151
thread method 1 50

Glycerine agar 162

134

Hackfleisch 124
Haemolytic tests

172

Ice-box for water samples 30
Ice-cream 119

B. coli in 120

B, enteritidis in 121

B. enteritidis sporogenes in 120

bacterial standards 121
Indicator organisms 3

Johne's bacillus 103

Lactose bile salt broth 164 ; L.B.A. 167
Leucocytes in milk 105, 108
Litmus milk 162
Loffler's blood serum 163

Malachite green agar 165, 186

Mastitis no

Meat, bacteria in 122

bacteriological examination of 126

chopped 124

enrichment examination method 128

examination for pathogenic bacteria

126

Media, Reaction Standardization 158
Milk, acid-fast bacteria in 103

acidity 96, 99

B. diphtheriae in 104

B. coli in 93, 98

B. enteritidis sporogenes in 93, 98

B. tuberculosis in 99

B. typhosus in 105

INDEX

199

Milk, bacteriological examination 85,

110

cellular content 105, 108
collection of samples 85
estimation of bacteria 92
samples, particulars required 87
sediment 95, 99
sources of bacteria in 90
streptococci in 107, 109
Streptococcus mastitidis in no
Milk-coolers and bacteria 91
-separators and bacteria 91
-strainers and bacteria 91
Mist-bacillus 104
Mud, bacteria in 81
Mussels, bacteriological examination 82
purification 81
bacteriology of 77

Neutral red glucose broth 165

factose bile salt agar 167
lactose bile salt crystal violet

167

Nutrient agar 161
broth i 60
gelatine 160

Oysters, bacteriological examination 82
purification of 80
standards for 78

bacteriology of 75

B. typhosus in 80

Para-Gaertner bacilli 131

Peptone water 162

Petri's sand filter 142

Pneumococcus 147

Ptomaines 129

Putrefaction, bacteriology of 127, 129

River water 58

Sausages 125
B. colt in 125
streptococci in 125
Sedgwick-Tucker tube 142
Sea water 75, 81
Sewage, bacteria in 12

bacterial content 70
bacteriological examination of

73

excretal bacteria in 70
influence of treatment 70
pathogenic bacteria in 73
samples 74

Sewer air 139

Shallow well water 56

Shell fish, bacteriological examination
74. 82

Smegma bacillus 104

Soil, B. anthracis in 64
B. diphtheriae in 63
B. pestis in 64
B. typhosns in 63
bacterial content 61
bacterial examination of 67
excretal bacilli in 64
pathogenic bacteria in 62
Spirilhim cholerae

identification tests 46
isolation from water 44
viability in sewage 70, 73
Staphylococcus albus in condensed milk

191
Staphylococcus aureus in condensed milk

191

Staphylococcus epidermidis albus 147
pyogenes 147
salivarius 147

Streptococci, agglutination reactions 178
characters of 7

in excreta 8

constancy of characters 177
differentiation of 171
distribution of 14
fermentation tests 174
Gordon's tests for 8
haemolytic tests 172
in air 146
butter 190
dust 138
ice-cream 120
milk 107, 109
mussels 79
sausages 125
soil 65
water 53, 182
isolation of 21

from water 37, 182
viability in water 14
Streptococcus anginosus 147
equimis 147
faecal is 147
mastitidis no
mitior 172
mucosus 172
pyogenes 147, 172
salivarius 147
Streptococcus group 7, 171
Sugar media 163
Swimming baths water 183

Timothy grass bacillus 103
Tuberculosis of udder 105

Udder tuberculosis 105
Upland surface waters 51

Volatile disinfectants 157

2OO INDEX

Water, bacterial content 27, 50 Water, quantitative examination 32

examination of 27, [79 Report of Standardization Coin-
standards for 58, 182 mittee 179
collection of samples 29, 30 river supplies 58
"deep" supplies 52 samples, particulars to record
examination for B. coli 34, 180 31

B. enter itidis sporogenes 38 shallow well 56

B. typhosus 39, 185 significance of bacterial results

streptococci 37, 182 47

Sp. cholerae 44 subsoil 56

in swimming baths 183 transmission of samples 29

objects of examination 28 upland surface 51

qualitative examination 34, 179 viability of B. typhosus in 48

CAMBRIDGE : PRINTED BY J. B. PEACE, M.A., AT THE UNIVERSITY I»BESS<

CAMBRIDGE PUBLIC HEALTH SERIES

Volumes now ready, continued

Sewage Purification and Disposal. By G. B. KERSHAW,
M.Inst.C.E. Demy 8vo. With 56 illustrations. 123 net.
" No book more worthy of perusal and study by those who are interested
in the subject of sewage purification has been published in recent years.. . .
A technical book which is at once readable and instructive." — Journal of the
Institute of Sanitary Engineers

Infant Mortality. By H. T. ASHBY, M.A., M.D. Demy 8vo.

With 9 illustrations. los 6d net.

" A painstaking digest of the vast literature on the subject which has of
late years accumulated.. . .Dr Ashby has done his work exceedingly well and
has given fair and unprejudiced acknowledgment to the several factors which
have contributed to the great fall in the mortality rate among infants." — Lancet

The Methods of Post-mortem Examinations. By J. BEATTIE,
M.D. Demy 8vo. With 8 plates and 3 text-figures. los 6d net.
" Our impression of the book is extremely favourable, and we can con-
gratulate the author on the issue of a most useful and practical volume

We can with confidence recommend it to our readers, and we feel sure it will
enjoy a wide and well-merited circulation." — Medical Times

Occupations. From the Social, Hygienic, and Medical points of
view. By Sir THOMAS OLIVER, M.D:, D.Sc., F.R.C.P. With 2 charts.
Demy 8vo. 6s net.

In the press
The Spread of Tuberculosis. By L. COBBETT, M.D.

In preparation

Ticks as Carriers of Disease. By G. H. F. NUTTALL, F.R.S.
Serum Diagnoses. By CARL BROWNING, M.D.

The Purification of Water in Sedimentation, Filtration and

Precipitation. By A. C. HOUSTON, M.B., D.Sc.
The Purification of Water by Ozone and Chlorine ; and

Domestic Filters. By G. SIMS WOODHEAD, M.A., M.D.
The Principles and Practice of the dilution method of

Sewage Disposal. By W. E. ADENEY, D.Sc.
Disinfection. By C. W. PONDER, M.D.

Housing in relation to Public Health. By C. J. COLEMAN, M.D.
School Hygiene. By E.T.ROBERTS, M.D.

Soils, subsoils, and climate in relation to Health. By G.

WALKER, M.B.
Meat Inspection. By W. J. HOWARTH, M.D., and T. DUNLOP

YOUNG, M.R.C.V.S.

Vital Statistics. By R. DUDFIELD, M.B., and G. UDNY YULE, M.A.
Foods, Sound and Unsound. By H. C. HASLAM, M.A., M.D., Sc.D.

Cambridge University Press
C. F. CLAY, Manager : Fetter Lane, London

also
London: H. K. LEWIS, 136, Gower Street, W.C.

2

THE CAMBRIDGE PUBLIC HEALTH SERIES

UNDER THE EDITORSHIP OF
G. S. Graham -Smith, M.D. and J. E. Purvis, M.A.

University Lecturer in Hygiene and University Lecturer in Chemistry and
Secretary to the Sub-Syndicate for Physics in their application to Hygiene
Tropical Medicine and Preventive Medicine, and Secretary

to the State Medicine Syndicate

The books in this series are written by experts, and the authors are
occupied, or have been occupied, either in investigations connected
with the various themes or in their application and administration.
It is hoped that the volumes will be useful to the medical profession
at home and abroad, to bacteriologists and laboratory students, to
municipal engineers and architects, to medical officers of health and
sanitary inspectors and to teachers and administrators.

Volumes now ready
Flies in Relation to Disease : Non-Bloodsucking Flies.

By G. S. GRAHAM-SMITH, M.D. Second edition. Revised and enlarged.
With 27 plates, 32 text-figures and 20 charts. Demy 8vo. i2s 6d net.
" That a second edition of this important work should have been demanded
within a year of the publication of the first is conclusive evidence of its merits. ...The
work is one of great value, and should be studied by every medical practitioner."

Medical Times

Flies in Relation to Disease : Bloodsucking Flies. By

E. HINDLE, B.A., Ph.D. With 88 text-figures. Demy 8vo. 125 6d net.
"We must congratulate Mr Hindle on the completeness of his work, on
his admirable illustrations, and on his exhaustive index." — British Medical
Journal

Isolation Hospitals. By H. FRANKLIN PARSONS, M.D.,

D.P.H. With 55 Illustrations. Demy 8vo. 125 6d net.

"The author has given to Medical Officers of Health, hospital superintendents,
and all persons concerned in the control and management of isolation hospitals, the
best and latest information on the subject of isolation hospitals, their utility, design,

cost, and the latest approved examples We are of opinion that this book is one

which few Medical Officers of Health and hospital superintendents can afford to be
without. We cannot recommend it too highly." — Sanitary Record

The Bacteriological Examination of Food and Water. By

WILLIAM G. SAVAGE, B.Sc., M.D., D.P.H. Second Edition. With
1 6 Illustrations. Demy 8vo. 75 6d net.

" It is rare to find in one individual such a happy blend of administrative
experience and specialised technical knowledge as is possessed by Dr Savage, and the

results are shown in every page of this admirable treatise No person interested in

the public health aspects of foodstuffs and water supplies can afford to be without this
excellent treatise." — Public Health

The Chemical Examination of Water, Sewage and Foods.

By J. E. PURVIS, M.A., and T. R. HODGSON, M.A. Demy 8vo. 9s net.
"This book is extremely clear and concise, containing within 228 pages the
methods of analysis for water, sewage, foods, disinfectants, gases, ragflock and urine....
The large number of typical analyses, accompanied by notes and descriptions of the
reasons for the final decisions, are exceedingly helpful in that difficult matter, the
interpretation of the results of an analysis — The book is exceedingly useful, and
the references and bibliography throughout the book are valuable to any worker
on these subjects."—; -Journal of the Royal Sanitary Instittite

[P. T. O.

93640

UNIVERSITY OF CALIFORNIA LIBRARY