A BIOLOGICAL ANALYSIS 
—OF THE— 
\ 
MONTREAL WATER SUPPLY DURING THE PERIOD PROM 
NOVEMBER, 1890, TO NOVEMBER, 1891. 
BY 
WYATT JOHNSTON, M.D., 
Lecturer in Bacteriology, McGill University ; Bacteriologist to the Quebec Provincial 
Board of Health. 
{Reprinted from the Montreal Medical Journal, August, 1894.) A BIOLOGICAL ANALYSIS OF THE MONTREAL 
WATER SUPPLY DURING THE PERIOD FROM 
NOVEMBER, 1890, TO NOVEMBER, 1891 * 
By Wyatt Johnston, M.D. 
Lecturer in Bacteriology McGill University ; Bacteriologist to the Quebec Provincial 
Board of Health. 
The following account of a biological analysis, made three 
years ago, has been abridged from my report addressed at the 
time to Mr. B. D. McConnell, then Superintendent of the 
Montreal Water Works, who took a deep interest in the inves- 
tigation. Chemical analyses were made at the same time by 
Prof. R. F. Ruttan and Prof. Phister. 
PLAN OF INVESTIGATION. 
I. Regular monthy examinations of samples of water from 
the following four localities : 
1. The lower reservoir. 
2. The settling basin. 
3* A point near the intake of the St. Cunegonde Water 
Supply. 
4. A point in the middle of the River St. Lawrence south 
of Nun’s Island. 
These examinations were made at the express order of the 
Water Committee with a view of determining whether the water 
* Published by permission of the Water Committee of the Montreal City Council. 4 
obtained from localities 3 and 4 would be preferable to that 
furnished by the present intake on the north shore of the St. 
Lawrence, just above the Lachine Rapids. 
In addition, I found it necessary to make : 
II. Examination of tap water obtained from various points 
within the city, from the upper reservior, and from the aqueduct, 
Fig. l. 
to see whether evidences of local contamination existed and to 
trace the effect of temperature, rainfall and water level. 
III. Examination of the water of the St. Lawrence and 
Ottawa rivers at points above Montreal, to see whether the. 
influence of the sewage from the towns along their banks was 
perceptible. 
IV. Examination of surface waters from other parts of 
Canada, and especially from uninhabited districts. 5 
METHODS. 
A large proportion of the work consisted in the estimation of 
the number of bacteria present. (Quantative bacterial analysis.) 
The nature of the bacteria was also studied, as far as the time 
limits of the analysis permitted. Cultures, for quantitative 
work, were for the most part made in slightly alkaline, 10 
p.c. beef peptone gelatine, made after Loeffler’s formula, and 
grown at 20°C. The samples were taken 10 to 20 feet 
below the surface, by means of an apparatus shown in 
figure 1, and were plated in flat glass vials. The cultures 
were, as a rule, made within a few minutes of the time of taking 
the samples, and in a few instances, when about an hour or two 
intervened, the samples were kept in an ice box. 
The sediments were all examined microscopically, and during 
four months the microscopical organisms present were estimated 
quantitatively by the Sedgwick-Rafter method. 
Before giving the details of the analysis, it might be well, in 
order to make the report intelligible to those who are not 
familiar with the local conditions of the Montreal water supply, 
to briefly mention the character of the water, and the topography 
of the district from whence it is obtained. 
Although taken from the north shore of the St. Lawrence 
river, the Montreal water supply is derived, during the greater 
part of the year, from the Ottawa, which enters the St. Law- 
rence from the north at a point about 20 miles above the intake, 
and forms a belt of dark water close to the shore, the border 
between this water and the clear green of the St. Lawrence 
proper being very distinct, though varying in position with 
changes in the direction and force of the wind and the rela- 
tive level of the water in the two rivers. During the winter 
owing apparently to an ice-jam, the Ottawa passes to the north 
of the island of Montreal, so that the Montreal supply during 
the months of January, February and March consists of nearly 
pure St. Lawrence water. 
SOURCE OF SUPPLY. 
Ottawa River. 
The Ottawa river drains an area of over 60,000 square miles 6 
(rather less than the Danube), most of which is entirely unin- 
habited. Its discharge has been estimated at 60,000 cubic feet 
per second. Its average width for the 100 miles above Mont- 
real is somewhat over half a mile. At 25 miles above the city 
it expands into the lake of Two Mountains, varying from 2 to 
4 miles in width, and 4 miles above the intake, into Lake St. 
Louis, 4 to 7 miles wide. There are rapids and falls 60 and 
30 miles above Montreal. At many points between Ottawa 
and Montreal navigation is impeded by enormous sawdust beds 
from the Ottawa saw mills. 
The population along its course, according to the census of 
1891, is about 300,000, or 6 per square mile, of which about 
100,000 is comprised in cities or towns of over 1,000 inhabi- 
tants, the remainder being rural. The chief centres of popula- 
tion and their distances above the Montreal intake are as 
follows : 
Pembroke 4,401 220 Miles. 
Renfrew 2,611 190 
Perth 3,136 180 
Smith’s Falls 3,864 175 
Aylmer 1,945 140 
Ottawa (and Hull) 55,429 125 
Buckingham 2,239 100 
Hawkesbury 2,042 60 
Lachute 1,751 50 
St. Anne 1,500 20 
Lachine 3,167 4 
The Ottawa water is dark, and contains a large amount of 
peaty pigment, giving the water, when in a deep column, a tint 
suggesting that of porter. Apart from this it is stated by Prof. 
Ruttan to contain almost no organic matter. It is much softer 
than the St. Lawrence water. 
St. Lawrence River.—The St. Lawrence drains an area of 
510,000 square miles (half as much as the Mssissippi). Its 
discharge, before receiving the Ottawa, has been estimated at 
500,000 cubic feet per second. Apart from the cities and 
towns, situated upon the Great Lakes or on streams draining 
into them, the total population of the towns and villages of over 
1,000, situated upon the river proper, amounts to about 55,000, 
of which Kingston (20,000) is really in Lake Ontario. The 7 
populations and distances above the intake at Montreal are as 
follows : 
Kingston 20,000 185 Miles. 
Gananoque 3,669 150 
‘Clayton 4,400 430 
‘Prescott 2,920 120 
‘Ogdensburg 11,662 120 
Cornwall 6,085 70 
‘Valleyfield 3,315 35 
‘Beauharnois.. 1,590 20 
Towns marked * are on the south side of the river. 
The river averages fully one to two miles in breadth during 
the whole of its course, and expands into Lake St. Louis, 4 to 
7 miles wide, just above the intake, and into Lake St. Francis, 
8 miles wide, 35 miles above. There are rapids at points 20, 
25, 30, 35, and 80 miles above the intake. 
The St. Lawrence water is clear and light green in colour, 
and is fairly hard.* 
In both these rivers the temperature falls to the freezing 
point in winter, even at points near the bed of the stream. 
I. Monthly Examination of Water Supply. 
Microscopical Analysis.—The method employed was, at 
first, that of simply allowing the sediment to settle in a 
conical glass, and by means of a pipette placing a little of 
it under a microscope. This gives a general idea of the con- 
stituents of the sediment, but affords no information as to 
the quantity in which the different organisms are present. 
In the Sedgwick-Rafter method (which unfortunately only 
became known to me after the analysis was completed) a given 
* The following table compiled from Dr. Kuttan’s analyses shows the average 
chemical composition of Ottawa and St. Lawrence water (quantities in part per 
million) : 
Ottawa. 
St. Lawrence. 
' 
O H-» 
Red. 
Color. 
Lovibond scale. 
h-» cn 
o **■ 
Yellow. 
0 04 
0.47 
Blue. 
52 
142 
Total. 
Solids. 
24 
09 
Loss on 
ignition. 
28 
74 
Ash. 
o © 
o b 
w- to 
Free 
Ammonia. 
Nitrogenous 
Matter. 
0.12 
0.09 
Albumenoid 
Ammonia. 
0.03 
0.09 
Nitnates. 
3.7 
0 5 
O 
15 nun. 3 | 
OOr 
o 
© * 
3 CH 
§ 
t—* O'. i ©1 
to | 4 hr?. i 
55 
102 
Hardness (as Cal- 
cium Carbonate ) 
GO H-* 
Cn Cn 
Chloride 
(as Chlorine.) 8 
quantity of the water, usually 500 cc., is filtered through sand 
and the sand with the organisms retained in it shaken up with 
a definite quantity of distilled water, 1 cc. of this is then placed 
in a glass cell, leaving a superficial area of 1,000 square milli- 
metres and a depth of 1 millimetre. By examining under a 
microscope, into the eye piece of which a diaphragm has been 
fitted covering exactly 1 square mm. with the objective 
employed, each microscopic field represents a fixed unit of 
measurement with reference to the original water, and the 
number of each different organism per cc. can be calculated 
from the average number present in each field. As a rule the 
genera only are determined. This method is not applicable for 
determining the number or character of the bacteria. 
During the period from March to November, 1891, the 
presence of the following organisms was noted. The numbers 
represent the number of different genera found in one sample 
and not of individual organisms per c.c. : 
Month 1891. 
Mar. 
Aprl. 
May. 
J une. 
July. 
Aug. 
Sept. 
Oct. 
Nov. 
Sample from : 
Reservoir 
* 
(5 
12 
ltt 
* 
8 
8 
5 
9 
Settling Basin. 
5 
3 
15 
9 
* 
10 
5 
* 
8 
St. Cunegonde. 
4 
4 
31 
18 
13 
11 
5 
5 
12 
St. Lawrence.. 
* 
3 
3 
12 
18 
14 
5 
5 
9 
* Not estimated. 
Of these, the following genera were the most frequent : 
Diatom ace.®. — Acnanthes, Amphora, Asterionella, Cyclotella, 
Diatoma, Encyonema, Epithemia, Fragilaria, Gomphonema, 
M- losira, Navicula, Nilzsehia, Pleurosigma, Stauroneis, Surirella, 
Stephan*'discus, Synedra, Tabellaria. 
Cyanophyce.*.—Anabcena, Oscillaria. 
Other Ai.g.e.—Chara, Cladophora, Ccelosphcerium, Conferva, Cos- 
marium, Palmella, Pleurococeus, Pediastrum, Vaucheria, Volvox, 
Penium, Protococcus, Scenedesmus, Tetraspora, Zygogonium. 
Fungi.—Creno th rix. 
Rhizopoda.—Actinocyclus, Actinophrys, Amitba, Gromia, 
Infusoria.—Bursaria, Can hesium, Dinobryon, Epistylis, Euglena, 
Heteronema, Monas, Paramcecium, Trachelocerca, Trachelomonas. 
Vorticella. 
Spongiaria—Sponge spicules. 
Vermes. Anguileula, Monostylus, Rotifer, Stylonychia. Stentor. 
Crustacea.—Alona, Cyclops, Daphnia. 9 
As I had not been able to employ the quantitative method 
during the year of analysis, I give the results obtained, per c.c., 
from tap water during the period from April 10th to June 4th, 
1892, in the following table : 
Date of examination 
Number of sample 
April 30. 
62 
May 6. 
63 
May 15. 
64 
May 28. 
65 
June 4. 
66 
Diatomace^e. 
64 
H4 
56 
42 
Acnanthes 
2 
0 
0 
0 
0 
Amphora 
3 
pr 
36 
0 
0 
0 
Asterionella 
21 
18 
12 
20 
Cocconeis 
0 
0 
0 
1 
0 
Cyclotella 
pr 
pr. 
2 
0 
1 
Cymbella 
1 
pr. 
0 
0 
0 
Diatoma 
2 
0 
pr. 
0 
0 
Encyonema 
0 
0 
2 
0 
0 
Fragilaria 
pr 
6 
pr. 
2 
0 
Gomphonema 
1 
pr. 
pr. 
0 
0 
Grammatophora 
0 
0 
0 
pr. 
0 
Melosira 
23 
24 
21 
2 
5 
Navicula 
9 
9 
pr. 
1 
300 
Nitzschia 
0 
4 
11 
2 
0 
Surirella 
0 
0 
1 
0 
0 
Synedra 
2 
4 
0 
22 
1 
Tabellaria 
0 
3 
9 
0 
0 
Algae. 
Chlorococcus 
0 
32 
0 
0 
0 
Protococcus 
0 
2 
1 
0 
0 
Zoospores 
pr 
pr. 
pr. 
pr. 
10 
Infusoria. 
Monas 
0 
2 
pr. 
0 
0 
Miscellaneous. 
Starch grains 
3 
2.5 
2 
2 
4.5 
I have omitted from the table the following genera which, though 
occasionally seen, were never present in an amount equal to 0.5 per c.c. : 
—Coscinodiscus, Pleurosigma, Stanroneis, Stephanodiscus, Oscil- 
laria, Arthrodesmus, Cladophora, Ccelosphcerium, Conferva, Pedias- 
trurn, Pleurococcus, Beggiota, Amoeba, Cercomonas, Trachelomonas, 
Spongilla and Cyclops. 
The organisms were more numerous in the warm than in the 
colder months. The higher animal forms being only met with 
during the summer. 
Pollen grains (most commonly from the pine) and vegetable 10 
fibres were usually present in traces, and were most constant in 
the samples from the reservoir. 
From the above results it will be seen that while the waters 
contain small amounts of the non-bacterial organisms common 
to all surface water, these were never found in sufficient 
quantity to affect the odor, taste, or hygienic quality of the 
water. Of the organisms, the diatoms Melosira and Asterionella 
were the only ones occurring constantly in any appreciable 
quantity. 
The green organism (Anaboena) which abounds in the 
water of Lake Ontario and the Bay of Quinte during the 
summer, was scarcely detected at Montreal, though owing 
to the infrequency of the periods of collecting samples it 
may have been missed. Though present in the reservoir during 
August and September very little appeared to enter the sup- 
ply pipes. 
The results of examination of sediments, on the whole, were 
decidedly satisfactory from a hygienic point of view. 
Starch Grains.—The only anomalous features presented by 
the sediments was the constant occurrence of starch grains in 
the sediment of most of the samples. These I first noticed in 
the May samples, they being present in the water from the 
reservoir, settling basin and St. Cunegonde, but not in that 
from the St. Lawrence. 
These grains were usually round or slightly oval, or in some 
cases presented blunted angles. They measured 12 to 30 
microns in diameter, stained blue with iodine solution and 
polarized with a central cross. Some showed a central fissure 
in the form of a slit or cross, and often lamination could be 
distinctly made out. 
I was at first disposed to regard them as an accidental con- 
tamination, due to the entrance of dust into the samples, but 
this was shown not to be case by the fact that upon filtering 
water directly from the tap through glass wool, compressed into 
a small strainer, the starch was invariably detected, while the 
materials employed as well as the glass-ware used, showed no 
signs of it. 11 
Upon consulting the standard works on water analysis, I was 
unable to find any reference to the presence of starch in water 
otherwise than as a consequence of contamination by sewage 
proper, kitchen refuse, or the waste of industrial establishments. 
On the other hand, all the other results of my analysis were 
strongly opposed to the theory of contamination of the water. 
Being myself unable to identify the grains satisfactorily with 
any of the known starches, I consulted Prof. D. P. Penhallow, 
of McGill University, who examined them carefully and called 
my attention to the fact that they corresponded in size and 
shape and structure to corn starch grains, and were much 
larger than any of the starch grains found in aquatic plants. 
He stated that, in his opinion, the only starch bearing aquatic 
plants at all likely to lead to dissemination of starch grains in 
the water were the yellow and white water lillies (Nymphcea 
and Nuphar) the starch grains of which, however, never ex- 
ceeded 13 microns in diameter, and were readily distinguished, 
by their form and arrangement, from the granules under con- 
sideration. 
If the grains were corn starch then they must have come 
from some starch factory or grist mill. 
There were, however, no starch factories or large milling 
industries along the banks of the Ottawa, and though some 
starch factories are situated upon the St. Lawrence, none of the 
grains had been found by me in that water. 
Upon estimating the number of starch grains per cc., I 
obtained the following results, for different seasons of the year, 
from samples of the water which happened to have been 
preserved : 
Month. 
Mar. 
Aprl. 
(May. 
June. 
July. 
Aug. 
Sept. 
Oct. 
Nov. 
Sample. 
Reservoir 
* 
pr. 
* 
2 
* 
* 
* 
3 
2 
Settling Basin. 
0 
* 
* 
pr. 
* 
1 
* 
* 
2 
St. Cunegonde. 
0 
* 
0.8 
4 
4 
* 
* 
5 
2 
St. Lawrence .. 
* 
* 
* 
0 
0 
* 
* 
0 
0 
* Not examined. 12 
The largest amount of starch ever found in any sample was 
7 granules per c.c., in a stagnant rusty sample, obtained from a 
street hydrant. 
The presence of the starch in the Ottawa water and its 
absence from the St. Lawrence, was a matter which completely 
puzzled me. Examination of the starch granules of the sweet 
Fig, 2—Starch grains from water. 
Fig. 3—Starch from white pine bark. 
Fig. 4—Starch from white pine bark 
after soaking in water. 
Fig. 5—Corn starch. 
flag root and wild rice, showed that these grains were altogether 
too small to be thought of as a possible source. 
At this point, Prof. G. P. Girdwood, of McGill University, 
suggested to me the possibility that as starch is present in the 13 
bark of some of the coniferous trees, it might be derived from 
the white pine lumber which, as already stated, is sawn in such 
large quantities as to block the Ottawa river in places with vast 
beds of sawdust. Upon my examining white pine bark, I was 
delighted to find not only that it contained large quantities of 
starch, but that these, though somewhat more angular, closely 
corresponded in size, shape and structure with the grains found 
in the water (and closely resembled corn starch). 
Upon soaking pine bark for two months in water, many of 
the starch grains in it assumed the rounded outline typical of 
the starch of the water sediments, whereas corn starch grains, 
after the same period of maceration, became fissured and tended 
readily to disintegrate upon slight pressure. 
The appearance of the various grains may be better under- 
stood from the accompanying illustrations, figs. 2, 3, 4 and 5. 
Starch grains similar to those of the pine were found, 
though less plentifully in the bark of the cedar, hemlock and 
spruce. • 
The following table gives the diameter in micro-millimeters of 
the various starches examined : 
Diameter 
in microns. 
Water Sediments 
11.4 to 28.0 
8 to 28.0 
5.8 to 27.0 
6.0 to 13.0 
5.7 to 13.0 
1.9 to 7.6 
3.8 to 13.3 
White Pine Bark 
Corn 
Sweet Flag 
Wild Rice 
White Water Lily 
Yellow Water Lily 
There is nothing to show that the starch forms a dangerous 
ingredient of the water. I have also found somewhat similar 
grains under circumstances which did not show any possibility of 
sawdust pollution, and unless great care is exercised one is 
liable to meet with them as a result of contamination of the 
glass-ware, etc., by dust. 
My excuse for giving the above results at such length, is 
that it does not seem to have been recognized as yet that starch 14 
grains may be observed in water independently of sewage or 
industrial pollution on the one hand, and of errors in manipulation 
on the other. 
Bacterial Analysis.—The opinion entertained by chemists of 
the Montreal water supply, at the time when this examination 
was undertaken, is fairly well expressed in Bulletin No. 15 of 
the Inland Revenue Department at Ottawa, which in referring 
to the relatively high proportion of organic matter, speaks of it 
as “ capable of sustaining and nourishing, to a much greater 
degree than in most water supplies, those minute organisms 
which, while in most cases harmless, are closely related to others 
known as disease germs. A water so largely impregnated with 
organic matter, as that of the Ottawa, would become a very 
efficient nidus for the propagation of morbific bacteria were such 
organism to find an entrance to it.”* 
It may be stated in a general way that a pure water should 
not habitually contain large numbers of bacteria. Although no 
hard and fast rule can be set, Miquel’s scale fairly expresses 
our present ideas upon the relation of the number of bacteria 
to the purity of water ; 
Exceptionally pure water contains 0 to 10 per c.c. 
Very pure “ “ 10 to 100 
Pure “ “ 100 to 1,000 
Mediocre “ “ 1,000 to 10,000 
Impure “ “ 10,000 to 100,000 
Very impure “ “ 100,000 and over. 
The number of bacteria in filtered water should not, accord- 
ing to Koch, habitually exceed 100 per c.c. 
I was agreeably surprised to find that the Montreal water, 
instead of teeming with bacteria, was conspicuously free from 
them, as compared with other bodies of running water, so 
that whatever might be the nature of the organic matter 
present it did not appear to be specially favourable to bacterial 
growth. 
The following table shows the average number of bacteria 
found in some well known surface waters, most of which are 
* McGill, Bulletin No. 15, Department of Inland Revenue, Ottawa, 15 
used as sources of drinking water. These marked * are filtered 
before being distributed : 
Locality. 
Authority. 
Ottawa ... 
St. Lawrence 
220 
! Montreal 
.»U ) 
800 St. Louis 
*Danube • 
2,000 Vienna 
Kowalsky. 
33,(too!Above Paris 
‘Thames 
Croton Aqueduct 
19,750 Above London 
4,280 New York 
P. Frankland. 
Health Report. 
Hudson 
3,0061Albany 
Potomac 
Neva 
Rhone 
3,774 Washington 
5,772|St. Petersburg 
75Geneva 
Thos. Smith. 
Poehl. 
Fol. 
Rhine 
20,300[Mulheim 
‘Main 
2,050|Frankfort 
Rosenberg. 
Frank. 
‘Spree 
Number of Bacteria found each month.—The following table 
shows the average number of bacteria per c.c. found each month 
in the reservoir, settling basin, St. Cunegonde and St. Law- 
rence samples : 
Date. 
Tempera- 
ture of 
water °C. 
Level of 
water at 
Lachinein 
feet. 
Reservoir. 
Bacl 
6C . 
.2 c 
co 
4J JO 
eria p 
i . 
S3 <v 
°§ 
&c 
er c.c. 
i a? 
co o 
h-3 S3 
. 
V) 
Combined 
Average. 
December 1st,’90 
4°. 
11 1 
8 
313 
473 
265 
284 
January 5th, ’91. 
0°. 
12 0 
31 
44 
30 
61 
41 
February 2nd... 
0°. 
10 9 
20 
89 
6:1 
29 
50 
March 5th 
0°. 
12 2 
185 
164 
316 
577 
310 
April 13th 
0°. 
13 0 
171 
347 
363 
161 
260 
May 4th 
10°. 9 
15 0 
79 
121 
156 
324 
167 
June 2nd 
13°.0 
13 0 
42 
189 
130 
210 
142 
July 2nd 
18°.3 
11 5 
30 
481 
197 
81 
275 
August 3rd 
21°.0 
115 
92 
119 
101 
85 
99 
September 7th.. 
18°.3 
101 
21 
81 
53 
53 
52 
October 1st 
13°. 1 
101 
40 
55 
29 
43 
42 
November 25th.. 
4°. 
10 5 
143 
1132 
1883 
363 
930 
The following summary shows the maximum, minimum and 
average number of bacteria per c.c. for each sample throughout 
the year, together with the dates upon which the maximum and 16 
minimum numbers occurred, and the total number of samples 
examined from each source : 
Number of 
samples 
examined. 
Source. 
Bacteria per c.c. 
Max. 
Min. 
Average. 
70 
Reservoir 
286 
9 
78 
(Nov.) 
(Feb.) 
67 
Settling Basin ... 
1900 
32 
278 
(Nov.) 
(Oct.) 
73 
St. Cunegonde 
2260 
12 
316 
(Nov.) 
(Oct.) 
71 
St. Lawrence 
600 
18 
189 
(Nov.) 
(Oct.) 
281 
The above tables show that during the greater part of the 
year the number of bacteria per c.c. of the water varies 
between 100 and 200. During the early part of the summer 
and in midwinter this number falls considerably below 100, and 
during the spring and early fall it rises for a short period to 
between 1,000 and 2,000. These temporary elevations coincide 
with a period of heavy rainfall which ushers in the winter, and 
with the melting of the snow in the spring, on both of which 
occasions the river level rises considerably. 
The interval of one month between the taking of samples is 
so great, that the temporary rise in the number of bacteria 
might pass unnoticed, if this sample did not happen to be taken 
exactly at the time when it occurred. Suspecting that this was 
the case in 1891, I made private examinations of the tap water 
at intervals of one week, with the result that a rise to 1940 per 
c.c. (compared with 347 per c.c. in the official sample taken a 
few days before) was observed, the number falling to 117 by 
the time the next official collection became due. The number 
obtained in the official settling basin being 121. 
It is evident that the 12 months covered by the analysis 
comprises the early winter increases in bacteria for both 1890 
and 1891, which makes the average number for the year higher 
than would otherwise be the case. 
This spring contamination of the water was also studied in Jtec.90 'fan,91 Fete. 'Mur. Clpr. .Mat/ Ju/fc Full/ (lift/. Sep. Oct. JVbie. 
Lies? 
Set.Bas. 
S*Ciui. 
StLan: 
lies’- 
Set Has. 
S- Cun . 
S* Law: 
Res? 
Set. Bos. 
S-Ciue. 
SrLan: 
Resv 
Set.Bas. 
S-Cujv. 
St Law: 
Res? 
Set.Bas. 
SSCu/v. 
StLan: 
Res? 
Set.Bas. 
St Cun. 
StLan: 
Res? 
Set.Bas. 
St Cutn. 
SfLan: 
Res? 
Set.Bas. 
St Cun. 
StLan: 
Res? 
Set.Bas. 
St Cun. 
StLan: 
Res? 
Set.Bas. 
St Can. 
StLan: 
Rss? 
Set.Bas. 
SfCu/i. 
StLan: 
Res? 
Set.Bas. 
St Cun. 
St La rr. 
Fig. 6-: 
Diagram showing the results of the monthly examination of water samples, 17 
tap water during April, 1892. The following table shows the 
variation in the number of bacteria : 
Date. 
Bacteria per c.c. 
April 2 
112 
9 
830 
16 
2400 
122 
:to 
4(5 
The two periods characterized by low numbers of bacteria 
(midwinter and early fall) correspond with seasons when the 
level is very low. 
These relations are shown graphically in Fig. 6. 
Although rainfall, when sufficient to produce a marked 
rise in the water level of the rivers, was found to be associated 
with an increased number of bacteria, due no doubt to the 
washings of the soil, no increase was noted corresponding to 
the ordinary local rainfall. 
Reservoir.— One is struck by the marked superiority of the 
reservoir water shown by its small number of bacteria, as com- 
pared with the other samples. During 9 months of the 12, the 
number of bacteria was below 100, while the average number 
was less than one-third of the number found in the settling 
basin. This, apparently, is due to the beneficial effects of sedi- 
mentation, although the reservoir is not well constructed for 
that process (not having separate inlet and outlet pipes), but 
chiefly serves to secure a head of water with constant pressure 
and to form a reserve in case of need. That the reservoir water 
does not deteriorate, and that its quality remains unimpaired 
in spite of a large accumulation of mud and slime at the bottom, 
is a matter which can be readily accounted for. We know now 
that the agencies which produce the series of oxidative and 
nitrifying changes, leading to the self purification of waters, are a 
special class of organisms (nitro-bacteria) which are most 
abundant in that very slime which is generally regarded with so 
much suspicion by the public. To secure the proper perform- 
Comparison of the Four Samples Examined. 18 
ance of this beneficial process, by which the albumenoid and 
ammoniacal bodies, products of pollution, are (perhaps after 
being first decomposed into more readily assimilable forms by 
the agency of the water bacteria) changed into the more stable 
forms of nitrates, it is necessary that there shall be a sufficient 
supply of dissolved oxygen in the water and a sufficient circu- 
lation to promote oxidation and check any tendency to anaerobic 
putrefaction. For this reason shallow reservoirs of 15 to 30 
feet in depth are better than deeper ones. 
Sunlight has been supposed to act powerfully in keeping in 
check any tendency to bacterial overgrowth, but although I 
have not yet been able to practically test the matter, it seems 
probable that the opacity of the Montreal water supply in 
summer would render the effect of sunlight very slight. 
That the improvement which reservoir waters undergo during 
sedimentation is not merely due to a mechanical sinking of the 
bacteria, is shown by the fact that the number found in the 
deeper strata does not show any corresponding increase. This 
was seen in the following observations : 
Depth. 
Bacteria per c.c. 
From 
surface. 
Above 
bottom. 
Max. 
Min. 
Aver- 
age. 
Lower Reservoir.. 
(South Basin).. 
10 
15 
66 
43 
54.3 
No. 564 
Oct. 2, 1891. . 
24 
1 
86 
55 
67.3 
(North Basin).. 
10 
15 
16.0 
No. 564a 
Oct. 2, 1891... 
24 
1 
20 
15 
17.5 
Lower Reservoir.. 
(South Basin).. 
5 
20 
236 
ISO 
203.0 
No. 570 
Nov. 2!), 1891. 
10 
15 
246 
214 
238.0 
20 
5 
248 
96 
172.0 
24 
i 
182 
148 
165.0 
Lake St. John 
(Roberval) 
5 
40 
57 
9 
24.2 
No. 555 
Oct. 7, 1891... 
40 
5 
27 
8 
17.6 
From what we know of nitrification in waters, the ideal bed 
for a reservoir should be coarse sand or gravel rather than of 
bare masonary or cement, but as a matter of fact the natural 
sediment from the water furnishes abundance of the nitrifying 
agent. 
Settling Basin.—This term as applied to the pond at the 
wheel house is a misnomer, as the current is always so rapid as 19 
to allow of very little settling, and, as a matter of fact, the 
number of bacteria found there was never noticeably less than 
that in the aqueduct. From a biological point of view the plan 
suggested by the Superintendent of having a separate channel 
for the water used in obtaining power for pumping, and of 
greatly enlarging the settling basin seems to be an absolute 
necessity. At present the water is pumped into the mains 
with very little settling at all, while only a small proportion of it 
ever passes through the reservoir. I might point out that the 
question of what should be the proper dimensions of the settling 
basin is a biological as well as an engineering one, and a series 
of examinations should be made to find out what amount of 
surface area would be sufficient to secure, by sedimentation, the 
requisite reduction in the number of bacteria*. 
St. Cuneyonde.—The samples from the St. Cunegonde source 
in the Nuns Island Channel showed about the same number of 
bacteria as those from the settling basin, and were decidedly 
inferior in quality to both the reservoir and St. Lawrence water. 
Evidently the theory of the supposed superiority of this water 
arose through a mistaken interpretation of the chemical analyses 
by the Inland Revenue Department, and simply consists in a 
lessened amount of organic matter due to larger dilution by the 
St. Lawrence water. As the organic matter, characteristic of 
the Ottawa water, has beeti shown by Dr. Ruttan to be of the 
nature of a harmless pigment (crenic and apocrenic acids), the 
most exact proportion in which it may be present is a matter of 
indifference from a sanitary point of view. 
That the mere passage over the rapids in anyway improves 
the water by oxidation has never been demonstrated, and as we 
now know that the oxidation of water is not simply a matter of 
aceration, but is due to the action of the nitrifying bacteria, 
there is no longer this theoretical argument in favour of this 
point of supply. 
On the other hand, a special investigation, made jointly by 
* The question of the undesirabie proximity of the garbage depot to the settling 
basin had not arisen at the time when this analysis was made, and I have since had 
no opportunity of investigating the matter. 20 
Dr. Ruttan and myself in July, 1891, brought to light facts 
which show that the intake of the St. Cunegonde supply is not 
very favourably situated. 
The discharge from the tailrace, which empties into the Nuns 
Island channel 150 yards above the St. Cunegonde intake, 
brings with it the contents of the river St. Pierre. This little 
stream receives the drainage of all the land lying to the north of 
the canal between Montreal and Lachine, with the result that 
its water half a mile west of Cote St. Paul was found to contain 
over 18,000 bacteria per cc. A little further on it receives 
the washings of the West End Abattoir. This addition gives 
the water a very offensive character, and I found it to contain 
172,000 bacteria per cc. In examining the tailrace water upon 
several occasions I never failed to detect floating portions of 
offal and animal debris. After receiving the tailrace water this 
number was reduced to 92,500 per cc. owing to the dilution.* 
As the discharge of a large volume of this filthy water at a 
point 450 feet above the St. Cunegonde intake which is situated, 
900 feet from the shore, was so obvious an objection, I made, 
jointly with Dr. Ruttan, an examination of samples obtained on 
July 7th, 1891, at 5 points in the line between the shore and 
the intake in order to see how far out the zone of pollution 
extended. The wind was off shore and its velocity 15 miles per 
hour. The water level was fairly, high in the channel. The 
water close inshore opposite the intake contained 69,000 
bacteria per cc. ; at 100 feet out it contained 669 per cc. ; at 
200 feet out it contained 238 per cc. ; and at 400 feet 157 per 
cc. The number obtained from a sample of tap water at the 
pumping station was 127 per cc. which one would expect in 
pure water. 
The chemical results obtained by Dr. Ruttan showed marked 
pollution inshore and at 100 feet, with slight pollutions at 200 
feet and none at 400 feet, thus corresponding closely with the 
biological result. 
It is evident that on that occasion the zone of pollution 
♦This contamination of the tailrace has no bearing upon the Montreal supply as 
he water only becomes polluted after leaving the settling basin. 21 
ceased between 200 and 400 feet from the shore or 500 and 
700 feet from the intake, and it is unlikely that under ordinary 
conditions the contents of the tailrace enter the St. Cunegonde 
supply. Still, as under altered conditions of the current, water 
level or bed of the river it is not impossible that this may 
occasionally happen, especially when the shallow Hats lying 
inshore are packed with ice. 
It would seem safer to divert the drainage of the St. Pierre 
into the city sewers, though I never found any evidence of such 
pollution in the samples examined. 
I was not able to detect any evidence of pollution from the 
tanneries either in the water or the ice of this locality, but the 
probability that the Verdun shore may soon become densely 
populated is a further objection to the site. 
An interesting point in the analysis was the increase in bac- 
teria, was almost entirely caused by a species apparently iden- 
tical with the colon bacillus. Corresponding with this increase 
there was a falling off in the proportion of the Bacillus fluores- 
cent liquefaciens, which formed from 30 to 40 per cent, of all 
the colonies in the pure water of the river and only 0.5 to 1.0 
per cent, of those in the polluted water of the tailrace. At 
100 feet out the proportion of B. fluorescent liq., rose to 12 per 
cent, at 200 feet to 25 per cent, and at 400 feet to 83 per 
cent. It would seem that any unusual deficiency of the propor- 
tion of this organism to the total colonies during summer should 
be regarded with great suspicion. 
St. Lawrence Water.—The results of the examinations do not 
show that this water is better from a sanitary point of view than 
the present city supply, as far as can be judged from the 
number of bacteria and the nature of the sediment. Although 
informed that the line of the pure St. Lawrence water would 
always be met with at a point 800 feet south of Nun’s Island, I 
have on two occasions seen the Ottawa water extend as far as 
1500 feet south of the island. Of the St. Lawrence water it can 
safely be said that it is a perfectly clean and pure river water. 
One point in favor of the St. Lawrence is that it is far less 22 
affected than the present city supply by temporary pollution 
due to heavy rainfall or melting snow. 
II.—Examination of Local Conditions Affecting the 
Montreal Water Supply. 
Tap Water.—In order to determine whether the water as 
supplied by taps was similar in quality to that of the mains, 
numerous samples were examined during July and August of 
1891. The taps were in all cases allowed to run for at least 
30 minutes before samples were taken and two or more samples 
were always examined, in order to make sure that the number 
obtained was typical for the day. Besides taking samples each 
day from one special tap which was allowed to run continuously, 
I made frequent examinations from taps in various parts of 
the city. 
The tap water was found to contain practically the same 
number of bacteria as the water of the settling basin and, as a 
rule more than that of the reservoir. The number of bacteria 
was found as a rule remarkably constant, irrespective of the 
points from which the samples were obtained. Usually, but not 
always, the taps on the circuit supplied by the upper reservoir 
(the water from which is pumped up from the lower reservoir) 
contained fewer bacteria than those in the lower circuit. I 
have given the results in the following table. 
Number of Bacteria per cc. 
Date. 
Lower Circuit. 
Upper Circuit. 
1891.. 
May 1 
“ 8 
306 
117 
“ 14 
210 
66 
“ 22 
146 
105 
June 23 
50 
48 
“ 30 „ 
30 
22 
Comparison op Upper and Lower Circuit 
As far as it goes this supports the view that the water is 
improved by standing in the reservoir. 
During July the daily examination showed for the upper cir- 23 
cuit a maximum number of 136 bacteria per cc. and a minimum 
of 28, the average being 68. During August the maximum was 
160 per cc. the minimum 17 and the average 55. 
A comparison was made of the water from the lower and 
upper circuits with the following results. 
Lower 
Circuit. 
Upper Circuit. 
Reservoir. 
Taps. 
Reservoir. 
Taps. 
Sept. 23 
39 
37 
41 
50 
Occ. 2 
53 
49 
29 
54 
Although this shows relatively slightly more bacteria in the 
upper than the lower circuits, the difference is not large enough 
to be outside the limits of experimental error. 
Aqueduct.—Two examinations of samples taken at 5 points 
along the aqueduct gave : 
Aug. 7. 
Sept. 12. 
Maximum 
173 
102 
Minimum 
93 
38 
Settling Basin 
224 
113 
Lachine Intake 
115 
80 
The variation is not sufficient to show any material change in 
the water during its passage from Lachine. 
Dead Eads.—In districts where the circulation in the mains 
is not complete complaints are often made of turbidity of the 
water. This turbidity appears to be due to rust from the mains 
but. as the consumers are inclined to consider this condition as 
unwholesome, I made on Aug. 24th, 1891, an examination of 
the water from 11 different districts supplied from dead ends. 
The average number of bacteria found per cc. was 94, and 
therefore such as to exclude any idea of a polluted or stagnant 
state of the water. The vital statistics from the streets supplied 
by dead ends do not show any greater frequency of typhoid than 
other parts of the city. Iron rust is, as we know used as a 
means of precipitant for freeing water of organic matter. 24 
III.—Study of the River Water at Points Above 
Montreal. 
Ottawa Water.—In order to study the influence o( the towns 
along the course of the river upon the character of the water, 
two sets of examinations were made in 1891, one on July 3rd, 
and the other on Sept. 24th. Samples were collected from the 
bow of a steamboat by means of a fishing rod and line to which 
small weighted bottles were attached, and the cultures made 
immediately. Duplicate samples were taken at 15 points on 
each trip, and a sample was also obtained from lake Des Chenes, 
10 miles above Ottawa. Owing to an accident, several of the 
cultures made during the first trip could not be made use of. 
The results obtained are given in the following table together 
with the distances below Ottawa. 
Distances below 
Ottawa. 
Bacteria per 
cc. 
Above Ottawa (C. 1\R. Bridge) 
0 Miles. 
170 
Gatineau 
5 
686 
Cumberland 
30 
1530 
Grenville 
66 “ 
48 
Carillon 
80 
60 
*Como 
IK) 
72 
*St. Anne 
100 
11 
Lynch’s Id 
120 
49 
*In lake of Two Mountains. 
These are shown graphically in Fig. 6. 
This showed a marked increase in the number of bacteria 
below the city of Ottawa, diminishing to the normal for river 
water by Grenville and reaching a minimum in the lake of Two 
Mountains, and increasing slightly in the river channel below 
St. Anne. None of the smaller towns appeared to have any 
perceptible pollutory effect on the water. 
In the second test on Sept. 24th and 25th, a much more 
thorough examination was obtained, but the results corresponded 
to a remarkable extent with those of the former examination. I 
have given the table in full in order to show the measure in 
which samples taken from the same points on two succeed- 25 
ing days resembled one another in regard to the number of 
bacteria: 
Up Trip 
Down Trip. 
Combined 
Locality. 
Average 
Max. 
Min. 
Aver. 
Max. 
Min. 
Aver. 
of both 
trips. 
0 
4 
5.2 
Mil’s below Ott’a. 
2 
5,80 
878 
479 
365 
250 
307 
393 
5 
528 
5(H) 
5(H) 
732 
329 
582 
520 
10 
400 
272 
877 
128 
140 
20 Cumberland. 
172 
180 
147 
170 
139 
152 
149 
80 Thurso 
814 
184 
257 
172 
72 
122 
204 
50 Montebello.. 
47 
40 
40 
45 
24 
34 
40 
00 L’Orignal.... 
40 
20 
88 
45 
24 
84 
33 
05 Grenville.... 
20 
19 
28 
80 Carillon .... 
87 
84 
30 
38 
26 
32 
34 
90 Como .. ) 
18 
10 
14 
28 
16 
22 
17 
92 Oka .... V 2s 
17 
8 
5 
0 
12 
100 St Anne 1 jsj § 
10 
0 
8 
16 
0 
8 
8 
105 Lynch's Id. 
21 
15 
18 
22 
14 
18 
18 
120 Lachine 
18 
8 
12 
This is shown graphically in Figs. 7 and 8. 
A point to which Dr. Ruttan was the first to call attention is 
that the thickly settled agricultural district composed by the 
counties of Pembroke and Russell, having a population of about 
100,000, drains into the Ottawa. An examination of water of 
one of the large streams for this district, the South Nation 
river, was made by us in May 1892, but no evidences of pollu- 
tion were detected. 
From this it is evident that any pollution due to the Ottawa 
or other sewage is effectually got rid of long before it reaches 
Montreal. The greatest improvement apparently takes place 
in the Lake of Two Mountains the bacteria being much fewer 
at the lower than the upper end. Attention may also be called 
to the fact that the number of bacteria in the water of the Lake 
of Two Mountain is lower than that of the present Montreal 
supply which on Sept. 23rd, gave 30 to 49 per cc. 
That St. Anne, and upper Lachine with the intervening 
population along the banks of the St. Lawrence do not form a 26 
possible source of infection for the Montreal water supply is by 
no means clear as our water is taken from the portion which flows 
by the north bank of the St. Lawrence. It seems advisable 
that a careful sanitary inspection of this district should be made 
Fig. 7—Diagram showing the condition of Ottawa water above Montreal 
(First examination, July 30th, 1891.) 
Fig. 8—Diagram showing the condition of Ottawa water above Montreal 
(Second examination, Sept. 24th, 1801.) 
and a map prepared showing the position of all privies, barns, 
etc., in order that any possible source of infection should be 
eliminated. The key to the safety of the Montreal drinking 
supply may be said to lie between St. Anne and the intake. 27 
The entrance to the intake is confidingly placed so as to catch 
all washings from the adjacent portions of the lower Lachine 
road. 
Question of typhoid infection.—The following inquiry into 
the possibility of water-borne typhoid in connection with the 
Montreal water may be of interest: 
Comparison of freqiwnci/ o/' Vi//j/toti/ 
at OttanxL a/a/ . Mo/it/rat 
Fig. a 
The number of cases reported each month at the health office 
are shown in figure 6. It is evident that if the frequency of 
typhoid fever depended upon general contamination of the water 
supply, it would, allowing for the period of incubation, be ex- 
pected to appear in the month following the greatest contamina- 
tion of the water. As no increase in typhoid occurred in these 28 
months, that disease being most prevalent when the number of 
bacteria in the water reached its lowest point, it is evident that 
the turbidity and increase in bacteria which periodically affects 
the Montreal supply is not of such a nature as to cause or predis- 
pose to typhoid infection. In this connection it was interesting 
to see if there was any relation between the frequency of 
typhoid at Ottawa and at Montreal. Unfortunately non-fatal 
cases of typhoid are not reported to the Ottawa health office, 
and it is a well-known fact that less than half of the cases are 
reported at the Montreal office. As the deaths from typhoid 
are reported however, I have taken these as my basis, calculating 
the mortality at 10 per cent. As shown by figure 9 there is 
not only no constant relation between the frequency of typhoid 
in the two cities, but that even the severe epidemic of typhoid 
at Ottawa in 1888, was not accompanied by any increase in the 
number of cases in Montreal. 
It appears therefore that general pollution of the Montreal 
water as may occur is probably of a harmless nature and does 
not form a source of infection. 
St. Lawrence above Montreal.—A double series of observa- 
tions was made in the same manner as in the case of Ottawa, 
the samples being taken on July 27th, 1891, between Brockville 
and Lachine, and on Sept. 30th, 1891, between Kingston and 
Lachine. 
The results with the distances above the Montreal intake at 
which the samples were taken, are shown in the following table 
and in figures 10 and 11 : 
Sample from 
Distance 
above 
intake. 
Bacteria per cc. 
, J 
Max. 
uly 27t 
Min. 
h—^ 
Aver- 
age. 
—Se 
Max. 
pt. 30 
Min. 
th 
Aver- 
age. 
Lake Ontario near 
lit!) miles. 
29 
10 
22 
Long Point 
180 “ 
48 
49 
Clayton 
175 “ 
33 
29 
Brockville 
125 “ 
70 
14 
44 
Galop Rap'd 
98 “ 
38 
31 
37 
18 
Head of Long Sanlt 
75 “ 
210 
70 
121 
70 
Foot of Long Sault 
08 “ 
150 
141 
151 
Cornwall 
05 “ 
155 
90 
130 
72 
50 
03 
Coteau 
35 “ 
77 
20 
47 
15 
10 
12 
Caughnawaga 
2 “ 
74 
49 
01 
33 
20 
27 29 
This examination showed an interesting increase in the 
number of bacteria on both occasions in the swift and relatively 
shallow stretch of river below Prescott, the number falling again 
in Lake St. Francis and rising somewhat below Lake St. Louis. 
Fig. 10—Diagram showing condition of St. Lawrence water above 
Montreal. (First examination, July 26th, 1869. 
Fig. 11 Diagram showing condition of St. Lawrence water above 
Montreal. (Second examination, October 1st, 1891.) 
It also shows throughout a relatively smaller number of bacteria 
in the water in September than in July. 
IV. Surface Waters in other Parts of Canada. 
Water from Uninhabited Districts.—A number of examina- 
tions made for the purpose of comparing the Ottawa water with 30 
similar peaty waters for uninhabited districts, may be briefly 
recorded here by means of the following table, which shows that 
the water of the large rivers of the far north, coming from a 
desolate and almost unexplored country, contain as many bac- 
teria as the Montreal water supply. It must be mentioned 
however that some of the samples were taken during a period of 
heavy rainfall late in the autumn. 
Date. 
Sample. 
Bacteria per cc. 
Max. 
Min. 
Aver- 
age. 
Tempera- 
ture of 
water °C. 
1891. 
Aug. 30 
Saguenay above Chicoutimi 
70 
41 
56 
18° 
Oct. 7 
Ouiatchouan 
134 
101 
118 
12° 
“ 
Ashuap-Mouchuan 
7(K) 
too 
470 
10° 
“ 
Mistassini 
094 
400 
474 
10° 
Other Canadian Water Supplies.—Finally it seems of some 
interest (in view of the scanty data available on the subject) 
to mention some analysis of other Canadian water supplies which 
I made during the summer of 1891, though the fact that these 
waters were not repeatedly examined makes it impossible to 
draw any definite conclusion as to their relative sanitary value. 
In each case several different samples were taken and the cul- 
tures were, in every case made upon the spot. 
Locality. 
Date 
Bacteria per cc. 
Number of 
samples. 
Max. 
Min. 
Aver- 
age. 
Kingston, Out 
Sept, mil, 1891. 
8 
99 
48 
65 
Quebec, Q 
Oct. 7th, “ . 
5 
112 
86 
90 
Sherbrooke, Q 
July “ . 
7 
263 
85 
212 
Halifax, N.S. 
July 17th, 1892. 
7 
218 
41 
99 
I mention these results partly in order to emphasize the fact 
that for a reliable analysis the water must be repeatedly exam- 
ined and samples obtained at different seasons. In a recently 
published biological analysis of 21 Canadian water supplies,* 
made in the spring of 1894 very different results were obtained, 
*E. B. Shuttleworth. Toronto Telegram, May 10th. 94. 31 
he number of bacteria found in the Quebec water, for example, 
being stated as 104o per cc. whereas it only contained 90 per 
c.c. on the occasion when I examined it. 
Description of Species of Bacteria Found—Qualitative 
Bacterial Analysis. 
From a sanitary point of view the most pressing questions in 
connection with my analysis of the Montreal water supply were 
those bearing upon the possibility of pollution. This can be 
determined as a rule better by quantitative than qualitative 
work, and I found that the study of the problems of this nature 
took up so much time that very little was left for the determin- 
ing of the species of bacteria present. 
Although 1 isolated over fifty different forms from the water, 
I was not able in all cases to study them thoroughly enough to 
identify them with existing species described by others. This 
identification is a matter of extreme difficulty except in the case 
of well known and easily recognized forms, and the difficulty 
is increased rather than diminished by the fact that some 
workers have published as new species forms which were already 
described, or described them in so vague and unsatisfactory a 
manner that it is impossible, from the meagre details given, 
to tell whether they are new or not. Microphotography does 
not seem to have greatly helped matters, and with the pigment 
producing (chromogenic) bacteria, it is impossible to tell from 
the descriptions given what the shade of colour really is and in 
how much its tint depends upon the medium employed. 
For these reasons, although I detected and described several 
forms which I consider to be new species, I have hesitated to 
publish them for fear of adding to the existing confusion. 
It seems to me that the great tendency which these organ 
isms have to form varieties and races makes it really of less 
importance except in the case of pathogenic forms to multiply 
the number of new species by emphasizing minute points 
of difference than to study their points of resemblance, and so 
form them into definite groups in which, while the members 
might differ slightly from one another, their main characteristics 32 
would enable them to be distinguished from the members of 
other groups. In other words, I would suggest the study of the 
affinities as well as the differences of the organisms. In that 
case, even if one might not be quite certain if the organism was 
a new species, he would know approximately where to place it. 
It may, perhaps, not be out of place to quote the following 
passage from my report made in 1891 (though no doubt the 
idea has occurred to others) : 
“ The result of the large number of disjointed efforts made 
in the direction of systematic description of the water bacteria 
makes it clear that the matter can never be settled on paper or 
by the isolated observations of individuals. What is wanted 
seems to be more co-operation among those working on the sub- 
ject. This would lead to a sounder basis of classification of the 
water flora and seems really to be the only feasible means of 
attaining that end. If a society or committee of those engaged 
in water analyses in different localities could be formed, and each 
member allotted one group to investigate, so that various organ- 
isms of the same group obtained from different localities could be 
compared by parallel cultures, the results when compared and 
published would soon form a recognized standard of comparison. 
This would not only help beginners, but would obviate to some 
extent the causes which tend to confuse the work.” 
Not feeling myself competent to take the lead in a project of 
this sort, I refrained from taking any steps in the matter, but 
it may be mentioned that with the co-operation of Professor 
Adami, of Montreal, an attempt is now being made to organize 
somewhat upon the lines just laid down a scheme for the co- 
operative study of the water bacteria. 
Bacteria Found in Montreal Water.—The forms which oc- 
curred were almost exclusively bacilli, only two species of 
micrococci being met with. 
During the pollution due to heavy rains and melting snow a 
considerable number of molds, were present. A form of 
Fusarium was once detected in tap water. 
The relative proportions of the species present often gave 
valuable indication of slight degrees of pollution when the total 33 
number of colonies was not sufficient to attract attention * and 
certain species, notably B. mycoides, appeared when the water 
was exposed to the washings of cultivated land. 
As a rule 5 to 7 forms were detected in each sample when 
the water was pure, while in impure samples from the Montreal 
harbour, I have isolated as many as 16 species from the sample. 
On the other hand when a large number of bacteria developed 
in stored waters which were pure nearly 90 per cent of the 
colonies would belong to one species, usually B. Fluorescent 
liquefaciens, and if during the summer the proportion of this or- 
ganism (which was normally from 30 to 40 per cent of the total 
colonies) fell to below 12 per cent, other proofs of pollution 
were usually forthcoming. 
A singular circumstance was that in winter this ratio fell to 
5 or 10 per cent., although the water was pure, the proportion 
suddenly rising again when the warm weather returned, while 
B. Aquatilis and other members of the yellow pigment- 
forming group formed the leading flora during winter. This 
transition is shown in the following table of analyses : 
No. 
Sample. 
Date. 
Temp. 
of 
water. 
Bacteria 
per c.c. 
O 
0Q 
. 
£ cr 
m 
j£ 
cr 
< 
m 
128 
Tap 
Tap 
May (5, 1891. 
6° 
646 
4 
•30 
129 
May 7 
6° 
860 
8 
35 
14(5 
Tap 
May 14 .... 
10° 
106 
15 
20 
1451 
Reservoir 
May 14 . . .. 
13° 
106 
15 
20 
140 
Basin 
May 13 .... 
115 
14(5 
12 
20 
138 
St. Ounegonde 
May 13 .... 
10° 
189 
15 
15 
131 
St. Lawrence 
May 13 .... 
9° 
245 
10 
25 
165 
Tap 
June 4. ... 
11°5 
140 
40 
2 
15(5 
St. Lawrence 
June 5 
14° 
123 
30 
4 
The following were among the common forms met with in 
Montreal water : 
B. arhoresccns, B. aquatilis, B. flourescens, B. fluores- 
cent liquefaciens, B. janthenus, B. glaucus, B. megatherium, 
* See page 21. 34 
B. multipediculus, B. mycoides, B. nacreosus, B. aurantiacus 
B. ramosus, B. aquatilis sulcatus, B. mesentricus vulgatus, 
B. mesentericus fuscus, B. proteus, B. fulvus, B. fuscus, 
B. ochreceus, B. plicatus, B. implexus, B. ruber. 
Among the rare forms may be mentioned B. Berolinensis, of 
which one single colony was met with. 
Spirilla were not detected, but it must be mentioned that the 
plan of cultivating in weak peptone solutions was not known at 
the time. 
I was able by means of the Parietti and P6r6 methods to 
isolate forms apparently belonging to colon group, but never 
succeeded with Montreal water in finding a perfectly typical 
distinctive culture of B. Coli or B. Typhi, whereas I found these 
present in some spring water at a village where typhoid was 
epidemic. 
CONCLUSIONS. 
From the result of this analysis it appears : 
1. That the Montreal water is of good quality, as compared 
with other surface water, and does not appear to be at present 
a source of danger to public health, though its future purity 
is not altogether assured. 
2. That from a biological point of view the St. Lawrence 
water is not superior to that of the Ottawa. 
3. That the St. Cunegonde site offers no advantages. 
4. That the reservoir water is superior to that of the settling 
basin, and that the Ottawa water is well adapted for storage in 
open reservoirs. 
5. That better facilities for settling water should be provided? 
and no water pumped into the mains without previous sedi- 
mentation. 
Inasmuch as the extreme severity of the winter makes 
the employment of filtration impracticable, it is necessary 
to watch very carefully any minor sources of possible 
especially those lying between the settling basin and St. Anne. 
It would be advisable to make experimental studies upon 
the dimensions and capacity of a suitable settling basin, and 
also to make arrangements by which the water could be 35 
examined regularly at weekly or fortnightly intervals in order 
that any variations from the usual standard of purity estab- 
lished by these analyses may be promptly investigated. It 
would also be well to see if a better quality of water could 
not be obtained from the Lake of Two Mountains, and to inves- 
tigate the amount and quality of water available in the lakes in 
the Laurentian Mountains, lying to the north, in case of a 
change of supply becoming necessary in the future. 
I have to record my thanks to Mr. A. Davis, Superintendent 
of the Montreal Water Works, for having obtained permission 
to publish the foregoing report, and also for kindly loaning the 
cuts which illustrate it.