direct determination can be made only with the greatest of difficulty, and the results of such correct determination have only a doubtful value. A certain number of the bacteria producing water-borne diseases, such as the typhoid bacillus and the cholera spirillum, have been successfully isolated and can occasionally be detected in a polluted water. Broadly speaking, however, it is not at all certain that such bacteria will be detected, even if they are present in a public water supply.

A few investigators, among whom may be particularly mentioned Dr. A. C. Houston, Director of Water Examinations, Metropolitan Water Board, London, have insisted that their methods of determination of these specific organisms are sufficiently accurate and reliable, so that tests made by them showing that these bacteria are not present are conclusive to them that there is no danger of such diseases being caused by the water. Most of the laboratory workers do not concur in this viewpoint. They assert that, while specific bacteria may often be isolated and shown to be present in the water, they are not at all willing to say, on the other hand, that the fact that they are unable to isolate this organism shows conclusively that they are not present. They believe that the methods used are sufficiently uncertain so that typhoid or cholera bacteria may be present and still not be detected.

There is one other important factor which must be considered. We do not know with exact certainty what diseases are disseminated by water. A certain number certainly can be attributed at times to a water supply. Others are uncertain. There may be still others which may be carried by water, but we are not fully aware that they can be so carried. To make satisfactory tests of a water supply, therefore, by the direct method, would require specific tests for each particular type of pathogenic bacteria which we know can be carried by water or which possibly might be carried by water. Quite apart from the accuracy and desirability of results of such methods, the work would probably become so laborious and time-consuming for practical use in ordinary water determinations that the method falls of itself as too unwieldy to be used.

In effect, any such direct methods, ideal as they may be, fail us in service, and we must have recourse to some indirect

method which will give us better service. The indirect methods available are, first, the determination of the total number of bacteria present in a given volume of the water, and, second, the determination of the absence or quantitative presence of certain specific bacteria which may be called indicators of the possibility of the presence of disease germs.

SIGNIFICANCE OF TOTAL BACTERIAL COUNT.

The oldest and simplest biological method of gauging the quality of a water supply is by counting the total number of bacteria by means of the colonies formed by growth on some solid nutrient medium. The method formerly used and still in most general practice is to count the total number of colonies produced by incubation on solid gelatine in forty-eight hours at a temperature of 20° Centigrade, and to call this the number of bacteria present in the quantity of water examined (one cubic centimetre). By this standard, German sanitarians about twenty-five years ago fixed on one hundred bacteria per cubic centimetre as the fair dividing line between a good and an unsatisfactory water, under their local conditions as studied. Any water showing less than these one hundred bacteria per cubic centimetre was considered satisfactory, but a bacterial count in excess of this amount showed a water of inferior quality. This standard has been very much used and until fairly recent times has been considered a reasonable and proper standard for some classes of water.

A number of rather obvious objections present themselves to the use of this standard for determining the water quality. The first natural objection is that there may be, and probably is, no definite relationship between the number of bacteria present in the water and the number of pathogenic bacteria which may cause disease. The bacteria in the water originate from any number of points. They come from the soil, the vegetation, and from animals; they grow in the water itself, and many forms multiply very rapidly in the water. A pure water may show very high bacterial results, and a polluted water a relatively small bacterial count.

In considering any new source of water without knowing its special local conditions, it is almost impossible to say from

the total bacterial count alone whether the water is good or poor or something intermediate. It even happens that in carefully filtering water the filtered effluent may show as many as, or even more bacteria than, the water applied to the filters. This comes from the fact that certain species of bacteria may grow in the filter itself, particularly in its underdrains, and from this growth of water bacteria, always of a harmless form, the effluent obtains. its large bacterial count. And yet this filtered effluent may be entirely free of any pathogenic forms.

Another factor militating against the useful purpose of this total bacterial count test is that the incubation was effected at a temperature of 20° Centigrade, about normal room temperature. Under such conditions the growth of the harmless water bacteria is facilitated, while the growth of the pathogenic bacteria, which naturally thrive at the body temperature, is partially inhibited. It has been attempted to meet this difficulty by specifying a different total bacterial count method of testing. This involves the incubation of the sample at a temperature of 37° Centigrade, using nutrient agar as the medium. The use of the high temperature, approximately that of the body, serves the purpose of encouraging the growth of various bacteria which thrive at the body temperature, and of inhibiting the growth of the ordinary non-pathogenic bacteria. This gives relatively small counts on the plate, but these bacteria, of smaller number, perhaps, may be a fairer measure of the possible presence of contaminating organisms.

Table I of "Comparison of Gelatine and Agar Counts for Various Coli Conditions" exhibits, opposite to each other, the bacterial counts per cubic centimetre obtained by the two methods above cited. It will be seen that the bacteria shown by the 37° agar test are very much lower than those shown by the 20° gelatine test. There is, however, no constant relation between the two. In some cases the agar may show only ten per cent. of the bacteria of the gelatine test. In other cases the count may be twenty-five per cent. or more of those shown by the gelatine test. It is, of course, not to be expected that there should be any constant relation, or even approximately constant relation, between these two tests. In addition, different waters and different conditions of the same water, even different seasons of the year will show a different relation in the count by the two tests.

TABLE I.

COMPARISON OF GELATINE AND AGAR COUNTS PER CUBIC CENTIMETRE FOR VARIOUS COLI CONDITIONS.

Counts shown here are the average for sixteen months

Reference-W. U. C. Baton, Journal Am. W. W. Assn., 1914, Vol. No. 1-p. 19.

An instructive series of investigation has recently been made by Dr. Ernst Quantz, at Göttingen, in Germany, and published in the Zeitschrift für Hygiene, 1914, vol. 78, p. 193. Table II shows the results of this investigation made entirely with various well waters. The well waters are divided by the author into three classes: Group I, as shown in the table, consists of wells which are practically impossible for pollution, and may be considered as safely good well waters. Group II are rather doubtful, and at times might possibly be polluted and at other times be pure. Group III are those classed by the author as particularly bad wells, and pretty nearly certain to be polluted.

TABLE II.

BACTERIAL Count of Well Waters.

By Dr. Ernst. Quantz-Zeit. f. Hyg.-Vol. 78-Page 193.

In examining these data a very general relationship is shown. Thus, taking 1000 bacteria per cubic centimetre as the limit, we see in Group I that 88 per cent. are under this limit; in Group II, 62 per cent.; and in Group III, 34 per cent. Taking 100 bacteria as the limit, in Group I, 31 per cent. are under this limit; in Group II, 18 per cent.; and in Group III, no per cent. This indicates very fairly that better wells, on an average, have a lower bacterial count than the poorer wells. The author's bacterial count was taken on gelatine at 20° Centigrade.

Individual wells, however, show no such relationship. In some cases good wells show a higher bacterial count than poor wells. Very often the rate of draft on the wells is a bigger factor in the number of bacteria shown than the possibility of pollution. Good wells which have been standing for some time show a very high bacterial count.

The most that can be deduced is that very low counts mean certainly very safe water and very high ones possibly unsafe

ones.

As the conclusion of some thirty years of experience, bacterial counts are still taken as showing roughly the variation in the same water. In other words, their chief advantage is found in comparing a water before and after it is filtered. Especially is this true of counts on gelatine at 20° Centigrade.

When taken alone bacterial counts are not considered ordinarily as a reliable indicator of the safety of the water.

SIGNIFICANCE OF B. COLI AS AN INDICATOR OF POLLUTION.

After eliminating as a satisfactory indicator of pollution the total count of bacteria which may originate anywhere at all, the next natural step is to restrict our investigation to those particular bacteria which seem likely to originate in the human body. The normal rule seems to be that only human diseases are transmitted to other human beings, and that bacteria, in order to cause infection, must originate in man. Infectious pollution in water supplies arises almost wholly from discharges from the human intestines, and any bacteria which can be assumed to originate in the human intestines would be a fair measure of such human pollution which potentially might be dangerous or infectious.