nation the same principle is utilised in the estimation of the number of bacteria in a given volume of water; it is carried out in the following manner :-About a hundred test-tubes, each containing 10 c.c. of sterile broth, are prepared; into the first tube, A,, 1 c.c. of the water under examination is introduced by means of a pipette. Into a second broth tube, A., 1 c.c. is introduced from tube A,; A, will then contain 11 c.c. of liquid, which is equally divided into a series of eleven broth tubes, marked respectively A,, A'3, A23, &c. The contents of A,, containing, like A2, 11 c.c. of liquid, is then equally divided amongst eleven more tubes marked A4, A4, &c. The number of such series of dilutions. that must be prepared will depend upon the number of microbes supposed to be present in the water, the object being to ultimately obtain a series of tubes, each of which shall not receive more than one microbe.

49

All these broth-tubes, with the exception of A, and A, (the whole contents of which were divided amongst the two series of tubes A13, A23, and A', A2, .), are incubated for some days, or even weeks, at from 30° to 35° C. If all the tubes subsequently exhibit turbidity, it shows that the dilution has not been carried far enough, and the process must be repeated; or, if that is impossible (which, as regards water, must be the case, for the sample, after standing days or weeks, is rendered absolutely worthless through the multiplication of the water-bacteria in the interval), a fresh sample must be collected and the process repeated de novo.

If, on the other hand, in one of the several series of dilutions described above some tubes become turbid, whilst others remain clear, it is argued that the turbid ones have received only a single microbe apiece; and if this be granted, it is obvious that from the number of

39

turbid tubes the number of microbes in the volume of water represented by the particular dilution-series can be inferred. Thus the series of tubes A13, A23, A33, A10, has received in all 1 c.c. of the original water, and, to take a simple case, if three out of these ten tubes became turbid, the other seven remaining clear, it would be inferred that 1 c.c. of the water contained three microbes, or thirty microbes in 1·0 c.c.

It is obvious that this undoubtedly ingenious method possesses a number of grave disadvantages; thus it is not only exceedingly laborious, but the possibility of failure is very considerable, unless skill and judgment are employed in arranging the necessary degree of dilution required by the particular sample of water under examination. Moreover, it affords little or no immediate indication as to the particular varieties of microbes present in the water. On the other hand, some real advantages attach to this method, viz.: First, that the cultivation is made in a liquid medium in which some bacteria will thrive which would not develop on gelatine-plates; and second, that the incubation can be effected at any desired temperature, whilst with gelatine-plates the temperature of incubation cannot exceed about 22° C.; on this account it is in exceptional cases necessary to employ this dilution method for the examination of water.

In actual practice Miquel has more recently adopted what he calls the 'mixed process.' The water is diluted to, say, 100, 1,000, 10,000, or 100,000, &c. times its volume, according as the particular sample is supposed to contain a smaller or larger number of bacteria. After this has been done, from one to two drops are introduced into a flask with a large flat bottom (about 9 centimetres in diameter) containing a layer, about 2 millimetres in thickness, of sterile and liquefied

gelatine-peptone. The whole is then gently agitated, allowed to solidify, and incubated at 20-22° C., and the resulting colonies counted and examined in the usual way. Miquel uses about a dozen such flasks for each water examination. The number of bacteria originally present in the sample can then be calculated from the number of colonies which make their appearance in the gelatine-films of these flasks. Excepting in the use of the cumbrous and otherwise inconvenient conical flask in question, this so-called mixed process of Miquel's does not differ in any single detail from the ordinary method of plate-culture as commonly practised, for it should be pointed out that preliminary dilution before plate-cultivation must invariably be resorted to in the case of all waters which, like sewage, polluted streams, &c., are very rich in bacterial life. Such preliminary dilution is best made with sterilised natural water, and not with sterilised distilled water, as the latter is liable to prejudicially affect some bac

teria.

NUMERICAL DETERMINATION OF BACTERIA IN WATER

BY GELATINE-CULTURES 1

The method of pouring gelatine-plates has already been given (see p. 30), and it only remains here to describe the process as applied to water examinations. Gelatine-plates, small round covered dishes, or Esmarchtubes, may all be employed for this purpose. The frequent presence of microbes in water which liquefy the gelatine renders the Esmarch-tubes less serviceable than

1 For the qualitative determination of the bacteria present in any given water, besides examining the colonies on the gelatine-plates with a low power under the microscope (see p. 35) and inoculating particular colonies into gelatine-tubes &c., recourse must be had to the special methods described on pp. 267, 276, when typhoid or cholera bacteria are suspected of being present.

the gelatine-plates, whilst most convenient of all are the covered shallow dishes (commonly called Petri-dishes) already referred to (see p. 34). After the ice-plate has been arranged, and all the necessary preparations made for the pouring of plates, the sample of water is taken. and violently shaken for several minutes, in order to disintegrate any aggregations of microbes, as well as to secure even distribution of the bacteria throughout the liquid. A definite quantity of the water is now removed by means of a sterilised and graduated pipette, and introduced into the gelatine-tube, which, during the operation, should be held in a slanting position. The cotton-wool stopper is then replaced and the contents gently agitated, and the stopper removed with all the precautions already described, and the plate poured in the usual manner. Supposing the water to be fairly pure as regards microbes, 1 c.c. may be taken for one plate, and 5 c.c. for a duplicate. In all cases at least two plates must be poured of each sample of water. If the water is suspected of containing a large number of bacteria, then it will be necessary to dilute, say, 1 c.c. of it 50 or 100 or 500 times, as the case may be, before pouring the plates. For this purpose a small sterile stoppered bottle containing, say, 50 c.c. of sterilised. natural water (not distilled water) may receive 1 c.c. of the original water. After thorough shaking, 1 c.c. from this bottle may be introduced into another similar bottle, and so on, until the attenuation is considered sufficient; plates may then be poured from the two last attenuations. As in all such manipulations success can only be attained after practice and much experience. The plates are incubated in the usual manner, and the counting of the colonies is conveniently carried out with the assistance of a counting apparatus (fig. 13).

This consists of a wooden stand A on which is sup

ported at a a a a the transparent glass plate B. The middle portion C of this glass plate is etched out into squares, some of which, situated in various parts of the field, are further divided up into nine smaller squares. When the gelatine-plate is ready to be counted, B is raised, the gelatine-plate so placed that it is covered by the etched surface C, and a small magnifying glass resting on three

D

feet is placed on C, and the colonies enclosed by each square counted. If the colonies on the plate are too numerous to count individually, an approximate estimate may be made of their numbers by counting those contained in a few of the large squares (which is most. accurately carried out by using the squares divided into the nine smaller squares) and then multiplying the average number on these squares by the total number of squares over which the gelatine-film extends; a very accurate result may be thus obtained if the water has been well mixed with the gelatine in the first instance. The number of colonies found is then calculated on 1 c.c. of the original water.

By using glass dishes 1 instead of plates there is much less risk of aërial contamination, whilst by introducing the water direct from the pipette into the melted. gelatine in the dish the loss is avoided of those bacteria which must necessarily remain in the gelatine left ad

1 See note, p. 34.