tine, kept fluid at about 35° C., and the mixture, after solidification in a thin layer, is incubated generally at a temperature of 20-22° C. in contact with air, but protected from danger of contamination: we need not go into the particular methods of sterilisation, protection, incubation, &c.; suffice it to say that in a few hours or days colonies of bacteria appear on the culture plates, and the number of individual bacteria in the measured sample of water is estimated from these, on the assumption that each colony has sprung from one germ.

The comparison of numerous researches made in recent years, and the experience gradually being gained in all branches of the technique of the subject, have slowly led to the detection of numerous fallacies in the almost established mode of procedure.

In the first place, it was soon apparent that the mere numbers of bacteria, per cubic centimetre of water, are in no sense a satisfactory guide to the fitness of such water for domestic purposes; it may be quite true that one revolts from a water proved to yield large numbers of colonies of bacteria, and one can understand that a water yielding even 500 colonies per cubic centimetre should be preferred to one yielding, say, 5000 colonies per cubic centimetre, but, so long as this choice is based on the assumption that mere numbers decide the safety or danger of the water, it is utterly fallacious. The 500 colonies may contain some which have been developed from pathogenic germs, while the 5000 may have all arisen from harmless forms. This consideration entirely invalidates all the older conclusions, which were made in some quarters, as to a given water being good or bad according as it yields few or many colonies per cubic centimetre on plate cultures; the only test is to determine what the bacteria of the different colonies are, and the only general deduction of any value to be drawn from mere quantitative bacteriological determinations is, that a water obviously containing a number of different species is, on the whole, more likely to have been subjected to contamination than one which contains but few different kinds.

A water should be suspected, therefore, and subjected to further examination, if it yields several different kinds of colonies unknown to the investigator.

As a matter of practical experience, it is certainly impossible to rapidly identify more than a few colonies in such cultivations; if a complete investigation of the life histories, &c., of all the forms were attempted, the bacteriological examination of a single sample of water might take years, and consequently this part of the subject is the one which awaits and invites the attention of numerous and energetic, properly equipped workers.

Then as to the primary assumption which lies at the base of all the older plate cultures. This was that each colony has taken origin

from one germ, isolated at the time of infecting the gelatine, and which developed in the medium during the period of culture.

In the first place, the conclusion that each colony has sprung from one germ is a mere assumption, and it is to be viewed with suspicion at the outset. Cramer* showed that bacteria have a habit of sticking together in the water, and several other observerst have shown that this is a real danger in all bacteriological analyses, and that individual colonies often result from the growth, &c., of not one, but many agglomerated spores or segments. Many observers have attempted to get over this difficulty by shaking the sowing in distilled water, before infection. Wolffhügel and Riedel suggest the possibity that there are dangers connected with this method also, e.g., removal of gases, oxidation, &c. It is even asserted that prolonged‡ mechanical shaking affects the growth of bacteria, but this concerns the transit, &c., of cultures rather than the point under discussion.

It has been suggested that the best method for ensuring separation from one another of the bacteria would be to pass the water through sterilised glass wool, as Elfving did for spores of fungi; § only there would be some loss. Thoroughly sterilised cotton wool, or even filter papers, may be used, but there is danger of washing traces of soluble

substances from these.

The question as to whether the colonies result from a single germ or from an agglomeration is, after all, not a matter of such grave importance as might at first sight appear, for if the precaution be taken, as it invariably should be, of violently agitating the sample of water immediately before making a plate cultivation, it is obvious that any conglomerate which may be present and does not yield to this treatment is, for practical purposes, a single source of infection, and will thus give rise to a single colony.

Another difficulty with plate cultures is due to some species causing liquefaction of the gelatine through the action of peptonising *Kommissions-bericht über die Wasserversorgung von Zürich und ihren Zusammenhang mit der Typhus-epidemie des Jahres 1884,' Zürich, 1885, p. 92.

6

+ Malapert-Neufville (Zeitschr. f. analyt. Chem.,' vol. 25, 1886, p. 39), Wollfhügel and Riedel ("Die Vermehrung der Bacterien im Wasser," Arb. a. d. K. Gesundheitsamte,' vol. 1, 1886, pp. 455-480); also Fol and Dunant (Revue d'Hygiène,' 1885, vol. 7, p. 183).

It may be assumed that the shaking carried on for a few minutes only before making a culture can have no effect on the life of the microbes; even the effect of long-continued shaking is very doubtful, the evidence being quite conflicting. On this point see Horvath (Pflüger's Archiv f. Physiol.,' vol. 17, 1878, p. 125), Naegeli (Theorie d. Gährung,' 1879, p. 88), Reinke (Pflüger's Arch. f. Physiol.,' vol. 23, 1880, p. 434), Büchner (Sitzungsber. d. K. Bayer. Akad. d. Wiss.,' 1880, pp. 382 and 406), Wernich (Desinfectionslehre,' 1880, p. 74), and further literature in these.

§ Elfving used cotton-wool (Studien ü. d. Einwirkung des Lichtes,' p. 31). See also Geppert (Ann. de l'Inst. Pasteur,' vol. 3, p. 673), who used glass.

ferments which they produce: this causes local floodings, and the running together of neighbouring colonies, or the submergence of some of them, and seriously interferes with the counting and estimation of the numbers.

The only mode of combating this difficulty consists in using such a volume of the infecting water as will yield a manageable number of such centres of liquefaction.

But these are by no means the only sources of fallacy incidental to the methods of gelatine-plate cultures. It has been implied by some of the earlier workers, rather than definitely assumed, that all the living germs in the sample of water mixed with the nutrient gelatine* will give rise to colonies, provided the plate culture is thin enough to ensure the access of oxygen to all parts, the sample small enough to ensure isolation of the individual bacteria or spores, and the temperature a suitably high one to promote rapid growth, without preventing the proper solidification of the medium.

As matter of fact, there are serious fallacies traceable to all these implications. Many bacteria are now known which are incapable of growing in presence of the oxygen of the air, while others will only withstand partial pressures of that gas; it may be safely concluded that the gelatine-plate cultures give no account whatever of these forms, although they may and often do occur in tap waters,† &c.

Moreover, even the thinnest layer of gelatine may so far hinder the access of oxygen to completely submerged aërobic forms as to retard their growth, and so they become dominated by the more rapid development of other colonies. This domination is not necessarily due to the mere flooding of the suppressed forms with liquefied gelatine : Garrét showed a short time ago that some bacteria, growing on gelatine side by side with other species, can inhibit the life-actions of the latter by the poisonous influence of their metabolic products, and Miquel§ claims to have proved similar actions in water, and even to have isolated the toxic principles, and rendered other water immune by their aid.

The quality of gelatine and peptone varies also. For hints in this connexion see Reinsch," Zur bakteriolog. Unters. des Trinkwassers" (Centr. f. Bakt.,' vol. 10, 1891, p. 415).

+ A good instance has recently been investigated by Perdrix ("Sur les fermentations produites par un Microbe anaérobie de l'Eau," 'Ann. Inst. Pasteur,' vol. 5, 1891, pp. 286-311).

"Ueber Antagonisten unter den Bakterien" (Correspondenzbl. f. Schweizer. Aerzte,' Jahrg. 17, 1887). Also Blagovestchensky ("Sur l'Antagonisme entre les Bacilles du Charbon et ceux du Pus Bleu," Ann. de l'Inst. Pasteur,' 1890, vol, 4, pp. 689-715).

§ "Dixième Mémoire sur les Poussières organisées de l'Air et des Eaux " (Annuaire de l'Observatoire de Montsouris,' 1888), and ' Manuel Pratique d'Analyse Bactériologique des Eaux,' 1891, pp. 153–155.

Again, as is well known, there are several forms which will not grow on gelatine at all,* and there are others which grow so slowly that they will not be counted in the estimations made by gelatineplate cultures, either because the colonies formed in the time are too small to be seen, or because they succumb to dominant forms—for it must never be forgotten that, among competing Schizomycetes, it is especially the early forms which gain the advantage, as elsewhere in

nature.

Finally, since the temperature has been shown to be such a determining factor in the growth and multiplication of bacteria, we may be sure that this item affects these plate cultures also, and it is well known that different numbers are obtained according to the temperature of incubation, and with reference to this point it is especially to be noted that the optimum temperatures for different bacteria may differ considerably.†

Taking all the facts into consideration, therefore, it is necessary to regard the gelatine-plate method as an imperfect one at best. But if we inquire whether there is a better one, we are bound to reply that there is not, at any rate for general purposes; but for special requirements it is possible to devise several modified methods for the culture of particular forms, and this has been done in certain cases.

§ III. The Vitality of Micro-Organisms in Water.

It is now time to enter upon the special literature dealing with the behaviour of selected forms of Schizomycetes in particular samples of water, and we propose to treat this somewhat more in detail, and in chronological order, so far as possible, because it bears directly on the subject of our enquiry.‡

This invesigation followed as the natural corollary to the discovery that some micro-organisms can multiply to a most extraordinary extent in waters almost entirely destitute of organic matter, like distilled water. The first recorded instance of such multiplication

* Some of these are of the highest importance in connexion with the chemical changes taking place in natural waters, e.g., the nitrifying organisms (Percy F. and Grace C. Frankland, "The Nitrifying Process and its Specific Ferment," Phil. Trans.,' 1890, B, p. 107; Winogradsky, 'Ann. de l'Inst. Pasteur,' 1890 and 1891; Warington, Chem. Soc. Journ.,' 1891, p. 484).

+ In this connexion it should be noted that the range of temperature for different bacteria is much larger than is commonly assumed. There are species which will grow at 0° C., and there are others which grow at temperatures as high as 50-70° C. See Fischer (" Bakterienwachsthum bei 0° C., &c.," Centr. f. Bakt.,' vol. 4, 1888, p. 89), Globig ('Zeitschr. f. Hyg.,' vol. 3, p. 294), Forster (Centr. f. Bakt.,' vol. 2, p. 337), Miquel ("Monogr. d'un Bacille vivant au-delà de 70° C.," 'Ann. de Micrographie,' Année I, Paris, 1888, pp. 4-10).

See Appendix C, p. 268.

in distilled water was made by one of us in 1885,* on which occasion it was found that in the course of forty-eight hours the number of microbes had increased from 1073 in the cubic centimetre to 48,100 in the same volume. Similar instances of multiplication in the pure spring water supplied to Munich were published shortly afterwards, by Leone, whilst the same phenomenon was observed by Cramer‡ in the case of the microbes present in the waters of the Lake of Zürich.

Similar phenomena formed the subject of more extensive investigations contained in three memoirs, which appeared almost simultaneously in 1886 by Wolffhügel and Riedel,§ by Meade Bolton,|| and by one of us.¶ Each of these not only confirmed the rapid and extensive multiplication of microbes, even in the purest natural waters, but differed from the predecessors in recording the results of experiments in which specific pathogenic forms were introduced into natural waters of different kinds, including sewage. It will be convenient to discuss, in the first instance, these three contributions to the subject along with another by Heraeus, which appeared shortly afterwards.

In Meade Bolton's paper,' ** after referring to the literature regarding the general bacterial contents of ordinary waters, and criticising the various methods hitherto in vogue, the following generalisations are made

1. Ordinary waters always contain some bacteria.

2. The numbers of individuals and species vary in different waters, and from time to time.

3. Certain forms predominate, because they can multiply readily in such waters, as is proved by their rapid increase when the water is allowed to stand for a few days.

4. After the climax of increased numbers has been obtained, the bacteria gradually diminish in quantity.

Bolton established the truth of these conclusions, and showed that the growth and increase of these water bacteria differ according to

* Percy F. Frankland, "The Removal of Micro-organisms from Water," Roy. Soc. Proc., vol. 38, 1885, p. 387.

+ "Sui Micro-organismi delle Acque Potabili, loro Vita nelle Acque Carboniche," 'Rendiconti della R. Accademia dei Lincei,' 4 Ottobre, 1885; 'Chem. News,' vol. 52, p. 275.

'Die Wasserversorgung von Zürich, ihr Zusammenhang mit der Typhusepidemie d. Jahres 1884,' Zurich, 1885.

§ Arbeiten a. d. K. Gesundheitsamte,' vol. 1, pp. 455–480.

Zeitschr. f. Hyg.,' vol. 1, p. 76.

Perey F. Frankland, "The Multiplication of Micro-organisms," 'Roy. Soc. Proc., vol. 40, 1886, p. 526.

**"Ueber das Verhalten verschiedener Bakterienarten im Trinkwasser" (Zeitschr. f. Hyg.,' vol. 1, 1886, pp. 76–114).