micro-organisms, and that it has never before been inhabited by such living matters; it is only reasonable to infer, therefore, that those of its ingredients which are capable of nourishing the particular micro-organisms which flourish in it are wholly untouched, whilst in the case of the river-waters the most available food supply must have been largely explored by the numerous generations of micro-organisms which have inhabited. them. Also far fewer varieties of micro-organisms are found in this deep-well water than in the case of the river-waters, hence those forms which are present will have a more undisputed field for multiplication in the absence of competing forms. This would also explain the greater capacity for multiplication which is exhibited by the filtered river-waters as compared with the water in its raw condition, a large number of varieties having been eliminated in the treatment which the water has undergone at the waterworks during storage and sand-filtration.

This remarkable phenomenon of bacterial multiplication, generally taking place more abundantly in pure or in waters containing a small number of microbes to start with than in impure or waters containing a large initial number, has been made the subject of some highly interesting and suggestive investigations by Miquel.1

Thus a sample of the Vanne spring-water, in which only 150 micro-organisms were present in 1 c.c. at the time of collection, on being kept for twenty-four hours at 20° C., contained as many as from 30,000 to 40,000, whilst a sample of the Ourcq canal-water, which is highly polluted and very rich in bacterial life to begin with, on standing for a similar length of time, exhibited no increase in the numbers present.

1 Manuel pratique d'Analyse bactériologique des Eaux, Paris, 1891,

p. 146.

The results obtained with different waters have been graphically brought together by Miquel in the following diagrams, in which the abscissae represent the duration of time, and the ordinates the number of

FIG. 16. MULTIPLICATION OF BACTERIA IN SPRING WATERS.

(Miquel.)

bacteria present in 1 c.c. All the samples were main

tained at from 29° to 30° C.

The above figure represents the behaviour of certain spring-waters containing initially only a small number

Q

of microbes. Miquel points out how rapid and enor mous the multiplication is which takes place in the first instance, and how, when once the maximum has been reached, the decline as rapidly follows, becoming, however, less marked as the age of the sample increases. In fact, Miquel goes so far as to say that a spring-water may be characterised by the power of rapid multiplication possessed by the bacteria present, as well as by the rapid decline in their numbers subsequently exhibited.

The following diagram, on the other hand, represents the phenomenon of multiplication exhibited by the micro-organisms in waters containing a large initial number of bacteria.

In this figure the scale employed is much larger, and the Dhuis spring-water is introduced in order to illustrate, by comparison, how greatly inferior is the power of multiplication possessed by the bacteria in the rivers Marne and Seine and in the Ourcq canal-water. As regards the Seine water collected at Ivry (above Paris), Miquel states that at times during the summer months it contains relatively few organisms compared with its bacterial contents at other periods, and that the multiplication exhibited in such cases resembles that observed in spring waters. When, however, the river-water contains as many as 20,000 to 30,000 microbes in 1 c.c., this power of rapid multiplication disappears (see foot-note, p. 220). In the diagram the Seine water is relatively pure, and the bacteria present behave, to a certain extent, like those in spring-water, only the multiplication is neither so rapid nor so extensive. The river Marne exhibits first a rise in the numbers present, then a decrease, followed by a subsequent increase, a phenomenon which is not unfrequently met with in the multiplication of bacteria in water. The

most interesting feature of all in this figure is the behaviour of the bacteria in Ourcq canal-water. This water, containing to start with about 8,000 organisms in 1 c.c., requires as long as twenty days before the numbers.

Days.

FIG. 17. MULTIPLICATION OF BACTERIA IN IMPURE WATERS.

(Miquel.)

present reach 60,000, after which point a slight decrease is noticed; but the number of bacteria subsequently remains almost stationary at this high level, and examinations made of this sample at the end of six months, a year, and even after, revealed practically no alteration. Miquel goes on to say that the bacteria in the Seine

water exhibited at the end of a much shorter time a marked decline in their numbers, but after some months the decrease becomes more and more slow, until at the end of from ten to twelve years the number of bacteria is equal to, or is only a half or a third of the number present at the time when the sample was taken. Miquel summarises his results by observing that a rapid but transitory power of multiplication characterises the bacteria in pure spring-waters, whilst in impure waters, or in waters rich in microbes, the multiplication is slow and persistent.

Miquel has pursued this subject of bacterial multiplication further, and has found that after a water has supported the multiplication of a particular species of micro-organism, the latter, on being reintroduced into the same water, will not only not again multiply, but in many cases will actually suffer rapid destruction. He compares the phenomenon to that of a zymotic disease, and the immunity which generally follows as a consequence. Thus the water which has been afflicted' with a plague' of a particular microbe acquires immunity towards further attacks of the same organism. This immunity he ascribes to the generation by the bacteria of soluble and toxic products which inhibit their further growth and multiplication. It is the absence of such toxic products in pure spring-waters that permits of the astonishingly rapid and extensive multiplication of the few bacteria which they initially exhibit, and causes these waters to present such a marked contrast in this respect to more contaminated surface waters. These toxic products are destroyed on boiling, for a water which will not support any further bacterial multiplication acquires this property after boiling (for experimental confirmation of this statement, see p. 229). According to Miquel, however, these products, at any