342 
was evaporated to dryness in the incubator and kept for some 
days at a temperature of 40°C. On February 20 the dried 
glass was charged with boiled and cooled solution, and 
plugged with cotton wool in the usual way. The liquid 
was examined microscopically on March 2, when it contained 
numerous torula cells, but no trace of microzymes. It there- 
fore appeared that the germinal particles of microzymes are 
rendered inactive by thorough drying without the application 
of heat. As, however, it could not be concluded therefrom 
that drying acted in a similar manner on the microzymes 
themselves, an experiment was made on this point also. 
XVIII.—March 4.—As it appeared probable that in the 
previous experiments with bacteria scum, desiccation might 
be prevented by the gelatinous matrix, a portion of the same 
scum was thoroughly washed with water, collected in an 
eprouvette, and dried for some days in the incubator. The 
eprouvette was then (March 4) charged with boiled and 
cooled Pasteur’s solution, and plugged with cotton wool. 
On March 11, the liquid was slightly hazy, but on micro- 
scopical examination was found to contain no trace of 
microzymes, The haziness was due to the presence of torula 
cells in great numbers. On the 18th the appearances were 
similar, but mycelium now existed in addition to torula. It 
thus appeared that fully formed bacteria are deprived of their 
power of further development by thorough desiccation; so 
that we may conclude that the contamination of water by 
apparently dry surfaces happens only in those cases in which 
desiccation is incomplete. It will be seen that this conclusion 
is quite ¢onsistent with the previous observations. 
Method of testing the zymotic property of water.—-As a test 
of the faculty possessed by all water which is not absolutely 
pure, of determining the growth of microzymes, Pasteur’s 
solution gives results so constant and satisfactory that it 
appears scarcely necessary to seek for better, although there 
is no doubt that many other liquids would answer the same 
purpose, and that some would react with greater delicacy. 
The method consists, as already indicated, in the addition 
of a small quantity of the suspected water to a relatively large 
volume of the solution. As it is very desirable that the con- 
ditions of experiment should be subject to as little variation 
as possible, in our test experiments we add one drop of water 
to a centimetre of solution, always using the same dropping 
pipette. As the eprouvettes commonly employed contain 
five centimetres when half full, this quantity is preferred, 
so that in the following paragraphs the term “ charged 
eprouvette ” is understood to mean an eprouvette which has 
