30 STORRS AGRICULTURAL EXPERIMENT STATION. 
multiplication of bacteria, even at the same temperature, as 
shown in the following tables, made it impossible to choose 
intervals that corresponded exactly in different experiments. 
After some preliminary tests it was found that the best inter- 
vals were as follows: Milk kept at 37°C. was plated every 2 
hours; that kept at 20° C. every 6 hours; that kept at 10° C. 
every 12 hours; and that kept at 1°C. about every 3 days. 
In making plates for analyses it was always necessary to di- 
lute the milk with sterilized water. Here arose the greatest 
problem in the series of experiments, and one which was never 
solved with perfect satisfaction, since it was not possible to be 
sure that the dilution chosen would be the proper one for the 
purpose of analysis. Experience has shown us that, in order 
to make a differential analysis of species, it is desirable to 
have from 200 to 500 colonies on each plate. If the number 
is higher than that, the differentiation is unsatisfactory, since 
the colonies look so much alike when crowded; if the number 
is lower than 200, the analysis cannot be regarded as a fair 
sample, and species present in small numbers will be missed 
entirely. The proper dilution to obtain this number of colonies 
on the plate varied, of course, with the number of bacteria in 
the milk. If the bacteria multiplied with the same rapidity in 
all samples kept at the same temperature, it would, of course, 
be possible, after some experience, to determine quite closely 
the desired dilution. But since the different samples of milk 
contained at the outset different numbers, and since the rate of 
multiplication at the same temperature varied in the different 
samples, and inasmuch as it was never possible to determine 
with accuracy how many bacteria were to be expected, in many 
cases the proper dilution was not obtained, and many experi- 
ments were rendered useless by this fact. We first carried 
through several experiments simply for the purpose of de- 
termining the numbers of bacteria, and then, from the gen- 
eral average of the numbers of bacteria in these preliminary 
experiments, made as close an approximation as possible to 
the proper dilution, at different hours, for producing from 200 to 
500 colonies to the plate. Where the bacteria in the later samples 
developed either more or less .rapidly than in the preliminary 
tests, the dilutions did not come out satisfactorily. Asa result, 
in the experiments described, a somewhat unequal value must 
