HFFECT OF TEMPERATURE ON BACTERIA. 29 
METHODS. 
The methods of experimenting used in the experiments de- 
scribed below are, briefly, as follows. Milk which has been 
used for the purpose has been ordinary market milk, obtained 
from one of the milk dealers in Middletown. The dealer in 
question keeps his dairy in a rather exceptionally good condi- 
tion, and the milk which he furnishes is above the average, It 
is quite fresh when delivered from the cart, being only one or 
two hours old, and contains, as will be seen from the experi- 
ments, a rather small number of bacteria, usually about 20,000 
per cubic centimeter. Our experiments thus far comprise two 
series. In the first the milk was divided at once into three lots, 
a series of bacteriological plates made immediately from the 
milk, and then the three samples placed at different tempera- 
tures. ‘They were again subsequently plated at regular inter- 
vals until they curdled. The temperatures chosen for the first 
series of experiments were 37° C., 20° C., and 10°C.; these tem- 
peratures representing body temperature, or hot summer heat, 
the temperature of an ordinary room, and that of an ice chest. 
In the second series of experiments the temperatures chosen 
were 20°C., 10°C., and 1°C. It would have been preferable 
to use all four temperatures with the same lot of milk, but this 
would have increased the number of plates to be studied so 
greatly as to make it impracticable to handle them all at the 
same time. The intervals chosen for making the bacteriologi- 
cal analyses were necessarily different for the three samples. 
In the milk kept at 37°C., the development of bacteria was 
very much more rapid than in either of the other samples. 
This made it necessary to analyze the warmer sample of milk 
at much more frequent intervals than the cooler samples. On 
the other hand, the bacteria in the milk kept at 10° C. developed 
very slowly, and a very much longer interval between the dif- 
' ferent analyses was necessary to make them at all comparable 
to the tests ef the other two samples of milk. It required con- 
siderable preliminary experimenting before it was possible to 
determine at what intervals the different tests should be made 
in the different samples. ‘The purpose was to obtain tests at 
such intervals that the development of the bacteria should be 
as closely as possible parallel in the three tests, so far as 
concerns numbers. ‘The great irregularity in the rapidity of 
