2 BULLETIN 563, U. S. DEPARTMENT OF AGRICULTURE. 
living organisms, the bacteriological method of analysis must take 
into account their distribution and development into colonies on the 
Petri plates. In this part of the method we encounter the difficulty 
of separating the bacterial cells and distributing them evenly. Their 
development is more or less influenced by the growth of different 
kinds of bacteria, one of which may retard the development of those 
near it. 
All these points are recognized by bacteriologists and are men- 
tioned here merely to call attention to the many difficulties which 
arise in making accurate bacterial counts and to point 
out that there must be variations in the result over 
which the analyst has no definite control. This con- 
dition has been recognized, consequently duplicate 
plates are made and results reported from the average 
counts of both plates. The variation in bacterial 
counts is particularly important and must be taken 
into consideration when a study is made of various 
samples of any material containing bacteria. If these 
variations are not considered, mistakes are easily made 
in the study of the distribution of bacteria. 
If the bacteria in ice cream are unevenly dis- 
tributed, and a bacterial analysis of a sample does not 
location of nine give results which will represent the whole mass of the 
s^npies^ taken cream? this fact will greatly complicate any study of 
can of ice the bacteria in the product. Consequently, before 
starting any further studies on the subject it was con- 
sidered advisable to carry out some experiments to throw more 
light on this point. 
METHOD OF SAMPLING AND PLATING THE ICE CREAM. 
Ice cream from various manufacturers was delivered in 1 -gallon 
cans at the laboratory. As soon as received, the can of cream was 
removed from the tub, the ice and salt wiped off, the cover removed, 
and the top layer taken off with a large sterile spoon. 
Three samples were taken from the topmost third of the gallon, three 
from the middle, and three from the bottom third, making a total of 
nine samples from each gallon, as shown in figure 1. Each sample 
contained about 30 grams and was removed with a small sterile scoop 
and placed in a sterile flask. After removing samples from positions 
1, 2, and 3, about one-third of the ice cream was removed with a 
sterile spoon and three more samples taken from positions 4, 5, and 6; 
similarly the three remaining samples were taken from positions 
7, 8, and 9. 
The flasks containing the samples were then placed in water at 
40° C. (104° F.) for 15 minutes in order to melt the ice cream, the 
