New York Agricultural Experiment Station. 



97 



on the samples from Farms A, B, C and D and this fact may possibly 

 be responsible for the change in the ratios. No milk found among, 

 these thirty-three dairies had a plate count of less than 10,000. 

 The important point to be noted in Table X, as in Table III, is the 



Table X. — Combined Summary of Bacterial Counts of All Samples of Milk 

 Studied by Agar Plate and Microscopic Methods. 

 Numbers computed for a cubic centimeter. 



* Each clump and each isolated bacterium counted as one; individual bacteria in 

 clumps not counted. 



fact that a wide difference exists between the plate and microscopic 

 counts when the bacteria are few in number and the fact that the 

 relative difference between the counts rapidly decreases as the 

 number of bacteria increases. 



DISCUSSION OF RESULTS. 



A study of the individual tables of bacterial counts reveals 

 the existence of a certain relationship between the counts obtained 

 by the plate method and the total number of individual bacteria per 

 cubic centimeter according to the microscopic method, when series 

 of counts are averaged together. Reference to Table IV shows a 

 series of low plate counts with a correspondingly low total indi- 

 vidual bacterial content per cubic centimeter. Successively higher 

 counts showing similar parallels are seen on Tables V, I and II 

 respectively. The existing parallels are, however, more striking in 

 a long series of counts than in a short one, owing to marked irregu- 

 larities in individual cases. 



A more instructive comparison is shown in Tables VIII and X 

 where the bacterial counts have been arranged and summarized 



