New York Agricultural Experiment Station. 101 



be recognized and explained. The first instance occurred among the 

 samples reported upon in Table I, April 29, where a plate count of 

 278,000 and a total microscopic count of 268,000 per cubic centi- 

 meter was obtained. The second case of this kind is found in 

 Table IV in a sample of night milk taken March 7, where a plate 

 count of 260,000 and a microscopic count of 100,000 per cubic centi- 

 meter was obtained. The third occurred among the samples which 

 were taken during the summer and which are summarized in Table X 

 but not given in detail. In this case the plate count was 2,150,000 

 per cubic centimeter while the total number of individual bacteria 

 by the microscope was 960,000 per cubic centimeter. In each of 

 these three cases the duplicate plate counts agreed well and there 

 was no evidence of contamination in any of them. 



Outside of error in technique there remains another possible 

 explanation. Organisms may have been present which were so 

 small that they were overlooked in the stained casein. If such a 

 condition is the true explanation it is so uncommon as to be 

 negligible, since there were only three samples in which it occurred. 



In the work done on the thirty-three dairies which was summarized 

 in Table X, sixty samples out of the two hundred and twenty-five 

 were passed as having too few bacteria to count under the micro- 

 scope. The corresponding plate counts of these sixty samples showed 

 forty-two to have less than 50,000 bacteria per cubic centimeter 

 while eight were between 50,000 and 100,000, eight between 100,000 

 and 200,000 and two higher than this. One of the latter was prob- 

 ably contaminated in the plating, as the duplicates were very irreg- 

 ular and gave an average count of 883,000 per cubic centimeter. Of 

 the one hundred and twenty samples given in Tables IV, V and VI, 

 one hundred and one showed so few bacteria that none would have 

 been found if only a few fields of the microscope had been examined. 

 Of the one hundred and one samples, sixty-five gave a plate count 

 under 10,000 per cubic centimeter, thirty-four were between 10,000 

 and 100,000 and two above this. The average of these one hundred 

 and sixty-one plate counts where no bacteria could be found in a 

 few fields of the microscope was 29,000 per cubic centimeter. In 

 other words it seems safe to assume that practically all samples 

 passed by the microscope as having too few bacteria to count when 

 five fields are counted would yield a plate count of less than 100,000 

 per cubic centimeter. 



