102 Report of the Department of Bacteriology of the 



A question which naturally follows is: How many samples giving 

 plate counts below 100,000 per cubic centimeter show microscopic 

 fields in which bacteria can be easily and quickly seen? Out of the 

 four hundred and fifty samples examined there were two hundred 

 and forty-six which gave plate counts below 100,000 per cubic centi- 

 meter and sixty-seven of these gave microscopic smears in which 

 bacteria could be readily found. In other words the plate method 

 passed sixty-seven of the two hundred and forty-six samples as 

 having less than 100,000 per cubic centimeter while the microscopic 

 examination showed that they had many more bacteria than this. 

 Thus the microscopic method is the more severe test and probably 

 the more accurate. 



It is more difficult to estimate the value of the microscopic method 

 when applied to milk with very low counts. Many of the samples 

 in Tables IV, V and VI are equivalent to the bacterial requirements 

 of certified milk, averaging about 5,000 per cubic centimeter by the 

 plate method, while the total number of individual bacteria by the 

 microscopic count averages about 221,000 per cubic centimeter. 

 This shows that the plate method does not reveal all the bacteria 

 and that the extremely low counts secured by means of the agar 

 plate count for certified milk are often misleading. The results 

 indicate the possibility of the application of the microscopic method 

 to counting the bacteria in certified milk, but a careful detailed count 

 would be required. Those who have not tried it may think this a 

 laborious operation but it is not a difficult nor a tedious task to 

 examine as many as one hundred fields of the microscope in which 

 only a very few bacteria can be found. 



Inequality in size of clumps and varying degrees of tenacity with 

 which different species resist separation by shaking were undoubtedly 

 some of the unmeasurable factors which were largely responsible 

 for the irregularities in results secured by the two methods of counting. 

 The average-sized clump found contained from four to sixteen or 

 twenty bacteria and clumps were frequently seen which contained 

 as high as seventy or eighty individuals. Ordinarily little difficulty 

 was experienced in counting all the bacteria in clumps no larger 

 than these. Clumps were occasionally found which contained more 

 than one hundred individuals and these frequently could not be 

 accurately counted, especially if they were very compact. In one 

 instance a dense clump covering more than one-half of the whole 



