98 Journal of Agricultural Research voi. xxii. no. 2 



to show that the scattered organisms in cheese sections were due, in a 

 large measure, to the breaking up and scattering of the clumps by the knife 

 edge. This explanation will hardly appear plausible to anyone familiar j 

 with the perfection of delicate histological sections prepared with a sharp 

 knife. 



The grouping and clumping of the organisms often cause difficulty in 

 accurately determining the number of organisms in the cheese sections. 

 This is especially true in sections of old cheeses in which the bacteria 

 tend to clump in large masses. The error can be overcome to a large 

 degree by counting or estimating a large number of fields, the larger the 

 number examined the smaller being the error in the final estimate. 



In a sample of green cheese where the organisms appear in large num- 

 bers, but are evenly scattered, it is impracticable to count the entire 

 field, and an ocular disk divided into quadrants may be inserted in order to 

 facilitate accurate counting. 



In all cases 20 or more fields should be counted, and especially where 

 the organisms are unevenly distributed. In such instances, typical 

 fields which represent the general flora should be located by studying the 

 entire section. 



COMPARISON OF DIRECT AND PLATE COUNTS 



Table I gives a few representative comparisons between direct micro- 

 scopic and plate counts made from cheese samples in various stages 

 of ripening. The plate counts average approximately one-twelfth the 

 direct count, but no common ratio has been found to exist between the 

 results obtained by the two methods. Wide variations in the ratios be- 

 tween the counts were found, but in general the ratios from green cheese 

 appeared to be larger than those from cheese more advanced in ripening. 



The above plate counts compare well with those found by other observ- 

 ers who have examined cheddar cheese, Russell (10) found from 62 to 

 665 million per gram. Harrison and McConnell (7) found the count to be 

 as high as 625 million per gram in the earlier stages of the ripening, while 

 Harding and Prucha (5) observed from 37 to 177 million per gram. 



Several explanations may be offered to account for the apparent 

 discrepancy between the results obtained by the two methods. The 

 plate count is an estimate based on observations of the growth of organ- 

 isms on some particular medium which, in cheese investigations, usually 

 contains lactose. Lactose has been generally used because media con- 

 taining this particular carbohydrate have been found to allow the develop- 

 ment of a larger number of colonies than do sugar-free media. Investigators 

 have based their cultural methods upon media giving the largest counts 

 rather than upon media which might serve as an index to the relative 

 number of types present. In comparing the microscopic counts with 

 results obtained with the plate method, it may be noted that the types 



