898 Popular Editions of Station Bulletins of the 



of the fresh, morning milk were taken each day from the milk of 

 one farm, and of the night milk, delivered at the same time, for 

 60 days. Not quite so many samples were taken from the other 

 farms; but the series made possible a detailed study of the daily 

 variation and development of the bacterial condition of the milk 

 in 225 samples. 



The milk from 33 other farms contributing to the city supply was 

 studied in less detail, not more than a week being given to any farm, 

 and only five fields in any sample were counted under the micro- 

 scope. The object in these more superficial studies was to make 

 a general survey of the situation and to determine the efficiency 

 of the microscopical method when used rapidly as it would be under 

 commercial conditions. 



The results of the counts showed little relation- 

 Relative ship between the figures secured by the two methods 

 accuracy when only single samples were considered; but 



of the two when series of samples are examined a relationship 

 methods. is shown. The count made under the microscope 

 is almost invariably much higher than that shown 

 on the plates and certainly represents the total number of indi- 

 vidual bacteria more accurately than the plate count, the results 

 in the latter case being invariably low because of clumps. Among 

 the 450 samples examined only three showed more bacteria by the 

 plate method than by the direct microscope count. 



The relative differences between the two counts are greater when 

 the bacteria are few in number. In samples of milk showing plate 

 counts of 10,000 per cubic centimeter, the microscope showed 

 approximately 44 times as many individual bacteria. A somewhat 

 fairer basis for comparison, however, is to consider each " clump " 

 of bacteria shown under the microscope as a unit only, since such 

 a collection of germs would, on the plate of nutrient medium, develop 

 only one colony, perhaps indistinguishable in any way from the colony 

 surrounding a single isolated bacterium. 



On this basis, the germ-poor milk referred to above showed only 

 17 times as many organisms in the sample under the microscope as 

 on the petri plate. When the milk contained more bacteria — 

 about 1,000,000 per cubic centimeter — ■ the count under the micro- 

 scope was only about 5 times as great as on the plates; or, if the 

 clumps were considered as units only, the microscope count was 

 slightly lower than the plate count. 



We cannot say definitely as yet, why there are such great differ- 

 ences in the counts by the two methods; but the fact that there are 

 differences is not at all surprising under the conditions that have 

 already been explained; and does not discredit either the old or the 

 new method. It is hoped by further work to secure a logical and 

 satisfactory explanation. 



