128 THE ENUMERATION OF BACTERIA IN MILK 



usual plate count it was essential that a constant proportion of 

 the bacteria capable of development on agar in forty-eight hours 

 at 37 C. must be precipitated during the process of centri- 

 fugalisation. A portion of the bacterial flora of milk, however, 

 does not produce visible colonies on agar under the usual condi- 

 tions, so that either these organisms must remain in suspension 

 or the error due to them be counterbalanced by some other 

 factor. 



No difficulty was found with the technique until the micro- 

 scopical examination was made. The representative field in 

 which the organisms were to be counted was difficult to find 

 owing to the widely differing content of various fields. In 

 order to minimise this source of error ten fields were taken at 

 random and the average calculated. 



In a series of market samples, for which the standard was 

 500,000 bacteria per c.cm. not a single sample was condemned 

 which passed the plate method; on the other hand, 17 per cent 

 were passed which were condemned by the plate method. 

 According to these results the direct method outlined above 

 would not be oppressive on the milk producer, and its adoption 

 would be tantamount to lowering the standard. In this series 

 the factor (c) for the conversion of microscopic counts to plate 

 counts varied within very wide limits, viz., from 0.4 X10 4 to 

 33. OX 10 4 , and the author is convinced that this is largely due to 

 the difficulty found in obtaining an even distribution of organ- 

 isms on the slide. Two observers obtained widely varying 

 results from the same slide; a condition fatal to accuracy. 

 Breed, 11 in 1911, improved this method by making a direct 

 smear of the milk and thus eliminating the centrifuge with its 

 many unknown factors. Breed's method consists essentially 

 in spreading a small volume of milk over a marked area and 

 examining under a high-power objective after washing out the 

 fat followed by suitable staining. Skar, 12 in 1912, independ- 

 ently developed a similar method which differs only in the 

 manner of staining and in allowing the fat to remain in the 

 smears. Rosam's method 13 differs essentially from Skar's 



