THE BACTEBIOLOGIC EXAMINATION OF MILK 419 



as in any similar work. As a matter of fact, different workers 

 may record varying numbers of bacteria in the same sample of 

 milk and even counting the same plates. This is due to several 

 factors. One of the most important ones is the uneven distribu- 

 tion of bacteria in milk. Tests made from one part of a sample 

 may show higher or lower counts than those made from another 

 part. It is impossible to overcome this difficulty, but it may be 

 stated that, as a rule, the difference in results is not large enough 

 to cause misinterpretation of the quality of the milk, unless the 

 number of bacteria is very close to the standards. In such bor- 

 derline cases one worker may condemn a milk because his enu- 

 meration is slightly higher than the permissible limit for a certain 

 grade, while another worker may consider the sample passable. 

 These cases, of course, occur rarely. 



The difference in counts is well illustrated by the following 

 figures of Slack: 



RELATION OF COUNT OF INDIVIDUAL WORKERS TO AVERAGE COUNT 



Relation to avera e count. A. B. C. D. E. Totals. 



Within 10 per cent , ... 7 4 5 4 2 22 



Between 10 and 20 pe. cent 4 5 3 11 3 26 



Between 20 and 30 per cent 7 5 1 6 3 22 



Between 30 and 40 per cent 2 2 1 2 4 11 



Between 40 and 50 per cent 1 2 4 2 4 13 



Between 50 and 100 per cent 3 1 3 5 4 16 



Over 100 per cent 1 2 3 



94 



19 

 Total... 25 21 17 30 20 113 113 



The American Public Health Association has published a 

 booklet entitled "Standard Methods for the Bacterial Examina- 

 tion of Milk." In this publication certain methods and media 

 are described which are to serve as guides for milk examinations 

 throughout the country. The results obtained in different labora- 

 tories are comparable only when uniform methods are followed. 

 Although the methods of the Committee of the American Public 

 Health Association are by no means perfect and are not approved 

 by some authorities, for the sake of uniformity they should be 

 universally adopted. In spite of care there are fundamental con- 

 ditions which render it difficult to obtain consistent results. It 

 should be remembered that the plate method as usually practised 

 does not reveal the presence of either anaerobes, soil bacteria, or 

 pathogenic bacteria. The relative representation of these groups 

 may be different in different milks, so that the proportion actually 

 counted bears no definite relation to the number actually present. 

 Furthermore, the variable possibility of breaking up clumps will 

 always be a factor beyond accurate control. 



Recently a comparative study of methods practised in differ- 

 ent laboratories in New York City and the results obtained wsre 



