152 BACTERIOLOGICAL EXAMINATION OF MILK. 



within the limit, and thus saves the trouble of plating them. 

 The committee of the A. P. H. A. in its 1908 report (4) 

 makes the following comment : 



" Several laboratories are now making this examination as a routine 

 procedure. One \vorker (Conn), who has done considerable experimen- 

 tation along this line during the past year sums up his results in attempts 

 to actually approximate plate counts as follows : 



1 The method seems to be fairly satisfactory for milk, the bacterial 

 content of which is not too low or too high. When the numbers are down 

 below ten thousand the method seems to be quite inaccurate, and when 

 the numbers run up into the millions I find also the numbers are not 

 very reliable. Within the limits of from thirty thousand to three hun- 

 dred thousand, however, the method seems to be fairly good. I have, 

 however, had the experience of occasionally finding samples of milk ' 

 which, when tested by direct microscopic methods, gave results very 

 different from those by the plate method. My general feeling is that a 

 direct microscopic method might give an idea as to whether the sample 

 is very good or very bad, but would not replace the plate method of 

 examination in determining actual numbers and would be of no use for 

 the examination of samples of milk whose number of bacteria is quite low.' 



"Another (Campbell), who has done much work with this method, 

 finds it chiefly useful as a preliminary test, not. plating samples which 

 are shown by the microscope to contain less bacteria than the city regu- 

 lation permits, in those samples which are plated it serves as a guide for 

 the proper dilutions. He finds it possible with this examination to cor- 

 rectly state in nearly every instance \vhen a milk contains less than 50,- 

 000 bacteria to a cubic centimeter, or when it contains over a million bac- 

 teria to the cc. On actual estimates within 50,000 of the plate count his 

 average is 66% correct on counts between 50,000 and 500,000 and 50% 

 correct on counts between 500,000 and 1,000,000." 



The subject of the microscopic examination of milk for leu- 

 cocytes and streptococci is discussed in Chap. VII. 



Tests for B. coli Methods for the examination of milk for 

 B. coli are in the process of evolution. The question of the 

 accuracy and significance of some of these tests has received 

 some attention by the committee of the A. P. H. A. in its 1908 

 report. The methods have hardly been tested out sufficiently 

 to warrant recommendation for general use. 



Examinations for typhoid fever and diphtheria bacilli. In 

 connection with the investigation of milk-borne epidemics of 

 typhoid fever and diphtheria, the desire is very frequently 



