REPORTS 



13 



present. This error would not be 

 troublesome if the groups were of 

 constant average size; but the best 

 information available shows that the 

 groups in ordinary market milk com- 

 monly vary in size so that they contain 

 an average of from 2 to 6 individual 

 bacteria. Some samples contain groups 

 of even smaller size than this, while 

 others, such as those bearing long 

 chain streptococci, may show groups 

 containing an average of 25 or even 

 more individual bacteria. The irreg- 

 ularity of this error (whose size is not 

 indicated in any way by the appearance 

 of the plates) should be kept in mind 

 in interpreting the results obtained. 



Reports. 



Because of the fact that agar plate 

 counts only represent a fraction of 

 the total number of bacteria present, 

 they should not be reported as showing 

 the "number of bacteria per cc." 

 Accurately speaking the counts from 

 agar plates give the estimated number 

 of colonies that would have developed 

 on standard agar per cc. of milk if an 

 entire cubic centimeter of milk had 

 been used for inoculation. Because 

 this statement of fact is cumbersome, 

 and also because a certain ratio exists 

 in each case between the colony count 

 and the total number of bacteria, it has 

 become a common practice to speak 

 of the plate counts as showing the 

 number of bacteria per cc. This is 

 very confusing now that microscopic 

 methods of counting have been de- 

 veloped which permits counts of the 

 actual bacteria to be made. These 

 counts average approximately five 

 times the size of the counts as made 

 by the standard agar plate technique. 



It is therefore recommended that all 

 agar plate counts obtained by the 

 standard technique shall not be stated 



ip the form "2,000,000 bacteria per 

 cc." but rather as follows: "ofiicial 

 plate count, 2,000,000." This latter 

 form of expression shall be considered 

 an abbreviated method of saying: 

 "a count of 2,000,000 colonies per 

 cc. as obtained by standard methods." 

 Moreover analysts shall be careful to 

 avoid giving a ficticious idea of the 

 accuracy of the ofiScial plate count. 

 There is ample justification for think- 

 ing them sufficiently accurate to justify 

 drawing conclusions as to the general 

 quality of a given sample of milk, and 

 when a series of samples from the 

 same source are examined the average 

 result may permit much more specific 

 conclusions to be drawn with con- 

 fidence. 



Specific data showing the actual 

 percentage error in these counts has 

 been difficult to obtain, and has only 

 recently been obtained by means of 

 comparisons made between microscopic 

 and agar plate counts. The conclu- 

 sions reached by Breed and Stocking, i' 

 are that the margin between two plate 

 counts made from similar samples 

 of market milk must be as great 

 as one to five before it becomes a 

 practical certainty that the larger 

 count actually represents the larger 

 number of bacteria. 



It is, however, self evident that 

 between any two samples the one 

 having the higher count probably 

 contains the greater number of bac- 

 teria, and this probability can be 

 made a practical certainty by the 

 examination of a series of samples. 

 It is therefore required that a series 

 of samples, preferably four or more, 

 be examined before judgment shall be 

 rendered as to the general quality 

 of a given milk supply. Under no 

 conditions is the practice sanctioned 

 of publishing exact counts from indi- 



