416 . MILK 



is a serious objection, as it delays obtaining results and ties up a 

 great deal of laboratory material. In the interest of accuracy, 

 however, too much weight should not be attached to these objec- 

 tions. A study of the influence of temperature and period of 

 incubation and, incidentally, of the influence of dextrose on the 

 number of colonies developing, made by Glenn and the writer, 

 led to the following conclusions: 



"1. Since pathogenic bacteria are difficult, in most cases im- 

 possible, to find in milk, a higher temperature of incubation has 

 no advantage over room temperature from this viewpoint. 



"2. Incubation at 20 C. is superior to incubation at 37 C. 

 because a higher count is obtained and a better differential count. 



"3. Dextrose is preferable to lactose as an addition to the me- 

 dium. 



"4. Milk is usually consumed before results of bacterial ex- 

 aminations are available. Bacteriologic and chemical examina- 

 tions should, therefore, have the principal object of improving and 

 controlling the general supply, and accuracy is of greater impor- 

 tance than quick results. The loss of a day in the interest of accu- 

 racy is irrelevant." 



The chart (Fig. 184) illustrates the results obtained. 



Conn found that the five-day count in most cases was slightly 

 higher than the two-day count, although in some cases it was 

 somewhat lower. Slack thinks that the addition of a carbohydrate 

 to the medium is irrelevant, and he has obtained as good results 

 with media containing no carbohydrate. 



Much also depends upon the number of colonies developing 

 on a plate. A large number gives too low a count, because some 

 colonies remain small and may be overlooked, while a small num- 

 ber of colonies on a plate may give too high a count, unless un- 

 diluted milk is plated. This is a difficult proceeding because the 

 opalescence of the milk renders colonies hard to distinguish. 

 Plates with 40 to 200 colonies are considered the best for a fair 

 estimation. At least two, better three, plates should be made 

 from each dilution, and the average taken as the proper number. 



A very important source of difference in results is the assump- 

 tion that each colony originates from one cell. This is not true, 

 and it is not known what percentage of error is caused by this 

 irregularity. Chains of streptococci and clumps of other bac- 

 teria frequently adhere with considerable tenacity, and a colony 

 may have originated from a number of cells rather than from one. 

 This objection to the colony count cannot be wholly overcome with 

 present-day technic. 



The apparatus for counting colonies should be of uniform type, 

 since fight and other factors may render small colonies discernible 



