142 EXCREMENTAL ORGANISMS 



reciprocal of the lowest quantity showing a positive result as 

 representing the number of B. coli per cubic centimetre. Thus, 

 0.1 c.cm.+, 0.01 c.cm+, 0.001 c.cm. — , was expressed as 100 B. 

 coli per cubic centimetre. When the average of a number of 

 samples from one source is calculated by this method (Phelps ®) 

 an accurate result is obtained providing the series is fairly 

 large (about 25), but McCrady ^ has shown that for individual 

 samples such assumptions are far from accurate. McCrady 

 calculates from the theory of probabilities that the most prob- 

 able number of B. coli present per cubic centimetre, if the above 

 result were obtained, would be 230 and not 100 as assumed. It 

 is possible that any number of B. coli per cubic centimetre would 

 produce this result and, in order to reduce the range of possibili- 

 ties and sharpen the probability curve, it becomes necessary to 

 employ more than one tube of each dilution. The greater the 

 number of tubes used the greater is the precision obtained. With 

 a milk of unknown origin that may contain up to 100,000 B. coli 

 per cubic centimetre it is obvious that even if only three tubes of 

 each dilution are used the total number of tubes for each sample 

 becomes so great as to be cumbersome. For this reason the tube 

 method of estimating B. coU in milk cannot be reconmiended. 

 The third point for consideration is the method of recording 

 the results. If desired, all tubes showing gas may be plated out 

 on rebipelagar or litmus lactose agar and the red colonies so 

 obtained put through confirmatory tests, but as such a pro- 

 cedure requires much time and labour it will be found more 

 convenient and fairly accurate to record all tubes as positive 

 that show more than 5 per cent of gas. Anomalies at the 

 higher end of the series should be ignored as they are probably 

 the result of overgrowths, but those at the lower end should 

 be corrected by moving the lower positive results to the next 

 higher dilution; thus, 1.0 c.cm. — , 0.1 c.cm.+, 0.01 c.cm.+, 

 0.001 c.cm!+, should be recorded as 1.0 c.cm.+, 0.1 c.cm.+, 

 0.01 c.cm.+, 0.001 c.cm.+, but 1.0 c.cm.+, 0.1 c.cm.-l-, 

 0.01 c.cm. — , 0.001 c.cm.+, should be recorded as 1.0 c.cm.+, 

 0.1 c.cm.+, 0.01 c.cm.+, 0.001 c.cm.-. 



