no Elements of Water Bacteriology. 



source of error as well as the relative importance of the 

 various other diagnostic tests is well shown in the table 

 (page 109) of the results obtained at Lawrence during a 

 period of eighteen months. 



It should be noted that considerable differences often 

 appear in these biochemical reactions between tubes of- 

 the same batch of a medium, inoculated with approxi- 

 mately the same amount of the same culture. This has 

 been shown very markedly for the amount of the gas and 

 the proportion of carbon dioxide in the dextrose tube by 

 Fuller and Johnson (Fuller and Johnson, 1899), Penning- 

 ton and Kiisel (Pennington and Kiisel, 1900), Gage 

 (Gage, 1902), and one of ourselves (Winslow, 1903). 

 Variations in nitrate reduction are often even more 

 marked, one tube, perhaps, showing a strong reaction 

 and another none. In important cases, therefore, it is 

 desirable to inoculate the subcultures in duplicate. 



These anomalies are most frequent with cultures freshly 

 isolated from water, and they may often be avoided, as 

 Fuller and Johnson (1899) have shown, by subjecting the 

 organism to a process of preliminary cultivation. For 

 this purpose the American Public Health Association 

 Committee recommends three successive cultivations in 

 broth at 20 degrees, each of 24 hours duration, inoculation 

 from the last broth tube of a gelatine plate which is incu- 

 bated for 48 hours at 20 degrees, inoculation of an agar 

 streak from one colony on the plate and incubation of 

 this streak for 48 hours at 20 degrees. 



