9 6 



ELEMENTS OF WATER BACTERIOLOGY. 



Number of samples tested ^ 5 T 7 2 



" giving preliminary fermentation. . . . 1036 

 Per cent of latter proved to contain coli 70 



474 



The work of Hunnewell and one of ourselves (Wins- 

 low and Hunnewell, i902 b ) showed that while 50 samples 

 of polluted waters all gave gas in the dextrose tube, only 

 40 out of 157 samples from presumably unpolluted sources 

 showed a similar fermentation. Whipple (Whipple, 1903) 

 examined a large number of surface-water supplies by 

 the tl presumptive test" and obtained striking results, 

 shown in the following table. The waters are arranged 

 in six groups according to the results of sanitary inspec- 

 tion, group I including waters collected from almost 

 uninhabitated watersheds and group VI waters too much 

 polluted to be safely used for domestic purposes. 



PERCENTAGE OF SAMPLES OF WATERS OF VARIOUS SANITARY GRADES 

 GIVING POSITIVE TESTS FOR B. COLI WHEN DIFFERENT AMOUNTS 

 WERE EXAMINED. 



(Whipple, 1903.) 



In view of these results Whipple suggests the following 

 provisional scheme of interpretation, which seems to the 

 authors to be, in general, sound. 



