20 THE UNIVERSITY SCIENCE BULLETIN. 



the bacterial count ranging between 400 and 1200 per cc. These 

 waters are open to strong suspicion and should not be drunk. 



In group III there are fourteen samples, or 17.8 per cent, hav- 

 ing from 1 to 4 B. coli in each 10 cc. of the water, with compara- 

 tively high counts. 



Of the forty-nine samples in which no B. coli was found, twenty- 

 two of them, or 44.9 per cent, show counts developing at body 

 temperature of over 350 bacteria per cc. and eighteen of these 

 have 500 or more bacteria per cc. Thirty-one samples show 

 counts under 500 per cc, and of these twenty-seven, or about one- 

 third of the total number of samples examined, are under 350 per 

 cc. That is, if we attribute any significance — and we must — to 

 the number of bacteria developing at body temperature, and 

 adopt even 350 per cc. as a dividing line between clean and dirty 

 water, we find that of the seventy-eight samples examined, only 

 twenty-seven, or about one-third, showed themselves to be suf- 

 ficiently clean for drinking purposes. 



The possible sources of pollution in these samples of water may 

 be grouped as follows: 



1. From the original water supply. 



2. From the water used in the ice. 



3. From the dirt adhering to the ice. 



4. From the hands or person of the individual handling the ice. 



5. From the buckets, wheelbarrows, etc., used in carrying and 

 handling the ice. 



6. From carelessness in washing the tanks as to thoroughness 

 and as to kind of water used in washing. 



