The Significance of B. Coli in Water. 137 



number fell to a normal value of 8 per cent. Correspond- 

 ing to this excess of B. coli in the city water, there were 12 

 cases of typhoid fever in December, 59 cases in January, 

 12 in February, and 9 in March, all during the early part 

 of the month. The authors conclude that " when filtering 

 a river-water as polluted as that of the Merrimac, it is 

 safe to assume that when B. coli is found only infre- 

 quently in i c.c. of the effluent, the typhoid germs, neces- 

 sarily fewer in number and more easily removed by the 

 filter, have been eliminated from the water." 



The results of the daily tests carried out at municipal 

 filter plants are frequently expressed in monthly or yearly 

 averages, as in some of the cases quoted above. It must 

 be remembered, however, that averages of this sort are 

 accepted only by courtesy and with the implied assump- 

 tion that conditions are approximately constant during 

 the period averaged. When it is said that an acceptable 

 effluent may show B. coli in three or four per cent of the 

 samples tested, the statement is true only for a series of 

 samples collected and examined at the same time. If in 

 a given month 3 per cent of the i-c.c. samples tested show 

 B. coli, the effluent may or may not be safe. If on each 

 of 20 days 3 B. coli or thereabouts were present in 100 c.c. 

 of the water, it is probably a safe one. If on 19 days no 

 B. coli were present, and on the twentieth day 100 c.c. 

 showed 60 B. coli, the average result would be the same, 

 but the water on one day was of a dangerous character. 

 With properly managed filter plants marked variations do 



