164 MICKO-ORGANISMS IN WATER 



during the "vvhole of the time the filter was working. 

 Eain and snow were excluded from the tank by a cover- 

 ing of canvas towards the end of 1888, and this with 

 the more even distribution of the sewage enabled, as 

 stated above, nearly double the daily quantity of sewage 

 to be treated. 



For a month previous to the application of sewage 

 to this tank it was, as already mentioned, filtering city 

 water at a rapid rate. The number of bacteria in a 

 c.c. of this water (obtained from the Merrimack 

 river) was seventy-six, and in the eSluent twenty-four. 

 After applying sewage which contains from half a 

 million to one million and a half bacteria per c.c, 

 the number in the effluent increased rapidly, till at the 

 end of ten days it was 387,000 per c.c, which in two 

 weeks fell to 1,443. The same general condition of 

 large numbers of bacteria coming through on the first 

 daily application of sewage was found to be the case in 

 all the tanks, irrespective of the filtering material or 

 the protection or exposure of the tanks. The number 

 in the effluent was found to varv from hour to hour, 

 and to depend upon the rate of flow. Thus in Decem- 

 ber, 1888, the number of bacteria in the morning 

 before sewage was applied was found to be forty-sis ; 

 after sewage had been on an hour, and the flow at the 

 outlet had increased twelve times, the number per 

 cc was 1,230. The next sample contained 1,748, 

 and when at the maximum rate of flow — i.e. 50 per 

 cent, greater than the last — there were found 1,640 

 per cc ; after this the number fell off in seven 

 hours to about 200. On another ■ occasion, in the 

 morning the number was ninety-eight ; a little after 

 the maximum rate of flow the number was 16,478 ; 

 two hours later it was 5,000 ; and after five more hours 

 it was 1,300. This circumstance will help to explain 



