254. BACTERIOLOGY. 
greatly purified is an indisputable fact. The increase in 
bacteria which occurs from contamination is also largely 
or entirely lost after ten to twenty miles of river flow. 
Nevertheless, the history of many epidemics seems to 
show that a badly contaminated river is never a safe 
water to drink, although with the lapse of time it be- 
comes less and less dangerous, nor will sand filter-beds 
absolutely remove all danger. These statements are 
founded upon the results of numerous investigations; 
thus the marked disappearance of bacteria is illustrated 
by the following: Kummel found below the town of 
Rosbock 48,000 bacteria to the cubic centimetre ; 
twenty-five kilometres further down the stream only 
200 were present—about the same number as before 
the sewage of Rosbock entered. On the other hand, 
the doubtful security of depending on a river purifi- 
cation is proved by such experiences as the following: 
In the city of Lowell, Massachusetts, an alarming 
epidemic followed the pollution of the Merrimac River 
three miles above by typhoid feces, and six weeks later 
an alarming epidemic attacked Lawrence, nine miles 
below Lowell. It was estimated that the water took 
ten days to pass from Lowell to Lawrence and through 
the reservoirs. As typhoid bacilli may live for twenty- 
five days in water, the Lawrence epidemic is easily 
explained. | Newark-on-Trent, England, averaged 
seventy-five cases a year from filtered water and only 
ten when it was changed to deep-well supply. 
The Purification of Water on a Large Scale. Surface 
waters, if collected and held in sufficiently large lakes 
or reservoirs, usually become so clarified by sedimen- 
tation as to require no further treatment so far as its 
appearance goes. The collection of water in large 
