700 BACTERIOLOGY OF THE SOIL, WATER AND AIR 
very much to chemical purification. In this respect they are unlike the 
slow sand filters which bring about mineralization of the organic con- 
stituents. It is customary to add from I to 1 part of chlorine gas per 
1,000,000 (0.25 to 1 mg. per liter) to the effluent from mechanical filters 
to insure the destruction of harmful bacteria. 
The efficiency of the purification of water or of sewage by the 
method of sand filtration is therefore to be measured chemically and 
bacteriologically. Chemically a complete transformation of complex 
organic compounds (ordinarily determined as albuminoid and "free 
ammonia") to nitrates is an indication that the digestive power of the 
filter is at par. Bacteriologically a disappearance of all bacteria 
derived from human or animal excrement and a great reduction of 
the total numbers of bacteria in the filtered water as compared with 
the unfiltered water is evidence of the bacterial efficiency of the filter. 
The chief source of danger in potable waters is bacterial contamina- 
tion from human sources. A simple inspection of water frequently 
fails to detect contamination, and even a chemical examination may 
not suffice to reveal pollution. Millions of typhoid bacilli may be 
introduced into a liter of water without inducing changes that could 
be detected visually or chemically. The bacteriological examination 
of water, therefore, is from ten to one hundred times more delicate 
than the chemical examination as a means of detecting contamination 
of water with human or animal waste. 
Bacteriological Examination of Water.— A bacteriological examina- 
tion of water requires relentless attention to details, from the collec- 
tion of the sample to its final analysis and interpretation. 
Collection of Sample. — It must be borne in mind that a small volume 
of water— 100 cc. or less— is ordinarily collected as a sample repre- 
senting thousands or millions of gallons, consequently sampling is an 
important detail in the bacteriological analysis of water. The col- 
lecting bottle must be clean and sterile, and the site at which the 
sample is taken must be representative. 
It is customary to obtain a sample of water from brooks, rivers and 
lakes at a distance from the shore, and preferably samples from dif- 
ferent depths should be taken. The bottle must be immersed below 
the surface before water is allowed to enter it, to avoid surface scums. 
If w^ater is taken from faucets or pumps the sample should not be 
collected until a suflRcient flow has been established to make certain 
that the fluid has come directly from the water mains, or from the 
well itself. 
As soon as the sample has been collected it should be examined; 
frequently this is impracticable, and the bottle should be surrounded 
with ice at once and shipped to the laboratory. Ice restrains bacterial 
development for some hours and this maintains the sample at approxi- 
mately its original bacterial content. 
Bacteriological Analysis of Water.— A bacteriological examination 
of water ordinarilv includes a determination of the numbers of bacteria 
