POTOMAC RIVER WATER SUPPLY RELATIVE TO TYPHOID FEVER. 261 
stand how this wlatinous coat on the surface strains out the bacteria 
o 
contained in the water. ^Yhere this is not formed, as in the Wash- 
ington filters, we must conceive that the organisms are taken out by 
some other mechanical, chemical, or biological process. Perhaps the 
sticky laj^er which coats each particle of sand forms a gelatinous trap 
to which the other bacteria adhere and eventually perish. Whatever 
the explanation, the Washington -sand filters actually take out 87.7 
to 95.6 per cent of the bacteria in the applied water. 
It is well known that bacteria, such as the B. prodigiosus can pass 
through a sand filter; also, that such comparatively large particles as 
3 ^east cells may^ do the same. Particles of clay in sufficient numbers 
to render the water turbid pass through the Washington filters, and 
it is difficult to see, from a bacteriological standpoint, how the filter- 
ing of water through 3 feet of sand without the formation of a 
‘ ‘ schmutz-decke ” can strain out such actively motile organisms as 
the typhoid bacillus. As a matter of fact, however, the sand filters 
eliminate a large proportion of the bacteria, both motile and non- 
motile, contained in the water. Further, sand filtration of other 
river waters giving a similar reduction of the total bacteria, including 
the colon bacillus, has been followed by a reduction of typhoid fever. 
METHODS USED IN MAKING AND RECORDING BACTERIOLOGICAL 
EXAMINATIONS OF WATER. 
All these examinations were made by Assistant Surgeon A. M. 
Stimson, United States Public Health and Marine-Hospital Service. 
The teclmique recommended by the committee on standard meth- 
ods of water analysis of the American Public Health Association on 
Januar}^ 9, 1905, was closely followed. 
Collection . — The specimens were collected in- small ground-glass 
stoppered bottles protected by tinfoil. The taps or pumps were 
allowed to run five minutes before the samples were taken and the 
bottles Vvere filled with the usual bacteriological precautions. 
The samples were at once brought to the laboratory and invariably 
plated within a few hours from the time they Avere collected. When 
the delays of an hour or more was necessar}^ after the arrival of a 
specimen at the laboratory it was placed in the cold room at 15° G. in 
the meantime. 
The following amounts, 0.01 cc., 0.1 cc., and 1 cc., were planted on 
standard nutrient gelatin, incubated at 20° G. and counted after 
forty-eight hours. 
The folloAving amounts, 0.1 cc., 1 cc., and 10 cc., were planted in 
lactose broth fermentation tubes and exposed to 40° G. for forty- 
eight hours, when subcultures Avere made upon lactose litmus-agar 
plates from one of the tubes shoAAung fermentation. If more than one 
of the three fermentation tubes shoAAud gas production, that one Avas 
chosen Avhich, from the appearance of the growth and from the 
