286 
TYPHOID FEVER IX DISTRICT OF COLUMBIA. 
is to improve it considerably as the result of oxidation and subsidence. 
In Tables 18 to 23, inclusive, winch have been compiled from the 
data given in Tables 16 and 17, ^vill be found mimbers representing 
the actual physical and chemical improvement of the water due to 
subsidence and filtration. This improvement is expressed in parts 
per million and also in per cent. 
On passing from the Washington City reservoir through the sand 
filters the Potomac water is still further purified. That such is the 
case mav be seen from Tables 20 and 21. 
It will be observed that the sand filters effect a reduction in the 
turbidity of the water amounting to 90 per cent. The general paral- 
lelism between turbidity and bacterial improvement, as shown by 
the curves in Fig. I, is also a matter of interest in this connection. It 
will also be seen from Tables 20 and 21 that there is a slight reduc- 
tion in the amount of total solids, especially in the amount of nfineral 
matter. The amount of chlorine remains practicall}’ the same as m 
the water of the Washington City reservoir. There is a falling off in 
the amount of the albuminoid ammonia as the result of filtration 
amounting to 43.7 per cent and an improvement in the water as 
shovm by the quantity of oxygen consumed amounting to 33.3 per 
cent. Perhaps the most striking chemical difference, however, 
between the water of the Washington City reservoir and the effluent 
is seen in the quantities of nitrites present in the two waters. It 
will be observed that the effect of filtration has been to accomplish 
the removal of 94.6 per cent of the nitrites present in the water of 
the Washington City reservoir, and yet, as ma}^ be seen from Table 
16, the nitrites show a gradual increase in the slow progress of the 
water through the reservoirs, so that in reality the water of the 
Washington City reservoir contains a larger quantity of nitrites 
than the river water at the Dalecarlia inlet. The effect of filtration 
is to remove the nitrites almost completely. In itself this change 
in the amount of nitrites from 0.0056 parts per million in the water 
of the Washington City reservoir to 0.0003 parts per million in the 
effluent of the filters is insignificant, and 3 ^et, considering the real 
significance attaching to the presence of even verj' small amounts of 
nitrites in natural waters, it is of reall}^ great importance as throw- 
ing light on the chemical changes accomplished hj filtration. The 
disappearance of the nitrites is accompanied b}^ a slight increase in 
the amount of nitrates present in the filtered water as compared with 
the water of the Washington City reservoir, and also by a loss of 
dissolved oxygen amounting to 16.7 to 21 per cent. These changes 
go to show that not only do the sand filters effect a mechanical puri- 
fication of the applied water by straining out any particles that may 
be held in suspension, but also^ that they effect a stiff further purifi- 
cation of the water through processes of oxidation. That such is 
