43 
water supply — the average of the annual rates was 18.88. Thus 
the reduction in the death rate from all causes since 1900 has been 
progressive. 
There was a greater reduction in. the rate between the 1900-1903 
and 1903-1906 periods than between the 1903-1906 and 1906-1909 
periods. So it is questionable how much of the reduction in the 
last three years has been due to the further improvement in the 
city’s water supply by sand filtration. In this connection it should 
be remembered that an additional storage reservoir (Washington 
City) was put into operation in 1902. 
These data are presented for what they are worth, in view of 
Hazen’s theorem: 
For every death from typhoid fever avoided by the purification of public water 
supplies two or three deaths are avoided from other causes.® 
SANITARY CONDITION OF RESIDENCES. 
In the following table is given the general sanitary condition of 
the residences at which the persons affected had lived at the time 
when the infection was considered to have been contracted, the 
condition of residences of the cases studied in 1907 and 1906 being 
placed in parallel columns: 
Condition of residence.^ 
Number of cases. 
1908. 
1907. 
1906. 
Good 
82 
98 
253 
Fairly good 
206 
228 
243 
Rather bad 
184 
140 
158 
Bad 
70 
57 
92 
Not determined 
0 
0 
1 
Total 
542 
523 
747 
a The definitions of the terms used in this table correspond to those used in our previous reports (Bulle- 
tins 35 and 44). 
The percentages of cases at residences of these four classes for the 
three periods were as follows: 
Condition of residence. 
Percentage of cases. 
1908. 
1907. 
1906. 
Good 
15.1 
18.7 
33.9 
Fairly good 
38 
43.5 
32.5 
Rather bad 
33.9 
26.7 
21.2 
Bad 
12.9 
10.9 
12.3 
a Sedgwick and MacNutt. Science, August 14, 1908, p. 215. 
