SEDGWICK AND WINSLOW. — BACILLUS OF TYPHOID FEYEU. 509 
exhibits two secondary maxima, one in December or January, the other between 
March and May. If our general theory be correct, there must in these localities be 
some special condition tending to produce typhoid epideuiics in the early winter and 
the early spring, which modifies the normal influence of tiie season. Fortunately, we 
know exactly what this influence is. These five cities — and of the thirty connnnnitles 
we have considered, these five only — draw their water supply from surface sources 
liable to gross pollution. The epidemics of March, 1S99, at Newark; of May, ISOl, 
at Chicago; of January, 1888, and December, 1889, at Paris, as well as the lesser 
winter and spring outbreaks in other years, were unquestionably due to the public 
water supplies of those cities. We have here then a special condition influencing 
the occurrence of epidemics in cities having surface water supplies and therefore de- 
ranging the normal course of prosodemic typhoid. Tlie heavy autumn rains and the 
spring floods consequent on the melting of the winter's snow, carry into surface waler 
supplies a larger amount of pollution than reaches them at any other time, — as is 
own by a comparison of the bacterial content of surface 
We may venture to generalize by saying that winter and sprin 
characteristic of 
ply is most subject to pollution ; they 
absent from communities which use filtered water or water obtained from adequately 
« 
protected watersheds. 
Finally, then, it appears that of the thirty communities we have studied, all but 
four, in which the number of cases is too small to furnish average results, give typhoid 
curves corresponding to one of three types, - the normal temperature distribution 
opical modification, and 
fication due to winter and spring water- 
epidemics. These latter types of distribution are explicable as tbe rosulln„t of a 
combination of the temperature factor with another. We ni.iy thur,.for.. concl.ulo 
that wherever a sufficient number of cases has been considered a dn-..d, relafum 
between typhoid fever and temperature appears to be general and uivarKible. 
IV. CONCLUSION OF THE AUTHOES THAT THE ^^^^^^^J:^ "" 
TYrilOID FEVER DEPENDS MAINLY UPOX SEASONAL ULMl LLA IlI.E. 
The increase of typhoid fever with a gradual rise in the mean air *-U'-<'.- of •^ 
given locality appear; to be a phenomenon so widespread and -S"-'- » '"-^ - 
beyond reasonable doubt some relation between the two factors, ^be ber rcl 
,1 L. .;.. ;n,n,..ct. must be determined by cons.derafons as to the xt.o 0,3 «f 
06 direct or inonecb mwv u^ v.v..^ . ^ i • i„o t^oJuU 
disease and as to the relation of temperature to .he var.ous veh.cle, mamly 
cerned in its transmission. 
