SEDGWICK AND WINSLOW. — BACILLUS OF TYniOID FEVER. 507 
III. INTEEPRETATIOX OF THE STATISTICAL KESUT/rS. 
■ 
An examination of the curves plotted as above described allows (hat a vcrj 
king parallelism exists between the monthly variations in temperature and tj i)lioid 
P 
prevalence. Of the thirty communities considered, elg-Ii teen show this 
be almost perfect; these are the Empire of Japan, the States of New York and iMassi- 
chusetts, the District of Columbia, and the cities of Atlanta, Baltimore, Bcilin, IJoslou, 
Buenos Ayres, Denver, Leipsic, London, iMobile, Montreal, New York, St. Paul, 
San Francisco, and Santiago. Three other typhoid curves — those for India, for 
Charleston, and for New 
perature in spring, and f 
it in autumn, but show a temporary decrease in the disease during the time of great- 
est heat. In all these twenty-one cases the connection between tlic two factors 
seems too close not to indicate a vital relation. In the northern cities — Montical, 
Boston, Denver, and St. Paul — the curve of typhoid is acute ; in cities with a more 
and more equable temperature the curve of the disease is progressively flattened, the 
limit being reached in tlie case of San Francisco. In the northerly localities the 
maximum occurs in September and October; in the southern cities, with a milder 
winter, it conies in August (Atlanta) or July (Charleston and Mobile). In the two 
cities of the southern hemisphere the curves of both typhoid fever and temperature 
are exactly reversed. In the case of the tropical and sub-tropical regions— India, 
Charleston, and New Orleans — it appears that the rise with the temperature, after 
beginning in the usual fashion, is checked by some other factor, perhaps strong sun- 
TI 
hght or extreme dryness. (See Plates I.-VITI.) 
It remains now to consider the nine cities which show more o 
and to see if their abnormalities are capable of explanation. 
Chicago, Cincinnati, Dresden, Munich, Newark, Oakland, Paris, Philadelphia, and 
Vienna. The first thing to notice in this connection, and the one all previous students 
of seasonal variation have neglected is the necessity of discriminating between sharp 
epidemic outbreaks of the disease and the slow succession of isolated cases which 
characterize that condition known to the older sanitarians ms '^ende 
endemic has been so misused and has become so associated with the idea of 
I 
mysteriousmiasm inherent in a geographical reg 
be safely 
■-i more scientific sense. At the s:une time a distinction, vital to t..e ei.,.,u,„u,.,„,. ., 
■mist be drawn between the infection winch reache. a number of i.erson.s at once 
tI>rot,gh a single medinm as water or n.ilk, and the slower, more con.plex process by 
winch a disease passes from person to person through a population, the patii o. 
