G. D. Maynard 
283 
an organism has an existence apart from its human host it is quite likely that this 
should be so. Cancer being a disease of fairly long duration, and of uncertain date 
of onset in relation to the time of death, there was no way of finding out whether 
the disease was contracted more frequently at one time of the year than another. 
If therefore meteorological conditions have an influence on causation, it must 
be sought in place rather than in time. For this purpose I have obtained from 
the U.S.A. Government publication Climatology of the United States, average 
meteorological conditions for 27 Registration cities with populations of over 
100,000 inhabitants. 
Corrected Cancer Death-Rate, 27 Cities, 1900—1904. 
Sunshine in hours r= - -0142 ± -1323 
Mean Temperature ?•=- -1478 ± -1294 
Mean Maximum Temperature r= - -22361 -1257 
Mean Minimum Temperature r= - -01481 -1323 
Rainfall r= - -2517 ±-1239 
Days of Rain /■= - -1621 ± -1288 
Absohite Humidity --2014 + -1269 
Cancer Death-Rate, for the 17 Cities. 
Sunshine >•= -1268 ±-1659 
Rainfall r= -3163 ±-1518 
Mean Temperature r=- -0732 ± -1685 
Mean Maximum Temperature r= - -1509 ± -1600 
Mean Minimum Temperature )•= -1254 ±-1659 
Rel. Humidity 8 a.m. ... r= -3101 ±-1519 
Rel. Humidity 8 p.m. ... r= -1603 ±-1600 
Range of Humidity r= -0724±-1685 
It will be observed that out of the seven correlations appearing in the table 
none of the values are really significant when compared with their probable errors. 
I had previously calculated the correlations for 17 of the larger cities and values 
of the same numerical order were obtained, but in many cases the sign was 
different. They are appended for the purpose of comparison. These values must 
therefore be considered as zero. 
It will be seen from these figures that so far as 27 cities of the U.S.A. are 
concerned there is no correlation between the average climatic differences obtaining 
in the city and the prevalence of cancer, as measured by its average death-rate 
over the five years 1900 — 1904. 
It has frequently been stated that cancer is less prevalent in hot climates, and 
there are facts to be obtained in support of this general statement. It is not how- 
ever necessary to assume that this is due directly to climate, and another explanation 
is, I think, probably the right one. Marked differences of race, temperament, and 
pressure of life are to be found in tropical and sub-tropical countries, as compared 
with temperate and cold ones ; the exception to this being the case now under 
discussion, the United States of America, 
