200 Correlation of Cancer and Diabetes Deathrates 
diabetes. Of course such an investigation as the present does not provide any 
measure of whether persons contracting diabetes are more hable to die of cancer, 
for the deaths would be registered as deaths from cancer, yet this not improbably 
is the vital problem *. The result reached, however, is consistent with a continuous 
increase of both cancer and diabetes not essentially related organically with each 
other. In this case Maynard's results for the United States would mean that 
in that huge population the towns and states were in heterogeneous stages of 
historical development, whether as to medical training, accuracy of record or 
cultural conditions, while the non-appearance of the high correlation in the Swiss 
data would merely signify that the units chosen there were homogeneous in such 
characters, and this considering the size of Switzerland is not improbable. In 
other words some American districts present with regard to cancer and diabetes 
the English conditions of 1870 and others the English conditions of 1910. The 
time correlation of the English deathrates may thus correspond to the geographical 
correlation of the American deathrates. If so, since, when the time factor is 
removed, the English diabetes and cancer deathrates show no organic correlation, 
we cannot use the English material to support the view that the American relation- 
ship is organic in character. The correlations that Maynard has indicated between 
both cancer and diabetes deathrates and the spread of insanity, suicides and the 
newspaper press, seem indeed to indicate not an organic relationship of the two 
diseases, but as Maynard himself has suggested a wider range of both with 
advancing cultural conditions. 
* The enquiry as to whether cancer patients are suffering or have suffered from diabetes should 
always be made and the answer recorded. 
