294 A Statistical Stvdy in Cancer Death-Rates 
has been said that Diabetes, or Glycosuria at least is frequent among the insane. 
There is, however, some connection between Diabetes and Insanity, as we shall 
proceed to point out. Dr Maudsley has recorded his experience that it is not 
uncommon to find a history of Diabetes in the parents or near ancestors of insane 
patients ; this is also our experience, especially among the affluent classes." 
G. H. Savage. 
" Dr Snow informs me that independently of actual brain-deposits, insanity 
is not uncommon among cancer patients, and that the incurable wards of the 
Cancer Hospital are seldom without two or three cases." {Diet. Med. Psych., 
Hack Tuke.) 
Several considerations have led me to offer this paper for publication in its 
present form. Firstly, I venture to hope that the subject will be considered 
of sufficient importance to induce other biometrical workers to assist in its 
elucidation ; secondly, the time at my disposal is limited while much remains 
to be done; and thirdly, much of the work is impossible to a resident living 
away from the big libraries, as many books of reference are unobtainable. Even 
such districts as have been dealt with could be better handled by workers 
conversant with the districts and their local conditions. Lastly, the publication 
of a preliminary paper may bring helpful criticism and so enable future work to 
be more profitably planned. 
It would much enhance the value of the Death Reports, if the Registrars 
could see their way to add special tables of the causes of death, in cases where 
the diagnosis has been confirmed by post-mortem examination. A list of 
associated factors found in each case might be added, much on the lines 
adopted by the London Asylums Board. Thus there would in time grow up 
a body of statistical material of the utmost value for determining the true age 
distributions of the various diseases, and for obtaining many other important 
pathological facts. 
I would take this opportunity of thanking Professor Pearson for much kind 
advice, and also Dr J. M'D. Troup and Mr G. Clark for their assistance during 
the compilation of this paper. 
Conclusions. 
1. That recorded differences in cancer and diabetes death-rates, as applying 
to different Districts and Cities of the U.S.A., as well as the increased rates 
observed in recent years, do indicate real differences in the prevalence of the 
disease. 
2. The correlations found to exist between cancer, diabetes, and insanity are 
not fortuitous and due merely to errors of observation or record. 
3. The statistics dealt with in this paper do not give any support to the 
suggestion that cancer is of infectious origin. 
