ON THE CORRELATION BETWEEN THE "CORRECTED" 
CANCER AND DIABETES DEATHRATES. 
By C. a. CLAREMONT, B.Sc, Biometric Laboratory, University College. 
In a paper due to Maynard published in this journal* it is shown that in the 
United States in the case of both states and great towns there is a marked correlation 
between the corrected deathrates from cancer and diabetes. The point is of very 
great interest, and additional investigations by Pearson and others | and by 
Greenwood and Frances Wood J confirm Maynard's result for the material used 
by him. On the other hand Greenwood and Wood show that it is not apparently 
true for Switzerland. 
Maynard writes §: "If the increased rates observed in the cases of cancer 
arid diabetes were due to a common cause then it is probable that their rates of 
growth will be found to be fairly highly correlated. Not being able to obtain 
rates for a sufficient period of time from the United States reports, the rates given 
in the Registrar's Report for England and Wales, in five-yearly groups for the 
35 years 1871 to 1905, were used and here p = -8060 ± -0893. This high correlation 
shows I think a strong probability that there is a common factor influencing the 
increase of both diseases. That this value is explicable on the assumption that 
the increased rates are merely apparent and due to more careful diagnosis is in 
view of the facts already mentioned almost inconceivable." 
In view of the development of the variate difference correlation method since 
the publication of Maynard's memoir, it seemed possible to test this point, and 
at the suggestion of Professor Pearson I undertook the necessary calculations ||. 
It is to be noted that it is partly the continuous rise in the deathrates from these 
two diseases during the last few decades that has suggested a possibly organic 
relation between them. This increase is effectively shown in the Diagrams of the 
corrected deathrates I and II for these diseases, and in the combined "spot" 
diagram for individual years, III, which brings out the contemporary rises. 
* Biometrilca, Vol. vn. pp. 276-304. 
■j" Journal of R. Statistical Society, Vol. 73, p. 534. 
J Journal of Hygiene, Vol. xiv. p. 83, 1914. § loc. cit. p. 289. 
II I have heartily to thank Miss B. C. B. Cave for revision of my arithmetic and the correction of 
several errors. 
