RETROSPECT AND PROSPECT 377 
Without citing similar statistics which may be derived from 
other states and most countries of the globe, it may be asked 
if these data really suffice to prove that cancer is actually in- 
creasing. In interpreting most statistics of the increase of cancer, 
allowance must be made for the changing age distribution of the 
population. Both the decline of the birth rate and the increasing 
duration of life make the proportion of people of middle age and 
beyond relatively higher. Hence a larger proportion of the 
population would now be liable to be affected by cancer than in 
previous years. 
Undoubtedly this circumstance explains a part of the statistical 
increase of cancer, but it does not suffice to explain all of it. 
Willcox in fact attributes only about one-third of the reported 
increase to this cause. If we study the death rate for any partic- 
ular age, say 55, estimating the proportion dying of cancer to all 
the population of that age we frequently find that the cancer 
death rate has increased materially in the last few years. This is 
true for most ages in the United States between the periods 
1903-07 and 1908-12 according to the United States Census. 
Data from the life insurance companies of Austria over the 
period from 1876 to 1900 fail to show any consistent trend of 
cancer mortality for most age groups. 
Dr. F. L. Hoffmann on the basis of his extensive and valuable 
collection of statistics on cancer mortality from several countries 
has concluded that there is an actual increase of cancer which 
cannot be explained either by changes in age distribution of the 
population or by improvements in the accuracy of diagnosis. 
Professor Willcox, however, has made a critical study of the prob- 
lem and has come to a quite different conclusion. Most of the 
statistical increase of cancer which cannot be explained by the in- 
creasing proportion of people of middle or old age may be ac- 
counted for, according to Willcox, by improvements in diagnosis, 
and the greater proportion of deaths which are now certified by 
competent physicians. The layman seldom reports cancer as a 
cause of death. Where physicians are relatively plentiful more 
deaths from cancer are put on record. Fewer deaths are now 
