Most physicians think this to be the case, but they have not taken the pains to measure the respective roles, by the exact methods of modern science. S. Adolphus Knopf of New York, an authority on tuberculosis, recognizes the importance of the heredity factor, but says that after this, the most important predisposing conditions are of the nature of unsanitary schools, unsanitary tenements, unsanitary factories and workshops. This may be very true; these conditions may follow after heredity in importance--but how near do they follow? That is a matter capable of fairly accurate measurement, and should be discussed with figures, not generalities. Taking the case of destitution, which includes, necessarily, most of the other evils specified, Professor Pearson measured the correlation with liability to phthisis and found it to be .02. The correlation for direct heredity--that is, the resemblance between parent and offspring--it will be remembered, is .50. As compared with this, the environmental factor of .02 is utterly insignificant. It seems evident that whether or not one dies from tuberculosis, under present-day urban conditions, depends mainly on the kind of constitution one has inherited. There is no escape, then, from the conclusion that in any individual, death from tuberculosis is largely a matter of natural selection. But by taking a longer view, one can actually see the change to which natural selection is one of the contributors. The following table shows the deaths from consumption in Massachusetts, per 10,000 population: 1851-60 39.9 1861-70 34.9 1871-80 32.7 1881-90 29.2 1891-1900 21.4 1901 17.5 1902 15.9 F. L. Hoffman further points out[60] that in Massachusetts, Rhode Island, and Connecticut, 1872-1911, the decline in the death-rate from tuberculosis has been about 50%. "The evidence is absolutely conclusive that actually as well as relatively, the mortality from tuberculosis in what is the most intensely industrial area of America has progressively diminished during the last 40 years."