the various professions; and again, had we these deaths classed according to the distinction which the individuals attained. In addition to this it would be necessary to ascertain (with some rough approximation, as I have attempted to do with regard to the greatest men of all times) the age at which they had accomplished sufficient work to entitle them to be enrolled in their special class. To take concrete instances, let us suppose that we wish to investigate the longevity of American lawyers. Now to be a lawyer in name only requires the candidate to have lived twenty-one years, and the average number of years which the average person of twenty-one years of age will continue to live is about forty; so that the mere fact that a man is a lawyer would bring his average age at death up to sixty-one years. I find in Mulhall's Dictionary of Statistics the statement made that the lawyers of Frankfurt die at the average age of fifty-four years, while merchants live to be fifty-seven years old. I know nothing about the authority of these figures, and am using them for illustration only. Assuming all the data to be correct (and twenty-one seems not too high an age for this purpose), this would seem to suggest that the ordinary lawyer of Frankfurt is not favored with abundance of years. In passing, it is interesting to note that these Frankfurt statistics of lawyers and merchants and other classes show a uniformly lower age at death than those of the more eminent representatives of their professions. This is just what we should expect; for to be included in the one group one must have lived only long enough to prepare and establish one's self as a lawyer or as a merchant; while for the other group one must in addition have had opportunity to cultivate one's ability to a riper fruitage, and in a keen, and often long competition gain public recognition. It thus follows that the average longevity of the most distinguished lawyers will be greater than that of ordinary lawyers, because it takes longer to enter the more select class. But this argument, like many others, should not be pressed too far; innate ability may accomplish in a brief period what for more moderate powers is the work of many years. Nonetheless, in the study of comparative longevity it is the average that is significant; and it is the fluctuation of the average that we aim to discover. Thus, in the investigation of the longevity of an unwholesome occupation, such as would be accepted by a life insurance company only at special rates, we should expect to find the age at death of such individuals less than that of other classes involving an equal period of apprenticeship; but, of course, not less than that of the 'population as a whole.' And, to continue with the main argument, if we wish to investigate the longevity of shoemakers we should again have to decide upon some age at which on the average a person has probably already acquired the dexterity requisite to be a shoemaker. Even if we fix this so low as ten years, at which time the expectation of life is forty-eight years, it would bring the average age at death of shoemakers to fifty-eight years. It has thus become extremely obvious that if we compared these ages at death with the average life-period it would be just as easy to prove that lawyers and shoemakers and merchants enjoy exceptional longevity, as to prove that great men do. The average longevity is low because of the very large infant mortality, which enters into the composition of this average. When once the first ten years of life are passed the further expectation of life increases quite slowly. Roughly speaking, for every ten years between ten and fifty years the added expectation of life is but three years for each decade. We therefore see that in the very nature of things no one class of adults can possibly live as much as thirty years longer than 'the population as a whole.' The differences with which we are dealing are differences of a finer order, of a small number of years, and being slight differences, must be sub-
