NOTICE. THE Council of the Royal Statistical Society wish it to be understood that the Society is not responsible for the statements or opinions expressed in the Papers read before the Society or inserted in its Journal. JOURNAL OF THE ROYAL STATISTICAL SOCIETY. JUNE, 1908. The PERADVENTURES of an INDIAN LIFE-TABLE. By SIR J. ATHELSTANE BAINES, C.S.I. [Read before the Royal Statistical Society, 14th April, 1908. THE description I am about to offer of the age-distribution and lifechances of the population of India may be appropriately prefaced by a quotation from one of the best-known of our Fellows, and, I believe, an original Member of the Society. "Nothing is more uncertain," he writes, "than the duration of life, when the maxim is applied to the "individual, but there are few things less subject to fluctuation than "the duration of life in a multitude of people." This last feature is the one with which I am chiefly concerned on the present occasion. We all admit, I think, that the experience acquired since Mr. Babbage wrote the above lines has led to the recognition of greater elasticity in the duration of life than was to be inferred from the information available in his time. To a considerable extent, indeed, the changes which have taken place are attributable to causes which were then but slightly operative. Improvement in hygienic conditions, for instance, has made its most marked and rapid progress within the last thirty or forty years, and is to be traced in not only the health but also the longevity of a population. In another direction, again, there are social tendencies, such as the decreasing fertility of marriage throughout Western Europe, which have markedly developed in extent and intensity even during the present generation, and the influence of which, though not yet apparent in the lifetable, reaches far beyond the existing population, and must vitally affect the future of the community. A very few figures will serve to illustrate my point, that tested by statistics, the duration of life, even in the older civilisations of the West, is influenced to a considerable extent by external causes, and that these causes were especially active in the later decades of the nineteenth century. I take, first, the expectation of life, or, to use Dr. Farr's terminology, the lifetime and the afterlifetime in several countries, at representative ages, VOL. LXXI. PART II. Χ the percentage of variation from the preceding life-table being also given for each of the ages selected : : An expanded form of the first part of the above is given in Appendix A. It will be noted that the changes have been less marked in England than in the other countries, but that in nearly all they have been greater amongst the females than amongst males, and amongst the young, who are more sensitive to external influences, than amongst adults. The course of the principal factors contributing to these changes is probably well known to most statisticians, and may be sufficiently indicated by the following summary, dealing with the facts of the last thirty years or thereabouts : : Legitimate Live Births 15 and 50 Years Old Deaths. England and Wales 28.3 19:4 1.7 22.1 21.1 7.7 7.5 6.4 The mutual interaction of these elements tends, no doubt, to retard the appearance of the resultant effect in the computations of a life-table, but it is certain that, in due course, it will make itself felt very materially. I have now to leave the comparatively safe anchorage of Western Europe, the circumstances of which are more or less familiar to us, and where the material for the preparation of lifetables is provided by a combination of census and registration, which, though still a good way from perfection, has undergone the test of many years' experience, and is steadily improving. The statistical voyage for which I am offering myself as personal conductor is to a country of which the population differs from our own almost as widely as our common humanity allows, in regard to the physical and the social conditions under which its millions pass their lives. The main interest of the subject, indeed, lies in this diversity, and it was on this assumption that I proceeded when, at the ebb-tide of our sessional contributions, which, as we all know, recurs with the regularity and at the season of the vernal equinox, I was asked to make good the normal default. In treating of the life-values of India, reliance has to be placed almost exclusively upon the material furnished by the census. The registration of births and deaths is carried out, it is true, in all the territories directly under the Crown as well as in several of the larger States; but, except in some of the chief cities, it is an operation of relatively recent introduction, and the returns, though not without their uses in a chronological series, as will appear later, are at present of little value as aggregates for the year. Then, again, the provincial censuses, which were taken between 1865 and 1872, were of a somewhat tentative character, and not conducted upon a system uniform throughout the country. The field of observation, then, has to be restricted to the results of the three Imperial censuses of 1881, 1891, and 1901, which dealt respectively with a population (returned by age) of 229,700,000, 286,600,000, and 293,300,000. this The return asked for was that of the age in completed years, being in accordance with general custom, though the current year is not unfrequently used, especially in the case of children. The age of infants in their first year was required in months, not for tabulation, but merely as a matter of administrative convenience. The entries were compiled by years up to 5, and in quinquennial groups onward to 60, beyond which a single aggregate was held to meet all requirements Parallel to the above table, a special return of the ages year by year was abstracted in 1891 and 1901 in each Province and State for about 100,000 of each sex, taken indiscriminately from the different divisions of the unit, to serve as an indication of characteristics likely to be obscured in the larger groups. I need not say that the returns are inaccurate. They are so even in the West, and our Registrar-General has recorded that "not improbably the greater number of adults do not know their "precise age, and can only state it approximately." So far as the inaccuracy is the result of ignorance, it should be noted that in Europe the progress of public instruction is reflected in the return, the inaccuracy increasing with the age. In India, from which throughout this review I exclude Burma, only 98 males and 7 females in a thousand of their respective sexes can even read and write. The number able to recollect their age is still less, and the horoscopes prepared at birth amongst the upper and middle classes of Brahmanists are rarely available for reference on this point. Now, the tendency most common amongst the illiterate in regard to their age is to state it as an even multiple of five, a feature not confined to the Indian returns, as the following summary will show : TABLE III.-Number per 1,000 of Total Population. The irregularity varies, it will be seen, with the degree of illiteracy, and to some extent with the age, the latter feature being, in the West, at all events, implied in the former, and tending to decline. The women, I am bound to point out, are, as a rule, more prone to favour the round numbers than the other sex. The Indian figures, it will be seen, show the error in its |