tables nearly that of 4486 to 14,888, and the actual numbers found by the enumeration in 1774, were 4486 and 14,713. In like manner; the number of the living above 70 was, by the same survey, found to be 625; and the tables give this number nearly the same. The expectation at birth, taking males and females together, is at Chester, by the tables, near 31; and therefore one in 31 ought to die annually. But the quotient arising from dividing the number of inhabitants (14,713) by 409 (the medium of annual burials from 1772 to 1781), will shew that in reality no more than one in 30 die annually.The reason of this difference is, first, that the births exceed the burials; and that, consequently, a table which takes the burials for its radix, must give the expectations of life too low.A second reason is, the emigration of males from Chester; in consequence of which, though more males than females are born, and though males are also more short-lived; yet fewer die at Chester, many dying in the army, navy, militia, &c. The effect of the first of these causes will be particularly exemplified hereafter, in the case of the kingdom of SWEDEN. Observations similar to these may be made on the tables in the following collection, formed from a register of mortality at Warrington in Lancashire, founded and conducted by the ingenious Dr. Aikin (then the phy sician there,) to whose kindness and communicativeness, as well as to Dr. Haygarth's, I have been much obliged. See Tables 41st and 42d. The expectation of a male just born, at WARRINGTON, is, by these tables, 20: of a female 254; and of males and females taken together, 23 nearly. S In the beginning of 1781 Dr. Aikin procured an enumeration of the houses and inhabitants in Warrington and its vicinity, consisting of the town of Warrington, the township as far as the lays are collected, Poulton, Fearnhead, and Woolston. The number of houses, including 74 uninhabited, was 2000; of inhabitants 9501, or 4, to a house The number of inhabitants divided by 302 (the annual average of burials for 9 years from 1773 to 1781) gives 314, but divided by 321, the annual average of burials for five years, from 1777 to 1781 (which, in this case, seems the fairest average) gives, 29. There is, therefore, in this town, a greater difference between the proportions dying annually, as determined by enumeration and by calculation from the register, than there is at Chester; and the reason is, that the two causes just mentioned operate more here. The births in particular (the annual average of which for the 5 years just mentio.ed was 411) exceed the burials much more at Warrington; and therefore the burials are much more below the the true average, and the probabilities of living exhibited by the table of decrements, much more below the true probabilities. Every one must be struck with the difference, in respect of longevity, which these tables exhibit between the inhabitants of Warrington and Chester; and it will appear more remarkable when it is considered, that about an 8th or 9th of the inhabitants included in the Warrington bills, are inhabitants of the country for a mile or two round Warrington. Chester appears, indeed, to be an extraordinary exception to the hurtful effects of towns on the duration of life. The probabilities of living in it, though lower than in country parishes, are considerably higher than in any other city where observations have been made. I am not qualified to explain the causes which give it this distinction. A probable account of them has been given by Dr. Haygarth, in a paper printed at Chester, and containing Observa tions on the Population and Diseases of Chester in 1774. It is farther observable, that these tables agree in exhibiting, in a striking light, the difference between the probabilities of living among males and females. But this difference will appear more evidently from the Tables for Sweden, of which I am next to give an account. There are two sorts of data for forming tables of the probabilities of the duration of of human life at every age. One is fur nished by registers of mortality shewing the numbers dying at all ages. The other, by the proportions of deaths at all ages to the numbers living at those ages discovered by surveys or enumerations.- Tables formed from the former of these data, are correct only when there is no considerable fluctuation among the inhabitants of a place, and the births and burials are equal. When there are more removals from than to a place, and the births exceed the burials, as is almost always the case in country parishes and villages, tables so formed give the probabilities of living too low. When the contrary happens, as is generally the case in towns, they give the probabilities of living too high. But tables formed from the latter of these data, are subject to no errors. They must be correct, whatever the fluctuations are in a place, and how great soever the inequalities may be between the births and burials. I know of no observations extant which furnish the means of forming such tables, except those published by the late Mr. Wargentin in the Memoirs of the Academy of Sciences at STOCKHOLM, in 1776; an ab stract of which I have given in the first additional Essay in this volume; and a continuation of which, from 1763 to 1776, Mr. Wargentin with the greatest goodness, communicated to me some time before his death. These observations are more curious than any that have have been yet published, and leave us little to wish for on this subject, except that similar observations were made in other king-: doms under the direction of men equally able and ingenious with Mr. Wargentin.-It is from the result of all these observations taken together, that I have constructed Tables 44th, 45th, &c. in the following collection.. The Tables for SWEDEN at large, compared with those for STOCKHOLM the capital, confirm, in a very striking manner, all that I have said in this volume, and other parts of this work, of the difference between the duration of life in great towns, and in the country.They likewise furnish the most indisputable evidence for the shorter, duration of the lives of males than of females; and it deserves particular notice, that the tables for Sweden at large differ, in this respect, but little from the tables formed from Dr. Haygarth's Observations at Chester. These observations give sufficient data for calculating, with some correctness, distinct tables of the values of lives among males and females, taken separately and conjunctly; but I have preferred for this purpose, the SWEDEN observations, because (as hath been just observed) more correct in their nature; and because also (being made on the inhabitants of a whole kingdom for 21. years, and the enumeration which gives them their chief value having been repeated at seven different periods) they are much more |