[19a] The first line of Table VIII shews the effect of the removal of families from the first division to the semi-suburban districts in divisions II and III. Such families take with them a good many children; hence the loss to the metropolitan division, and the gain to the two divisions mentioned of both boys and girls.
[19b] The following statement may make this fact clearer:—
| Division. | Births recorded 1841–45. | Population in 1871 born 1841–45. | Proportion as 100 to— |
| I. London | 316037 | 289951 | 91.7 |
| V. South-western | 266860 | 126219 | 47.3 |
| VIII. North-western | 392151 | 279606 | 71.3 |
The survivors, according to the English Life Table No. 3, should be about 62.2 per cent. The low rate of mortality in the South-western counties renders it certain that, but for migrations, the ratio of enumerated population to the corresponding births would be higher than 62.2 in that division; in the other divisions it would be lower.
[20] I am of opinion that the gain in this column, so far as regards London death-rates, is due to the departure of many women when in bad health, some of whom die in the country. But for the effect of such departures, the recorded deaths and the apparent net gain by migrations would reach higher numbers.
[22] The procedure for instance in 1881 might be:—Take the English population at each age in 1871 as already corrected, and introduce the numbers of births returned in 1876–80 and 1871–75 as the first two terms of the series, adding a reasonable allowance for non-registration. Deduct the deaths in 1871–80, duly apportioned under periods of birth. The gross number of either sex enumerated in 1881 being known, the net loss or gain by migrations can then be ascertained. Apportion this in somewhat similar proportions to those observed in 1851–60. The final results will exhibit approximately the distribution of population by ages in 1881.
[23a] It will be noticed that I have paid no regard to the interval of time between the census day and the January preceding, treating the exact decennium as being a near enough equivalent of the interval between census and census.
[23b] These numbers represent the births in the periods mentioned, plus an allowance for omissions, viz. 1¾ per cent. on births in 1866–70, and 2, 2¼, and 3½ per cent. on births in the earlier periods respectively.
[23c] The correction adopted in the case of female births is rather larger than in that of male births. These figures are those returned, plus 2 per cent. on the births in 1866–70, and 2¼, 2¾, and 4 per cent. respectively on the births in the earlier periods.
[24a] Births, plus allowance for those unregistered, say 2 per cent. on births in 1866–70, and 2¾, 3¼, and 4 per cent. respectively on births in the earlier periods.