A. O. POWYS 
259 
fertility and also that of greatest physical development, at least so far as stature 
is concerned*. I have therefore determined the correlation for each of these 
sections with the following results : 
Age 
at Marriage 
15 to 2 J/, years 
25 to 39 years 
Mean Age at Marriage 
20-0104 years 
28-7099 years 
Mean Duration of life beyond 
Age 46 20-6745 „ 
23^6708 „ 
Standard Deviation H^'- ^^^"f ^''^f J " 
[Duration of Life -6221 „ 
3-5277 „ 
1-0769 „ 
Correlation ... 
-•0865 ±-0124 
+ -1844f01822 
Coefficient of Regression 
-•02284 
+ -05629 
(Origin at Age at Marriage (Origin at Age at Marriage 
Regression Straight Line 
•< 14 years 
■< 24 years 
( ?/= 20-8118 --0228.V ( ?/ = 23-4057 + •05629.r 
In these two series we see 
that the correlation is 
low, though sensibly within 
the limits of probable error. The coefficient of regression shows that there is 
little difference in the average duration of life beyond the reproductive period 
by marriage at any age under 2.5 years — the mean being 20-01 years; also that 
there is but little difference by marriage at any age between 25 and 39 — the 
mean of which is 23'67 years. But there is a most decided advantage in longevity 
by deferring niari-iage until after age 25 — a difference of 3| years — and as these 
are the results of over 4000 observations, they must be accepted as reliable and 
not merely as accidental. In comparing these results with those of Section A, 
it must be remembered that in these observations the average lifetime is curtate, 
i.e. only completed years have been counted — the year of death is not considered — 
and that the age 46 is the origin. To bring these into line, therefore, with those 
of Section A, \\ years must be added to the present results. We see from 
Table XVII. that the average number of offspring to marriages contracted under 
age 25, and in which the wife at least survives to age 46, is 7'40, and that of 
marriages contracted between ages 25 and 40, 4'46. If now we turn to Fig. 3 
of Section A we find from the parabola that the average durations of life after 
age 45 to mothers of 7'40 and 4'46 children are almost identical, viz. 23"4 years, 
whilst (adding the corrections referred to above, viz. 1 J years) the average durations 
of those who marry before and after 25 years of age are 22-2 and 25'2 years 
respectively. These differences I think show that the age at marriage, as well 
as the number of offspring, is a factor in connection with the duration of life 
beyond the reproductive period. During the reproductive period it is of course 
obvious that as each successive birth has its risk, mothers of large families imperil 
their lives more than those of smaller ones — not simply in proportion to the 
number of births, but after the first confinement in a constantly increasing ratio. 
From the " Vital Statistics " of New South Wales, I have been enabled to deter- 
mine the risk incidental to each confinement. In one table is to be found the 
order of confinement — first, second, etc., whilst in another are to be found the 
* Biometrika, Vol. i. Part i. p. .33. 
33-2 
