A. O. POWYS 
271 
Fig. 1.5. Women Married umler 20. 
8 
7 
6 
^ 4 
'i 3 
2 
1 
la.' 
■b' 
—r^ 
r 
a 
25 
30 
0 5 10 15 20 
Duration of Marriage {Years). 
Kegression Straight Lines : 
aa' (Census), j/ = -3198 + •29894.r 1 Origin at Marriage. 
hb' (Deaths), j/= •7585 + •24796.r j Unit of a; = l year. 
between duration of marriage and number of offspring was determined from the 
figures in the above table with the following results : 
Mean Duration of Marriage 13"7295 3'ears. 
„ Size of Family 4-5031 children. 
^ 1 1 • • (Duration of Marriage 5'9345 years. 
standard Deviation t-. •, ^ 1 -i 1 
[Size of Family 17972 children. 
Correlation 1 0000 
Coefficient of Regression -30286 
Regression Straight Line (Origin at Marriage. Unit of x = I year) 
y = -3450 + •302,S6 x, 
where x= duration of marriage, and y the number of offspring. 
For the purpose of comparison the regression straight line was determined 
from the deaths on a similar basis (age at marriage under 25 years, and duration 
of marriage in quinquennial groupings), and was found to be y = '7360 + '2484 x. 
On comparing the results from the census figures of those marrying at under 
20 with those marrying under 25, we see that in both cases, as in the results 
obtained from the deaths, that the initial fertility is preserved throughout the 
reproductive period, and that it is slightly greater when marriage is contracted 
under age 25 than when contracted under age 20 ; that is, that it is greater at 
age 20 to 25 than under 20. This is in accordance with results arrived at else- 
where* in connection with first births. It is also seen that the mean annual 
* Biometrika, Vol. i. Part i. p. 33. 
