Miscellanea 
515 
deaths of females under 1. So then the mean age at death of males and females together for 
each period was determined, the two contributory means being weighted in this proportion 1'25 
to 1. Legitimates and illegitimates were dealt with separately. Then the mean age for the 
whole period 1877-95 for both legitimates and illegitimates was determined. In getting these 
values the three periods 1877-81, 1882-90, 1891-95 were weighted in the proiwrtions 1:2:1 
respectively. In this way was obtained the mean duration of life of legitimates and illegitimates 
irrespective of sex. It remained to combine into a single average the legitimates and illegitimates. 
I found that during the period under discussion there died, on the average, 6'556 legitimate in- 
fants under 1, to one illegitimate infant. Averaging, then, the legitimate and illegitimate means 
previously obtained, with weights of 6 '556 and 1, respectively, I found for the mean duration of 
life of all infants dying under age 1, during the period 1877-95 in Prussia, a value of 113-14 days. 
With a -standard month of 30 days this equals 3'771 months, or -3142 year. Reckoning 365 days 
to the year we have from the days directly the mean equal to •30997 year. 
I conclude that the deaths occurring in the first year of life may be centred at '3 year with 
sufficient accuracy for ordinary purposes*. 
We may turn now to two other practical problems which arise in connection with infantile 
mortality. These are 
1. At what age .shall the deaths occurring in the first month of life be centred ? 
'2. At what age shall the deaths recorded in census returns as falling in the age period 0-5 
years be centred? 
For a solution of the first problem I have taken the Prussian statistics for deaths in the first 
month of life, during the period 1882-90 (Table I.); I find the following values for the means : 
Age at Death of those Dying before completing the First Month of Life. 
Legitimate Illegitimate 
S ? ? 
10-82 days 11-25 days 12-41 days 12-75 days 
From these results it seems reasonable to conclude that the deaths during the first month of 
life may be centred, with sufficient accuracy for practical purposes, at -3 month. 
In order to test this result I have calculated the mean age at death of English infants dying 
under age 1, from the following statistics given in Newsholme's "Vital Statistics t." 
Annual Deatli-rate per 1000 Living at each Alonth of Age. 
Healthy Districts. 
Age in 
Months 
Death-rate 
Age, in 
Months 
Death-rate 
0 
447-51 
6 
70-54 
1 
145-49 
7 
65-97 
0 
102-05 
8 
61 -85 
3 
87-16 
9 
58 -.32 
81-09 
10 
.55-28 
5 
75-54 
11 
52-86 
* The error made by centering each element of the frequency distributions of Table I. at the mid- 
point would tend to raise slightly the mean age at death for tlie year. Hence it is probable that 
•3 year comes closer to the true value than -.31 year. 
t Second Edition, p. 105. 
