28 



Eeva Jalavisto 



of external bias arising from the 100-year time span necessary 

 for studies of total life expectancy has also to be applied in the 

 paternal age series. 



The percentage distribution of deaths at various ages (0, 1, 

 2-4, 20-29 and 2-29 years) in different age groups of fathers 

 is calculated. The paternal age groups are formed so that the 

 limits are five years higher than in the maternal series, and the 



30r 



25 



20 



^ 15 



10 



Total 



30 



c25 



X Dead during first year of life 



• ■■ second 



O •■ •• 3rd & 4th 



D •• between ages 2-29 years 

 ^ 20-29 •• 



<30 30-34 35-39 UO-UU 245 <30 30-34 35-39 40-U 245 



Paternal age, years 



Fig. 5. Distribution in various paternal age groups of deaths of 



children under 30 years of age. Total series, 5,590 cases; 2,735 



were born before 1800. 



last group extends ten years beyond the highest maternal age 

 group. 



The result is seen in Fig. 5. It may be seen at a glance that 

 differences in paternal age do not affect the distribution of 

 cases at any age : there is hardly any excess of infant deaths 

 with fathers of advanced ages. The difference between the 

 maternal and paternal age distribution of infant deaths 

 strengthens the impression that the maternal age effects are 

 not artifacts, and that this method of study is practicable. 



