No. 574] SHORTEB ARTICLES AND DISCUSSION 



637 



r w , = .781±.028. 5 



Thus on a scale of — 1 to +1 the interdependence of mean 

 ages of men and women is very close indeed. 8 Expressing the 

 same relationship in terms of regression by the well known 

 formula 



tion standard deviations of m — males and / = females, 

 /=— 56.474 + 1.009 m. 



Thus we see that each month's increase in average male age is 

 followed by a month's increase in mean female age. The fit of 

 the straight line to the empirical means as shown in the diagram 

 is excellent — considering the small number of the district means 

 from which the equation is deduced. 



Thus the available data show that the moan age at marriage of 

 women instead of being less variable from district to district than 

 that of men is actually more variable — both absolutely and rela- 

 tively. 



In short, there is, as far as our data go, no evidence for the 

 assertion that while the time of marriage of men is closely de- 

 pendent upon the complex of social and economic conditions that 

 of women is. practically independent of them. 



We have published this note in the hope that it may gnggari 

 to some one with the opportunities of obtaining really adequate 

 data an investigation of the problem which has several rather 

 important points of interest. 



where the bars imlir; 



population 



and the sigmas popula- 



J. Arthur Harris, 

 R ox an a H. Vivian 



5 The difference method applied to the ungrouped material gives 

 r = .763 ± .030. 



