1886.] Family Likeness in Stature. 55 



Figs. 5 and 6. 



HEICHT 





b/ 



INCHES 



MEAN STATURE OF y% 



MEAN STATURE OF ~A 



72- 



CHILDREN OF MID-PARENTS / , 



BROTHERS OF MEN OF A 



71- 

 70- 



OF VARIOUS HEIGHTS A A 



VARIOUS HEIGHTS A W 



-from, Ft. F.F data A)A <*> i 



" = % AL/o* 



jYom/3pec-ial data /)f \ 



69" 



MEAN STATURE ill! 



MEAN STATURE ! ! ! 



68- 







67- 







66- 







65- 



/ / 





64- 



A/ 



*A 



parents have a deviate of the mean of the deviates of all of their 

 family centres will be +§x\ It follows that*lf one parent only has 

 that deviate + x, and if the stature of the other parent is unknown, 

 and, therefore, on the average, mediocre, the mean of the statures of 

 their children will be half the above amount, or J. I cannot test this 

 conclusion very satisfactorily by direct observation, for my data are 

 barely numerous enough to enable me to deal even with the mid- 

 parentages. They are consequently insufficient to deal with a question 

 involving the additional large uncertainty of the stature of one of the 

 parents. I have, however, tabulated the data, but do not think it 

 worth while to give them. They yield a ratio of regression of 0*40 

 instead of 0*33 as above. I disregard it, and adopt the latter, namely, 

 iv=± 



(3). From the stature of men of the same height to the mean of the 

 statures of their mid-parents. By treating the vertical columns of 

 Table III in the same way as we have just dealt with the horizontal 

 lines, we obtain results of the same general form as in the last para- 

 graph but one, though of different values. 



Taking the height of a group of men of the same stature (viz., the 

 " Adult Children ") as given in the line that forms the heading* to 

 the table, we find the median stature of all their mid-parents, whence 

 I deduce in this case tv=^. The apparent paradox that the same 

 table should give results by no means converse in their values for 

 converse degrees of kinship, will be more conveniently examined 

 later on. 



(4.) From the stature of men of the same height to the mean of 

 the statures of all their brothers. In seeking for this I shall at first 

 confine myself to the more accurate special data, reserving to the end 

 a comparison between their results and those derived from the R.F.F. 

 The entries in the column headed "medians" in Table V are 



