Take the red thread through any value on the scale of stature, say 74", then the average 

 stature of persons having all the kinsfolk described below of that stature would be obtained 

 by drawing a vertical line through the mark indicated by the kinsfolk till it meets the red 

 thread, and carrying through this meeting point a horizontal line back to the scale of stature, 

 which provides the average desired. For example the average nephew of two uncles, one 

 paternal and one maternal, each 6 ft. 2 in. would be 6 ft. 06 in., but the average nephew of 

 four uncles, two on each side, each 6 ft. 2 in. will be 6 ft. 1-8 in.* Again if both parents 

 and paternal and maternal grandparents were of this stature, the grandsons would have 

 progressed and be on the average 6 ft. 3"1 in. The statures are recorded for males, the 

 corresponding female statures may be obtained by subtracting T g w ths from the male statures. 

 In starting with females the male stature equivalent to that of the female must first be 

 obtained by adding o- :i rds of its value to the female stature. Thus a woman of 5 ft. 9 in. 

 counts as a man of 6 ft. 3 in. 



The reduction from my life-size diagram to the present small dimensions costs much in 

 accuracy of reading, but serves to bring out the point that the regression ultimately changes 

 to a progression. 



* The regression coefficient used on the genometer was •9614. 



