HEREDITY 



479 



Unfortunately again, the first and most exhaustive studies 

 were made outside of our field, and mostly in that of human 

 characters, so that the best material for the study of heredity 

 lies in this field. But later studies, in wider fields, lead us 

 confidently to believe that the same general principles control 

 transmissions everywhere, in all races and with all characters. 

 Accordingly the writer will employ any studies that have been 

 made, in whatever fields, that offer valuable material either for 

 elucidating principles or for illustrating methods of study. The 

 fullest data of all are those collected by Galton, dealing princi- 

 pally with stature, and they will be freely employed for both 

 purposes. 



SECTION III THE REGRESSION TABLE 



As already seen (chap, xii), when a representative popula- 

 tion of any race is arranged in the form of a frequency distribu- 

 tion we are able to deduce exceedingly accurate expressions for 

 variability. 



If now this distribution be separated and assorted according 

 to parentage, we shall have a series of distributions, each of 

 similar parentage, the whole presenting the best facilities possi- 

 ble for the study of heredity. Such a tabular arrangement con- 

 stitutes what is called a " regression table," and inasmuch as 

 all regression tables present the same general features, we look 

 upon them with confidence as affording reliable data for the 

 study of this most important but otherwise most difficult and 

 apparently self-contradictory subject. 



In all regression tables a scale of values (measurements, 

 weights, numerals, or other valuation) is provided at one side 

 for the parents, and a corresponding scale along the top for the 

 offspring, or vice versa. 1 



The offspring, considered as adults, is then distributed, each 

 individual being recorded opposite the value representing his 



1 Obviously the parental measurements may be arranged along the top and 

 those of the offspring at the side. Every observer follows his fancy in this 

 respect, and as a matter of fact the tables are made in both ways. The writer 

 has become more accustomed to the one described in the text, and for no other 

 reason prefers to arrange the parental values at the side. 



