BIOMETRY 373 



parent is gifted by nature, the more rare will be his good fortune if he 

 begets a son who is as richly endowed as himself. " This so-called law 

 of filial regression is represented graphically in Figure 66 in which the 

 actual stature of individual parents is shown by the oblique line, 

 the stature of children by the dotted curve, and the mean stature of 

 the race in the horizontal dotted line. 



Statistical vs. physiological methods. One of the chief aims and 

 results of statistical studies is to eliminate individual peculiarities and 

 to obtain general and average results. Such work may be of great 

 importance in the study of heredity, especially where questions of the 

 occurrence or distribution of particular phenomena are concerned; 

 but the causes of heredity are individual and physiological, and 

 averages are of less value in finding the causes of such phenomena than 

 is the intensive study of individual cases. 



By observation alone it is usually impossible to distinguish between 

 inherited and environmental resemblances and differences, and yet 

 this distinction is essential to any study of inheritance. If all sorts 

 of likenesses and unlikenesses are lumped together, whether inherited 

 or not, our study of inheritance can only end in confusion. The 

 value of statistics depends upon a proper classification of the things 

 measured and enumerated, and if things which are not commensur- 

 able are grouped together the results may be quite misleading and 

 worthless. 



Statistical studies insufficient. Unfortunately Galton and Pear- 

 son, as well as some of their followers, have not always carefully dis- 

 tinguished between hereditary and environmental characters. Fur- 

 thermore much of their material was drawn from a general population 

 in which were many different families and lines not closely related 

 genetically. Consequently their statistical studies are of little value 

 in discovering the physiological principles or laws of heredity. Jen- 

 nings (1910) well says, "Galton's laws of regression and of ancestral 

 inheritance are the product mainly of a lack of distinction between 

 two absolutely diverse things, between non-inheritable fluctuations 

 on the one hand, and permanent genotypic differentiations on the 

 other." In the case of man we have few certain tests to determine 

 whether the differential cause of any character is hereditary or environ- 

 mental, but in the case of animals and plants, where experiments iiay 

 be performed on a large scale, it is possible to make such tests by (i) 

 experiments in which the environment is kept as uniform as possible 

 while the hereditary factors differ, and (2) experiments in which, in a 



