126 BLASTOGENIC VARIATIONS. 



tion for all the tissues in the body might thus be just as 

 great as for an individual who to all appearances closely 

 resembled his parents, but in this case chiefly in exter- 

 nal characters and not internal. 



If amphimixis be so largely responsible for the varia- 

 tions observed in offspring, what, then, are the rules 

 which govern the amount and range of these variations ? 

 For the answer to this question we are again primarily 

 indebted to the labours of Mr. Galton. He set himself 

 to determine the exact average relationships between 

 the two parents and their offspring. It might be 

 thought that this was so simple and obvious as to ren- 

 der it waste of time to put it to the test of experiment. 

 It might be thought, in fact, that the average char- 

 acters of offspring are a mean between those of the 

 parents. But this is far from being the case. As Gal- 

 ton first showed, by means of extensive observations on 

 the size of sweet-pea seeds obtained from plants which 

 had been grown from seeds of known size, the average 

 characters of the offspring show a considerable regres- 

 sion towards the mean characters of the race. That is 

 to say, in the present instance, the size of the filial seeds 

 was, on an average, more mediocre than that of the 

 parent seeds. By means of his data, above referred to, 

 concerning the stature of families, Galton was able not 

 only to thoroughly substantiate this phenomenon of re- 

 gression, but to calculate with some degree of exactness 

 the actual amount of regression occurring between 

 various kinsmen. For example, he found that if 

 parents were sorted into groups according to their 

 stature, then the stature of their sons, on an average, 

 deviated only two-thirds as much from the mean stature 



