96 THE PRINCIPLES OF HEREDITY 



which were immensely advantageous to the races that 

 possessed them. The first was this tendency which causes 

 parents to reproduce their Kke with variations. 1 The second 

 was the tendency which causes the reproductive elements to 

 be more or less completely indifferent to the action of the 

 environment so far as their hereditary tendencies were 

 concerned. These two tendencies would enable races to 

 adapt themselves to adverse circumstances, instead of under- 

 going degeneration. More or less complex organisms would 

 arise, which, in order to live in the environment in which 

 they were produced, would have to be very similar to the 

 parents, and which could not be similar unless they followed 

 nearly the same developmental footsteps. Recapitulation 

 would thus provide for adaptation to the parental environ- 

 ment, and variability would provide for adaptation to changes 

 in the environment. An inevitable corollary to recapitu- 

 lation and the tendency to vary about a mean would be the 

 occurrence of regressive variations. Regressive variation 

 would then render the evolution of still more complex beings 

 possible. An even greater degree of complexity would 

 demand a special selective means of eliminating useless 

 structures and variations. Bi-parental reproduction would 

 then be established, its actual origin being due, perhaps, to 

 attempts at cannibalism amongst unicellular organisms. 



161. It will be interesting to note some phenomena on 

 which the theory of regression set forth in these pages has a 

 bearing. Mr. Galton's Law of Inheritance, which attributes 

 an average of one-quarter of the total heritage of the child to 

 each parent, one-eighth to each grand-parent, and so forth, may 

 be true as regards the particular species investigated by him, 

 if we suppose that by the " contribution " of an ancestor to the 

 heritage is meant reversion to (i. e. failure to recapitulate 

 beyond) that ancestor. But the law is manifestly untrue as 

 regards other species that have been evolved at a different 

 rate. The parents of the pansy, for example, contributed, as 

 a rule, in Darwin's time, nothing to the heritage. They 

 contribute somewhat more now. The average amount of 



1 As to the intracellular mechanism by means of which the offspring 

 are rendered variable, I do not venture to offer an opinion. Certainly, 

 judging from the evidence, the transmission of acquirements, the direct 

 action of the environment, and bi-parental reproduction have nothing to 

 do with it. From a germ-cell, whether fertilized or nut (as in partheno- 

 genesis), arise somatic cells of many different varieties. There is no 

 reason to doubt that the same process of evolution, which caused the 

 somatic descendants of the germ-cell to differ among themselves, has 

 caused the germinal descendants to differ also. 



