1 3 2 RETROGRESSION 



and this process would continue indefinitely. An obvious interpre- 

 tation of the occurrence of polar bodies is that they are abortive 

 cells abortive because the ovum has retained almost the whole 

 of the cell-body and the contained food, for lack of which the 

 polar bodies perish. Neither of these interpretations can be tested ; 

 they are, from the nature of the case, merely working hypotheses. 

 In science guesses should be avoided if possible; but, at any 

 rate, the guess that the extension of the polar bodies is a 

 mere detail in the preparation of the ovum for conjugation is 

 at least as likely to be correct as the guess that reduction 

 implies a halving of ancestral contributions. Moreover, when uni- 

 cellular organisms unite, mixing their nuclei but not their cyto- 

 plasm, as happens in some cases, there is no extrusion and death 

 of polar bodies. The notion that polar bodies are merely cells 

 which perish from lack of cytoplasm and nutriment is supported, 

 besides, by the fact that, before disintegrating, their chromosomes 

 usually multiply by splitting as in ordinary cell-division. Probably, 

 therefore, extrusion and death are not essential parts of the process, 

 but only the means by which a double or quadruple supply of 

 nutriment is conferred on the ovum wherewith to start the new 

 cell-community. However, as we see, reduction has been held to 

 indicate a halving of ancestral contributions an enormous but 

 untestable deduction. 



219. On the other hand, suppose by an ' ancestral contribution' 

 is meant, not a unit representing an ancestor in the germ-plasm 

 and added to it by him, but a variation which has occurred in it 

 at the time he came into existence. Then, since each individual, 

 on the average, contributes a quarter of the heritage of his child, 

 we must suppose that he varied from his parents to the extent of one- 

 quarter of the sum of his characters. His parents did the same, 

 and so does the child. Therefore, on the average, the child varies 

 to the extent of three-quarters from its great-grandparents. But 

 were this true of all characters, so rapid would be the rate of racial 

 change, that a human being, for example, would become non- 

 human in three generations. If it be argued that the variations 

 of the individual from his parent do not imply variations from 

 remote progenitors, then how are the several 'contributions' 

 to be distinguished from one another ? How is it known that, on 

 the average, each parent contributes a quarter, each great-grand- 

 parent a sixteenth, and so on ? 



220. The immense rate of change indicated by Galton's 

 statistics is due apparently to the fact that he investigated only 



