OTHER THEORIES OF SPECIES-FORMING 275 



of this germ cell would continue the struggle among determinants, and 

 it would be expected that the strong determinant would continue to 

 gain further advantage until the structure it represents reached its 

 maximum efficiency. Similarly, a determinant that was for some 

 reason deprived of its fair share of nutriment at any time would be 

 weakened and would produce in cell division weakened daughter 

 determinants. These in turn, unless especially favored, would wage 

 a losing fight and continue to grow smaller and weaker. Each indi- 

 vidual that might develop from such germ cells would have the charac- 

 ters whose determinant had been weakened in a reduced and 

 progressively degenerating condition. Finally, certain determinants 

 might starve entirely, and the part for which they stood would dis- 

 appear entirely from the ontogeny of the indi\ddual arising from these 

 germ cells. 



In this way Weismann tried to explain the gradual dwindling 

 and the final elimination of useless organs. So also he would explain 

 definitely directed or orthogenetic variations, because germinal selec- 

 tion, once started in a given direction, continues automatically till 

 the goal of adaptiveness is reached. 



The most potent objections to the theory of germinal selection are 

 as follows: 



1. There should be, according to this theory, certain pronounced 

 tendencies in variability in definite directions, whereas fluctuating 

 variations nearly always distribute themselves evenly about the mean 

 or mode, and the same specific mean or mode is stationary in succes- 

 sive generations. 



2. The theory implies too rapid and too general modification of 

 parts and therefore does not accord with the fact that species are 

 decidedly constant, except for occasional mutations, over long periods 

 of time. To meet this objection Weismann proposes a new self- 

 correcting mechanism that checks too rapid a development of char- 

 acters. 



3. The over- or undernourishment of determinants might con- 

 ceivably induce size changes in characters already present, but could 

 hardly be responsible for the origin of qualitatively different characters. 



4. Actual experiments in over- and underfeeding of animals have 

 been carried on by certain experimenters in order to test out the theory 

 of germinal selection. In the experiments of Kellogg, for example, 

 involving feeding silkworm larvae only one-eighth of the normal 

 amount of food, the only result was that the mature individuals were 



