APPENDIX E 245 



atavism, must be very slight— so slight as usually to be inappre- 

 ciable. Therefore it is only by observing the retracement, not in 

 an individual, but in a line of individuals, that it becomes plainly 

 noticeable. It is by taking advantage of such retracement that 

 " Reversed Selection," as it has been termed, eliminates a structure, 

 which a change of environment has rendered not only useless, but 

 worse than useless, more rapidly than would otherwise occur under 

 the mere absence of selection. For example, Natural Selection 

 has resulted in the evolution of eyes. In animals dwelhng in 

 absolute darkness, e.g. certain cave-dwellers, the eye has become 

 not only useless, but worse than useless, since it is an extremely 

 prominent and tender, and therefore vulnerable, part of the 

 organism. In some such animals we observe that the eye is 

 better developed in the embryo than in the adult. Clearly here 

 the animal in its ontogeny retraces some of the steps it has already 

 made. Clearly, also, if ontogeny be a recapitulation of phylogeny, 

 such retracement was made in the phylogeny as well. It follows 

 that when a structure, useless both to the embryo and the adult, 

 is better represented in the former than in the latter, it must have 

 undergone retrogression through the action of Reversed Selection, 

 and that during the phylogeny, after being useful, it became not 

 only useless, but worse than useless. 



The second proposition, that an individual may so vary from 

 his parent as not to recapitulate the later stages of the phylogeny, 

 and that this constitutes atavism, is the main proposition of the 

 present thesis ; but I have yet to prove that this atavism is the 

 cause of true retrogression. 



True atavism can seldom be observed in such of the higher 

 animals and plants as have been evolved under Natural Selection, 

 not because it does not occur, but simply because it is usually 

 masked and slight. It is masked, because such complex beings 

 seldom or never retrogress in all their characters at once, and, 

 therefore, such reversion as may occur in this or that particular, 

 is associated with evolutionary variation in other particulars. It 

 is slight because, since such species have evolved but slowly, 

 reversion to a not very remote ancestor does not result in any 



