PROGRESSION AND RETROGRESSION 25 



complete and accurate history is quite another thing. On the 

 contrary, since variations occur in every stage of development, in 

 every structure, and in every generation, since like other features of 

 the germ-plasm these variations may be persistent, since, therefore, 

 they tend to accumulate during the passage of generations, it is 

 certain that in no single species and even in no organ or structure 

 is recapitulation complete or accurate. The structure we have just 

 considered reached its culmination in M, and began to retrogress 

 in N. Now, obviously, since development is by recapitulation, N 

 cannot retrogress except by an act of incomplete recapitulation ; 

 that is by failing to reproduce the variation by means of which M 

 advanced a step beyond L. Obviously again in that case N reverts 

 to L; in effect he is L; and M, unless his variation becomes latent, 

 a contingency we shall consider later, 1 disappears from the history. 

 Of course, N's act of retrogression may be larger, and may carry 

 him back to K or some remote ancestor, say F or E. O's retro- 

 gressive variation will carry him yet farther. Z, in whom retro- 

 gression is complete, will, in effect, be A, in whom the structure had 

 not even its beginnings. In all M's descendants, therefore, the 

 recapitulation will be incomplete. Even if the structure be evolved 

 again in some of Z's descendants by a fresh series of progressive 

 variations, it will still be incomplete. A chapter of the life-history 

 will be missing even though one like it be added subsequently. 



46. For the sake of clear thinking it must be noted that when 

 we allude to individuals as persisting or disappearing from the 

 life-history, we are using a mere figure of speech. As we have seen, 

 the structures of offspring are not derived from the structures of 

 their progenitors. In reality we are considering alterations in the 

 germ-plasm, which, though capable of change, is continuous and 

 potentially immortal. Germ-cells are unicellular organisms, with 

 unicellular germinal descendants. Multicellular individuals, as we 

 have seen, are merely episodes in the unending career of the germ- 

 plasm, dwellings which it periodically builds about itself and from 

 which it departs as they tend to become outworn. To be precise, 

 then, the alterations in the germ-plasm are such that individuals 

 recapitulate, with more or less accuracy, the evolution of the race 

 of individuals that spring from the slowly changing germ-cells. 



47. The case we have imagined uninterrupted progression 

 followed by uninterrupted retrogression is purely ideal. Really 

 in nature progression tends to alternate irregularly with retro- 

 gression. Thus, though the hand of every child tends to vary 



1 See 186-9 ; see also chapter vii. 



