26 THE METHOD OF DEVELOPMENT 



somewhat from that of his parent, the human hand has not altered 

 appreciably during thousands of years. Even when progression or 

 retrogression ensues as a whole, individual members of the race 

 may reverse the process temporarily. Thus B, C, D may exhibit 

 progression ; E retrogression, through reversion to C ; while F may 

 resume the progression and so resemble D. 



48. Moreover, structures possess breadth and thickness as well 

 as length. Their progressions and retrogressions, therefore, are in 

 three dimensions, and are founded on a number of variations, 

 internal and external, quantitative and qualitative, which is practi- 

 cally unlimited. Each of these variations may be independent of 

 all others ; even single cells may vary independently ; so that while 

 progression is occurring in some respects, retrogression may be 

 occurring in others. The apparent result, if we think of the struc- 

 ture as a whole, is chaos. But we are able to avoid the seeming 

 confusion if we think of each variation separately. We then 

 perceive that it must be progressive or retrogressive, that it must 

 consist in a prolongation or an abbreviation of the life-history of the 

 part as presented by the parent. 



49. Lastly, as we have already noted, offspring vary from their 

 parents, not only at the end of development, but during the course 

 of it not only as adults, but also as embryos. Not only do they 

 add sentences (progressive variations), or omit sentences (retro- 

 gressive variations) from the end of the life-history as related by 

 the parent, but they interpolate, or omit, or alter sentences in the 

 body of the work. These interpolated progressive and retrogres- 

 sive variations, like those occurring at the end of development, 

 may or may not be inherited ; that is, they may or may not repre- 

 sent persistent changes in the germ-plasm. But many of them are 

 persistent. They accumulate, and in the course of ages so alter the 

 life-history, especially in its earlier parts, which have longest been 

 exposed to change, that it may, and usually does, become un- 

 recognizable. Nevertheless it ever remains a real history, a real 

 recapitulation, though in part a recapitulation of past error, of 

 additions and subtractions, and of interpolations which are com- 

 pounded of both additions and subtractions. As a result, the 

 human embryo, for example, differs so greatly from its prototypes 

 of the life-history, that we cannot, with any degree of accuracy, 

 trace the early ancestry of our race by watching the develop- 

 ment of the individual. At first it is unicellular ; then it faintly 

 resembles very low multicellular types; next it reproduces in 

 succession, more and more clearly, higher and higher types till it 



