124 



SCIENCE. 



[N. S. Vol. I. No. 5. 



enough to have anj^ evohitionary signifi- 

 cance we should surelj' find all the mem- 

 bers of a species descended from a few re- 

 mote ancestors, and these few the common 

 ancestors of all. If one metazoon is de- 

 scended from pre-Cambrian unicellular an- 

 cestors, the same unicellular individuals 

 were the common ancestors of all the meta- 

 zoa, and we may be confident that there 

 were not verjr many of them in each gen- 

 eration. It is quite possible that they were 

 even so few as a single pair or even one. 



There is nothing very novel in all this. 

 Galton has himself devoted an appendix to 

 the mathematical study of the extinction of 

 family names, although he and other writers 

 on inheritance seem to forget it when thej^ 

 assume that the remote ancestors of two 

 persons, A and B, were, like the parents, dis- 

 tinct individuals, and that the offspring must 

 have twice as much ancestry as either 

 parent, and, therefore, twice as much va- 

 riety, unless there is some way to cancel out 

 half of it at each step. 



I called attention to the bearing of this 

 convergence of ancestry on the problem of in- 

 heritance in 1883, in words which still seem 

 to be a clear statement, although the views 

 on variation of both Galton and Weismann 

 are based on the unfounded assumption 

 that each sexual act brings together two to- 

 tally dissimilar sets of factors, instead of fac- 

 tors which are identical in innumerable 

 features for each one in which thej^ differ. 



My statement is as follows : " In order to 

 breed together, animals must be closely re- 

 lated ; they must belong to the same species 

 or to two closely allied species. Since the 

 individuals which belong to two closely re- 

 lated species are the descendents of a com- 

 mon and not very remote ancestral species, 

 it is clear that almost the whole course of 

 their evolution has been shared by them in 

 common ; all their generic characters being- 

 inherited fi-om this ancestor. Onlj^ the 

 slight differences in minor points which dis- 



tinguish one species of a genus from another 

 have been acquired since the two diverged, 

 and not even all of these slight difterences.- 

 * * We know that the duration of even the 

 most persistent species is only an infinites- 

 imal part of the whole historj' of their evo- 

 lution, and it is clear that the common char- 

 acteristics of two allied species must out- 

 number, thousands of times, the differences 

 between them. It follows that the parents 

 of any possible hybrid must be alike in 

 thousands of features for one in which they 

 differ. * * Crossing simply results in 

 the formation of a germ bj^ the union of a 

 male and a female element derived from 

 two essentially similar parents, ^^•ith at most 

 only a few secondary and comparativelj' 

 slight differences, all of which have been 

 recentljT acquired. ' ' 



I trust that you will not think me un- 

 warranted in the assertion that due consid- 

 eration of the substance of this extract might 

 have saved us much unprofitable discussion 

 of the causes of variation, for I hope I have 

 made it clear that these must be sought in 

 the modern world and not in the remote 

 past; that, as I expressed it in 1883, " the 

 occurrence of a variation is due to the direct 

 action of external conditions, but its precise 

 character is not." 



I sought hy these words to express the 

 familiar fact that the stimulus under which 

 a Yiial action takes place is one thing, 

 while the character of the action itself is 

 quite another thing. 



This fact seems, Irom its verj^ simplicity, 

 to slip out of the minds of naturalists, and I 

 should like to improve this opportunitj^ to 

 approach it fi-om another standpoint. 



We have been familiar for many years 

 with two views of the nature of the process 

 of development from the egg. 



One school of embryologists holds that the 

 organism arises ft-om the egg by virtue of 

 its inherent potency; that the constitution 

 which the germinal matter has inherited is 



