310 PROBLEM OF THE ORIGIN OF SPECIES. 



ever, need not necessarily alter our conception of the part these variations 

 may play in evolution. It simply brings them into line with the general 

 phenomena of variation.^ 



We must conclude, in the light of the evidence herein gathered and corre- 

 lated, that all variation represents solely the response of plastic organic mate- 

 rial to stimuli according to the method of trial and error. Permanent variations 

 are also of the same origin, and, as shown in Chapter V, are produced or 

 primarily arise in the germ plasm, and later or secondarily appear in the 

 soma, through the production of the soma from the germ. It is shown, too, 

 that these permanent variations or divergences may be small or large, but they 

 obey the law of the distribution of error in their proportionate appearances and 

 distance from the mean. There is, then, no necessary incongruity between 

 gradual small variation and rapid large variation in the origin of species, but 

 the two are the extremes of the same process. That variations are not pre- 

 determined, but are epigenetic, arising as the result of the interaction of 

 preexisting series of stages in the organism and its environment, seems to be 

 the only possible method of interpreting these experiments. 



The Origin of Spkciks. 



On the basis of our present knowledge the problem of the origin of species 

 divides itself into two chief questions — the origin and the preservation. Biol- 

 ogists at present are divided as to the origin, one school, the neo-Lamarckians, 

 maintaining that variations may first arise in the soma through mechanical 

 or chemical causes, and later be transferred to the germ ; the other, that they 

 arise in the germ plasm, and later appear in the soma. All acknowledge that 

 the variation, to become transmissible, must ultimately come to be incor- 

 porated into the germ plasm. There exists at present not one single fact 

 to show the inheritance of acquired somatic variations or their incorpora- 

 tion into germ plasm. Moreover, the neo-Lamarckians are at fault in over- 

 looking the very fundamental fact that the variations found to arise in the 

 soma are possible only because the germ-plasm constitution has developed 

 a soma in which variation in certain directions is possible and impossible in 

 others. That is to say, we find a particular variation because the germ plasm 

 has so constituted the soma that it can vary in certain directions and not in 

 others, and in the same direction as the germ plasm itself is varying. It is a 

 curious fact, moreover, that the idea of the inheritance of acquired characters 

 is supported almost entirely by the data of anatomy and paleontology, two 

 lines of investigation which, least of all, are capable of giving data upon the 

 question. Both deal with end results, and can show only what has happened. 



^ In this paper I have avoided any consideration of the hypothesis of unit characters 

 and their isolation as postulated by De Vries. The data and conclusions upon this and 

 allied problems will be presented in another paper. 



