422 



A^A TURE 



[September 3, 1908 



workers are absolutely cut off from the reproduction of 

 the race. According to Weismann, there is precisely the 

 same bar to the inheritance of somatic change. 



The racial or phylctic life of all organisms is conceived 

 by him as a series of germ-cells the activity of which is 

 limited to varying, and the survival of which in any genera- 

 tion depends on the production of a successful soma or 

 body capable of housing, protecting, and feeding the germ- 

 cell. Most people would a priori declare that a community 

 where e.xperience and action are separated must fail. But 

 the bee's nest, which must be allowed to be something 

 more than an illustration of Weismann 's theory, proves the 

 contrary. 



It is clear that there must be war to the knife betwee.n 

 the theory of Weismann and that of the somatists — to coin 

 a name for those who believe in the inheritance of acquired 

 characters. A few illustrations may be given of the 

 strength of Weismann 's position. Some trick or trivial 

 habit appears in two successive generations, and the son 

 is said to inherit it from his father. But this is not 

 necessarily a case of somatic inheritance, since according 

 to Weismann the germ-plasm of both father and son con- 

 tained the potentiality of the habit in question. If we keep 

 constantly in view Weismann 's theory of continuity, the 

 facts which are supposed to prove somatic inheritance cease 

 to be decisive. 



Weismann has also shown by means of his hypothesis of 

 "simultaneous stimulation"' the unconvincingness of a 

 certain type of experiment. Thus Fischer showed that 

 when chrysalids of Arctia caja are subjected to low 

 temperature a certain number of them produce dark- 

 coloured insects ; and further that these moths mated 

 together yield dark-coloured offspring. This has been held 

 to prove somatic inheritance, but Weismann points out 

 that it is explicable by the low temperature having an 

 identical effect on the colour-determinants existing in the 

 wing-rudiments of the pupa, and on the same determinants 

 occurring in the germ-cells. 



It does not seem to me worth while to go in detail into 

 the evidence by which somatists strive to prove their point, 

 because I do not know of any facts which are really 

 decisive. That is to say, that though they are explicable 

 as due to somatic inheritance, they never seem to me 

 absolutely inexplicable on Weismann's hypothesis. But, 

 as already pointed out, it is not necessary to look for special 

 facts and experiments, since if the mnemic theory of 

 ontogeny is accepted the development of every organism 

 in the world depends on somatic inheritance. 



I fully acknowledge the strength of Weismann's posi- 

 tion ; I acknowledge also most fully that it requires a 

 stronger man than myself to meet that trained and well- 

 tried fighter. Nevertheless, I shall venture on a few 

 remarks. It must be remembered that, as Romanes " 

 pointed out, Weismann has greatly strengthened his theory 

 of heredity by giving up the absolute stability and per- 

 petual continuity of germ-plasm. Germ-plasm is no longer 

 that mysterious entity, immortal and self-contained, which 

 used to suggest a physical soul. It is no longer the 

 aristocrat it was when its only activity was dependent on 

 its protozoan ancestors, when it reigned absolutely aloof 

 from its contemporary subjects. The germ-plasm theory 

 of to-day is liberalised, though it is not so democratic as 

 its brother sovereign Pangenesis, who reigns, or used to 

 reign, by an elaborate system of proportional representa- 

 tion. But in spite of the skill and energy devoted to its 

 improvement by its distinguished author, Weismannism 

 fails, in my opinion, to be a satisfactory theory of evolu- 

 tion. 



.Ml such theories must account for two things which are 

 parts of a single process but may logically be considered 

 separately : (i) The fact of ontogeny, namely, that the 

 ovum has the capacity of developing into a certain more 

 or less predetermined form ; (ii) The fact of heredity — the 

 circumstance that this form is approximately the same as 

 that of the parent. 



The doctrine of pangenesis accounts for heredity, since 

 the germ-cells are imagined as made up of gemmules re- 

 presenting all parts of the adult ; but it does not account 



1 I borrow this convenient expression from Plate's excellent book, " Ueber 

 die Eedeulung des Darwin'schen Selectionsprincips." IQ03, p, 81. 

 - " An Examination of Weismann," 1893, pp. 169, 170. 



for ontogeny, because there seems to me no sufficient reason 

 why the gemmules should become active in a predetermined 

 order unless, indeed, we allow that they do so by habit, 

 and then the doctrine of pangenesis becomes a variant of 

 the mnemic theory. 



The strength of Weismann's theory lies in its explana- 

 tion of heredity. According to the doctrine of continuity, 

 a fragment of the germ-plasm is, as it were, put on one 

 side and saved up to make the germ-cell of the new 

 generation, so that the germ-cells of two successive genera- 

 tions are made of the same material. This again depends 

 on Weismann's belief that when the ovum divides, the two 

 daughter cells are not identical ; that in fact the funda- 

 mental difference between soma and gcrm-cells begins at 

 this point. But this is precisely where many naturalists 

 whose observations are w^orthy of all respect differ from 

 him. W'eismann's theory is therefore threatened at the 

 very foundation. 



Even if we allow Weismann's method of providing for 

 the identity between the germ-cell of two successive genera- 

 tions, there remains, as above indicated, a greater problem 

 — namely, that of ontogeny. We no longer look at the 

 potentiality of a germ-cell as Caliban looked on Setebos, 

 as something essentiallv incomprehensible ruling the future 

 in an unknown way — "just choosing so." If the modern 

 germ-cell is to have a poetic analogue it must be compared 

 to a Pandora's box of architectonic sprites which are let 

 loose in definite order, each serving as a master builder 

 for a prescribed stage of ontogeny. Weismann's view of 

 the mechanism by which his determinants — the archi- 

 tectonic sprites — coinc into action in due order is, I assume, 

 satisfactory to many, but I confess that I find it difficult to 

 grasp. The orderly distribution of determinants depends 

 primarily on their arrangement in the ids, where they are 

 held together by " vital affinities." They are guided to 

 the cells on which they are to act by differential divisions, 

 in each of which the determinants are sorted into two 

 unequal lots. Thev then become active, i.e., break up into 

 biophores, partly under the influence of liberating stimuli 

 and partly by an automatic process. Finally the biophores 

 communicate a " definite vital force " to the appropriate 

 cells.' This may be a description of what happens ; but 

 inasmuch as it fails to connect the process of ontogeny 

 w'ith physiological processes of which we have definite 

 knowledge, it does not to me seem a convincing explana- 

 tion. 



For myself I can only say that I am not satisfied with 

 Weismann's theory of heredity or of ontogeny. As re- 

 gards the first, I incline to deny the distinction between 

 germ and soma, to insist on the plain facts that the soma 

 is continuous with the germ-cell, and that the somatic cells 

 may have the same reproductive qualities as the germ-cells 

 (as is proved by the facts of regeneration) ; that, in fact, 

 the germ-cell is merely a specialised somatic cell and has 

 the essential qualities of the soma. With regard to 

 ontogeny, I have already pointed out that Weismann does 

 not seem to explain its automatic character. 



The Mnemic Theory. 



If the mnemic theory is compared with Weismann's 

 views it is clear that it is strong precisely where these are 

 weakest — namely, in giving a coherent theory of the rhythm 

 of development. It also bears comparison with all theories 

 in which the conception of determinants occurs. Why 

 should we make elaborate theories of hypothetical deter- 

 minants to account for the potentialities lying hidden in the 

 germ-cell, and neglect the only determinants of the exist- 

 ence of which we have positive knowledge (though we do 

 not know their precise nature) ? We know positively that 

 by making a dog sit up and then giving him a biscuit we 

 build up something in his brain in consequence of which 

 a biscuit becomes the stimulus to the act of sitting. The 

 mnemic theory assumes that the determinants of morpho- 

 logical change are of the same type as the structural -altera- 

 tion wrought in the dog's brain. 



The mnemic theory — at any rate that form of it held 

 by Semon and by myself — agrees with the current view, 

 viz., that the nucleus is the centre of development, or, in 

 Semon 's phraseology, that the nucleus contains the 



1 "The Evolution Theory," En^. trans., i. 373 ei seq. 



NO. 2027, VOL. 78] 



