388 



SCIENCE 



[N. S. Vol. XXVIII. No. 717 



fully admit that the structure of an ovum 

 may in this way be altered, and give rise 

 to a variation which may be the starting- 

 point of a new species. 



But how can a new species originate ac- 

 cording to an epigenetic theory? How 

 can a change in the later stages of onto- 

 geny produce a permanent alteration in 

 the germ-cells? Our answer to this ques- 

 tion will depend on our views of the struc- 

 ture of the germ-cells. According to the 

 mnemic theory they have the quality which 

 is found in the highest perfection in nerve- 

 cells, but is at the same time a character 

 of all living matter— namely, the power 

 of retaining the residual effects of former 

 stimuli and of giving forth or reproducing 

 under certain conditions an echo of the 

 original stimulus. In Semen's phraseol- 

 ogy germ-cells must, like nerve-cells, con- 

 tain engrams, and these engrams must be 

 (like nerve-engrams) bonded together by 

 association, so that they come into action 

 one after another in a certain order auto- 

 matically, i. e., in the absence of the orig- 

 inal stimuli. 



This seems to me the strength of the 

 mnemic theory— namely, that it accounts 

 for the preformed character of germ-cells 

 by the building up in them of an organ- 

 ized series of engrams. But if this view 

 has its strength, it has also its weakness. 

 Eoutine can only be built up by repetition, 

 but each stage in ontogeny occurs only 

 once in a lifetime. Therefore if ontogeny 

 is a routine each generation must be 

 chemically connected with the next. This 

 can only be possible if the germ-cells are, 

 as it were, in telegraphic communication 

 with the whole body of the organism; so 

 that as ontogeny is changed by the addi- 

 tion of new characters, new engrams are 

 added to the germ-cell. 



Thus in fact the mnemic theory of de- 

 velopment depends on the possibility of 

 what is known as somatic inheritance or 



the inheritance of acquired characters. 

 This is obvious to all those familiar with 

 the subject, but to others it may not be so 

 clear. Somatic inheritance is popularly 

 interesting in relation to the possible in- 

 herited effects of education, or of mutila- 

 tions, or of the effects of use and disuse. 

 It is forgotten that it may be, as I have 

 tried to show, an integral part of all evolu- 

 tionary development. 



weismann's theoet 

 Every one must allow that if "Weis- 

 mann's theory of inheritance is accepted 

 we can not admit the possibility of somatic 

 inheritance. This may be made clear to 

 those unfamiliar with the subject by an il- 

 lustration taken from the economy of an 

 ant's nest or beehive. The queen,^ on 

 whom depends the future of the race, is 

 cut off from all active experience of life: 

 she is a mere reproducing machine, housed, 

 fed and protected by the workers. But 

 these, on whom falls the burden of the 

 struggle for life and the experience of the 

 world generally, are sterile, and take no 

 direct share in the reproduction of the 

 species. The queen represents Weis- 

 mann's germ-plasm, the workers are the 

 body or soma. Now imagine the colony 

 exposed to some injurious change in en- 

 vironment; the salvation of the species 

 win depend on whether or no an improved 

 pattern of worker can be produced. This 

 depends on the occurrence of appropriate 

 variations, so that the queen bee and the 

 drones, on whom this depends, are of cen- 

 tral importance. On the other hand any 

 change occurring in the workers, for in- 

 stance, increased skill due to practise in 

 doing their work or changes in their struc- 

 ture due to external conditions, can not 

 possibly be inherited, since workers are 

 absolutely cut off from the reproduction 



° Nor do the drones sliare the activity of the 

 workers. 



