INHERITANCE OF PARENTAL MODIFICATIONS 71 



soma cells have no descendants. 1 They die without off- 

 spring. Moreover, apparently no germ cell has ever been 

 anvthinQ- but a nascent o-erm cell. It has never been a 



^ o o 



muscle cell or a nerve cell. Muscle cells, or any other highly 

 differentiated soma cells, do not change into germ cells. 



We can leave out of account the processes of asexual 

 reproduction (fission, budding, reproduction by asexual 

 spores), for, while modifications of the soma of the parent 

 could pass from parent to offspring by these processes of 

 asexual reproduction, the modifications, if unable to be inher- 

 ited through sexual reproduction, would be lost whenever, 

 perhaps after several asexually produced generations, sexual 

 reproduction should intervene ; and we know of no species 

 of multicellular animal, or higher plant, which reproduces 

 indefinitely by asexual methods. Sexual reproduction inter- 

 venes sooner or later. The fact that asexual reproduction 

 occurs does not, then, alter the general argument in regard 

 to the inheritance of parental modifications. 



The phenomena of regeneration would be of some inter- 

 est in this connection, if we knew well-authenticated instances 

 of animals regenerating their reproductive organs, forming 

 from soma cells the new germ cells to take the place of those 

 which had been lost. We do not know, however, that such 

 regeneration is customary, or even possible, in any group of 

 animals. Certainly it is not of sufficient frequency to be 

 taken into account as a means by which soma cells might 

 impress their character upon germ cells and thus secure the 

 inheritance of parental modifications. 



1 As the diagram shows, the body (soia) of the ' parent " and the body (sowa) 

 of the " child '' are a sort of second cousins, or rather " first cousins once removed, 1 ' 

 being related only through the germ cells of the parent's parents. 



