CHAP, xvm A "FAIRY TALE OF SCIENCE"? 243 



germ plasm has been described as so continuous or 

 so stable, that it has threatened to make all the off- 

 spring of the same pair identical with each other 

 if the two parental germ plasms are simply added 

 together. But now, wise nature casts away half 

 the qualities or potentialities of the germ plasm, 

 when it throws away half the substance, and the 

 dividing line is drawn at random, or at any rate is 

 never twice the same. Weismann's later view, to 

 which Romanes had thought that he was bound to 

 come and on which Romanes looks with less dis- 

 favour seems to involve the same difficulty. How 

 can cell segmentation divide the germ plasm into 

 different potentialities, corresponding to differences 

 exhibited later in the different members of the litter 

 or family, if we are to hold to the high stability of 

 germ plasm ? Or how on earth can we reconcile this 

 with the doctrine that amphimixis is the only source 

 of variations ? Moreover, are we to understand that 

 germ plasm, "which grows very rapidly," never grows 

 at all, or never segments at all, after birth ? If it did, 

 apparently it would be constantly changing its quali- 

 ties. It would be highly z/Tzstable. 1 



Nature then, according to Weismann, had been 

 playing an immense game of permutations and com- 

 binations, if not since the dawn of life itself, yet 

 ever since the first origin of multicellular organisms, 

 whether plant or animal. All of these become uni- 

 cellular at the beginning of the embryonic process, 



1 The polar bodies had to serve as the explanation of a second 

 difficulty one of size. It also is mysterious. On it also Weismann 

 has changed his ground. And by that change also he secures greater 

 approbation from Romanes. 



