270 



EVOLUTION AND ANIMAL LIFE 



somes and the assumption that they represent the physical 

 basis of heredity i. e., that they influence and determine into 

 what the fertilized egg shall develop. Fifteen years ago 

 Wilhelm Roux showed with convincing clearness that the 

 complicated facts of nuclear division, the careful longitudinal 

 halving of the chromatin thread and its equal distribution 

 between the two daughter cells, can be explained only on the 

 basis that the chromosomes possess different structure in dif- 

 ferent parts of their extent, and that these structures, repre- 

 senting tendencies in development, are distributed in definite 

 ways to the daughter cells. Were this not the case, a simple 

 direct mass division of nucleus and cytoplasm instead of the 



complicated process of 

 karvokinesis with its 



/ 



consequent much greater 

 expenditure of energy 

 would serve all pur- 

 poses. 



In the light of this 

 probable individuality 

 and morphological or- 

 ganization of the chro- 



FIG. 154. Formation of the polar bodies shown Tr , osornps t ], p method of 

 diagrammatically. (After Korschelt and Heider.) S > ] 



their reduction in num- 

 ber, preparatory to the fusion of their germ cells, becomes 

 of the greatest significance; to those who may deny this in- 

 dividuality and definite architecture, the phenomena can have 

 no great importance save as concerns a general mass reduc- 

 tion in the amount of the chromatin present in the germ nu- 

 clei. It may be assumed as true, in the majority of cases 

 now accurately known, that the reduction takes place some- 

 where in or near the last two divisions of the germ cells previous 

 to their fusion that is, in the egg, in the divisions forming 

 the polar bodies, and in the sperm, in the last two divisions of 

 the spermatocyte which produce the four spermatids^ out of 

 which develop as many mature spermatozoa. The phenomena 

 are exactly homologous in both cases, as has already been 

 pointed out, differing only in the minor details which do not 

 affect the end result. 



Two peculiar features mark these divisions off from all the 

 others which precede and follow them. One of these is the 



