576 THE AMERICAN NATURALIST. [Vol. XXXIX. 



cussion of " Reduction of Chromosomes "). The important 

 point for us is the behef that the appearance of the bivalent 

 chromosomes during reduction is due to the temporary union of 

 somatic or sporophytic chromosomes in pairs and further that 

 the reducing divisions distribute the members of the pair, which 

 are believed to be descendants of the maternal and paternal 

 chromosomes of the previous generation, as organic entities to 

 the generation which is to follow. 



It is difficult to overestimate the importance of this general- 

 ization. If the program prove to be correct as stated above and 

 if the chromosome is established beyond doubt as a self-perpet- 

 uating organ of the cell and a bearer of hereditary characters 

 we have then the possibility of studying the actual manner in 

 which these structures are passed on from one generation to the 

 next and perhaps determine the ratios or combinations through 

 which the distribution is effected. The difficulty of making an 

 exact determination of ratios in any form so far studied lies in 

 our inability to distinguish the chromosomes of maternal and 

 paternal origin. There is much evidence that the pairs of 

 somatic and sporophytic elements, which form the bivalent 

 chromosomes of the reduction mitoses of animals and plants 

 respectively, are of different parentage but we do not know 

 whether or not there is any rule in the arrangement of the pairs 

 on the spindles of these mitoses although this is hardly to be 

 expected. Cannon (: 02, : 03a) and others have held that the 

 mitoses of reduction brought about the complete separation of 

 the maternal and paternal chromosomes so that two of the 

 resultant four nuclei contain chromosomes from one parent and 

 two from the other, and the germ cells are in consequence abso- 

 lutely pure in character. But this view was soon shown by Sut- 

 ton (:o3, p. 233; accepted by Cannon, :03b) to be at variance 

 with the facts of breeding for if germ cells of hybrids are abso- 

 lutely pure there could be no further change by cross-breeding 

 and the first cross would be repeated over and over again with- 

 out any divergence from the type, which is contrary to experi- 

 ence and fact. The pairs of chromosomes are probably arranged 

 in every possible order and the maternal and paternal elements 

 are distributed in every possible combination by the reducing 



