THE CHROMOSOMES IN HEREDITY. 233 



It is well known that in the eggs of many forms the maternal 

 and paternal chromosome groups remain distinctly independent 

 of each other for a considerable number of cleavage-mitoses, and 

 with this fact in mind the author was at first inclined to conclude 

 that in the reducing divisions all the maternal chromosomes must 

 pass to one pqle and all the paternal ones to the other, and that 

 the crerm-cells are thus divided into two cateo-ories which mis'ht 



O O . ^3 



be described as maternal and paternal respectively. But this 

 conception, which is identical with that recently brought forward 

 by Cannon, 1 was soon seen to be at variance with many well- 

 known facts of breeding ; thus : 



1. If the germ-cells of hybrids are of pure descent, no amount 

 of cross-breeding could accomplish more than the condition of a 

 first-cross. 



2. If any animal or plant has but two categories of germ- 

 cells, there can be only four different combinations in the off- 

 spring of a single pair. 



3. If either maternal or paternal chromosomes are entirely 

 excluded from every ripe germ-cell, an individual cannot receive 

 chromosomes (qualities) from more than one ancestor in each 

 generation of each of the parental lines of descent, e. g., could 

 not inherit chromosomes (qualities) from both paternal or both 

 maternal grandparents. 



Moved by these considerations a more careful study was made 

 of the whole division-process, including the positions of the 

 chromosomes in the nucleus before division, the origin and for- 

 mation of the spindle, the relative positions of the chromosomes 

 and the diverging centrosomes, and the point of attachment of 

 the spindle fibers to the chromosomes. The results gave no evi- 

 dence in favor of parental purity of the gametic chromatin as 

 a whole. On the contrary, many points were discovered which 

 strongly indicate - that the position of the bivalent chromosomes 



'Cannon, W. A., "A Cytological Basis for the Mendelian Laws," Bull. Torrey 

 Botanical Club, 29, 1902. 



2 Absolute proof is impossible in a pure-bred form on account of the impossibility 

 of distinguishing between maternal and paternal members of any synaptic pair. If, 

 however, such hybrids as those obtained by Moenkhaus (Moenkhaus, W. J., " Early 

 Development in Certain Hybrid Species," Report of Second Meeting of Naturalists at 

 Chicago, Science, XIII., 323), with fishes can be reared to sexual maturity abso 

 lute proof of this point may be expected. This observer was able in the early cell 5 



