342 COOK 



chromosomes are not the ultimate units of the nuclear structure, 

 but are merely aggregates of granules of chromatin. In the 

 final stage of conjugation (mitapsis) the chromosome aggregates 

 no longer appear distinct, but are subdivided into small clusters 

 of granules called chromomeres. The chromomeres are strung 

 out like beads in single file along two slender, protoplasmic 

 threads which finally lie parallel and close together, so that the 

 individual chromomeres can be paired off and fused with each 

 other. Instead, therefore, of thinking of conjugation as a 

 simple bulk fusion of protoplasm or of nuclei, we must view it 

 as involving a long line of many scores, hundreds, or even 

 thousands, of contacts or combinations between the much smaller 

 granule-groups or chromomeres. Chromomeres appear, there- 

 fore, to have important physiological functions as specialized 

 contact points in the fusion and reorganization of the protoplasm, 

 and do not need to be thought of as bearers of hereditary char- 

 acter-units. 



There remains one other stage of elaboration of mathematical 

 hypotheses of heredity, to treat the chromomeres as permanent 

 entities of descent and deduce the infinitely multifarious diver- 

 sities of individuals in nature from the infinity of combinations 

 and rearrangements of which the chromomeres may be capable. 

 This theory is complete and unimpeachable mathematically, 

 but is as indefensible biologically as its predecessors ; for like 

 them it rests on the assumption that the bringing of the chro- 

 matin granules into contact in mitapsis has no significance in 

 descent. It takes for granted that nothing of importance oc- 

 curs when the granules appear to fuse, and that they separate 

 again without mixture, interpenetration, or combination, of the 

 granular or fluid constituents of the protoplasm. 



The character-unit assumption requires us to imagine some way 

 in which the particular granules could create or bring about the 

 existence or the accentuation of the particular character, whereas 

 the other interpretation, by lines of descent, does not needlessly 

 destroy the unity of the problem of heredity. It avoids the 

 necessity of elaborate and gratuitous hypotheses in a field which 

 science is scarcely prepared to enter. As in the adjoining 

 regions of instinct and memory, it is easy to ascribe the phe- 



