■992 



SCIENCE. 



[N.S. Vol. XVI No. 416 



nary somatic mitoses in that it separates 

 whole chromosomes by a transverse division 

 ('reducing division' of Weismann). Wholly 

 independently of Mendel's conclusions a con- 

 siderable number of cytologists (vom Rath, 

 Riickert, Hacker) early reached the conclu- 

 sion that the chromatin-masses from which 

 arise the ' Vierergruppen ' (tetrad-chromo- 

 somes, or their equivalents) represent double 

 or ' bivalent ' chromosomes, each of which was 

 conceived to arise by the union (synapsis), 

 end to end, of two single chromosomes. An 

 actual conjugation of chromosomes in synapsis 

 was inferred by Riickert in some cases (e. (7., 

 in Pristiurus) , and more recently described in 

 £L far more detailed way in Peripatus and cer- 

 tain insects by Montgomery (1901), who 

 reached the remarkable conclusion that ' in 

 the synapsis stage is effected a union of pater- 

 nal with maternal chromosomes, so that each 

 bivalent chromosome would consist of one uni- 

 valent paternal chromosome and one univalent 

 maternal chromosome.' The ensuing trans- 

 verse or reducing division, therefore, leads to 

 the separation of paternal and maternal ele- 

 ments and their ultimate isolation in separate 

 ^erm-cells. This conclusion rested upon evi- 

 dence too incomplete to warrant its acceptance 

 without much more extended investigation — 

 ■it was, indeed, more in the nature of a sur- 

 mise than a well-grounded conclusion. Dur- 

 ing the past year Mr. W. S. Sutton, working 

 in my laboratory, has obtained more definite 

 .evidence in favor of this result, which led 

 him several months ago to the conclusion that 

 it probably gives the explanation of the Men- 

 delian principle. In the great ' lubber grass- 

 hopper ' Brachystola the chromosomes of the 

 spermatogonia were found to be grouped in 

 ■eleven pairs of different sizes, which reap- 

 peared in essentially the same relation through 

 at least eight successive generations of these 

 <!ells. In synapsis the graded pairs are con- 

 verted into similarly graded bivalent chromo- 

 somes that appear to arise by a conjugation, 

 or union at one end, of the two members of 

 «ach of the earlier pairs. Cogent reason is 

 ■given by Sutton for the conclusion that the 

 chromosome-pairs consist each of a paternal 

 and a maternal member. It is known that in 



fertilization chromosomes are contributed in 

 equal numbers by the two gametes (' Van 

 Beneden's Law '). Boveri's recent remark- 

 able experiments on sea-urchins have proved 

 that a definite combination of chromosomes is 

 necessary to complete development, and 

 strongly suggests, if they do not prove, that 

 the individual chromosomes stand in definite 

 relation to transmissible characters taken 

 singly or in groups. Every nucleus, how- 

 ever, contains two such combinations ; for the 

 facts of parthenogenesis and merogony prove 

 that either the paternal or the maternal group 

 alone may suffice for complete development. 

 It is a natural conclusion from these facts that 

 the constant morphological differences of the 

 chromosomes observed in the grasshopper are 

 correlated with constant physiological dif- 

 ferences. If such be the case it appears highly 

 probable, though the argument can not here 

 be presented in all its weight, that those of 

 corresponding size, associated in pairs, are the 

 paternal and maternal homologues {sit venia 

 verho) ! Sutton has pointed out that if this 

 be indeed the case, the union of these homo- 

 logues in synapsis, and their subsequent sepa- 

 ration, which this preliminary union involves, 

 in the reducing (second maturation) division, 

 leads to the members of each pair being iso- 

 lated in separate germ-cells ; and this gives a 

 physical basis for the association of dominant 

 and recessive characters in the cross-bred, and 

 their subsequent isolation in separate germ- 

 cells, exactly such as the Mendelian principle 

 requires. 



A similar conclusion was subsequently, but 

 independently, reached by Mr. W. A. Cannon, 

 of the Department of Botany, though by a 

 different and less direct path of approach. A 

 study of hybrid cotton-plants, which are fer- 

 tile, showed the maturation-divisions to be en- 

 tirely normal, in contradistinction to the sterile 

 hybrids of Syringa, where Juel has shown that 

 the maturation-divisions are abnormal in 

 character. It thus appeared that a sifting 

 apart of paternal and maternal elements, such 

 as Mendel's law demonstrates to occur, can- 

 not be explained on the hypothesis of irregu- 

 larities in the maturation-divisions (as had 

 been suggested by Guyer's earlier work 



