SUTTON 



it tend in diverse directions, must in 

 time, guided by natural selection, re- 

 sult in an appreciable difference in a 

 definite direction bet\veen a chromo- 

 some and its direct descendant and 

 hence between the characters asso- 

 ciated with them. In this we have a 

 sus^S^estion of a possible cause of in- 

 dividual variation in homologous chro- 

 mosomes which we have already seen 

 reason to suspect (pp. 221 and 226). 



Finally, we may briefly consider 

 certain observations which seem at first 

 sight to preclude the general applica- 

 bility of the conclusions here brought 

 out. If it be admitted that the phe- 

 nomenon of character-reduction dis- 

 covered by Mendel is the expression 

 of chromosome-reduction, it follows 

 that forms which vary according to 

 Mendel's law must present a reducing 

 division. But the vertebrates and 

 flowering plants— the very forms from 

 which most of the Mendelian results 

 have been obtained— have been re- 

 peatedly described as not exhibiting a 

 reducing division. Here, therefore, is 

 a discrepancy of which I venture to 

 indicate a possible explanation in the 

 suggestion first made by Fick -- and 

 more recently by Montgomery.-^ This 

 is to the eff^ect that in synapsis as it 

 occurs in vertebrates and other forms 

 possessing loop-shaped chromosomes, 

 the union is side by side instead of 

 end-to-end as in Arthropods. In ver- 

 tebrates, two parallel longitudinal 

 splits, the forerunners of the two fol- 

 lowing divisions, appear in the chro- 

 mosomes of the primary spermatocyte 

 prophases. Both being longitudinal, 

 they have been described as equation 

 divisions, but if it shall be found pos- 

 sible to trace one to the original line 

 of union of the two spermatogonial 

 chromosomes side by side in synapsis, 



" Fick, R., "Mittheilung ueber Eireifung 

 bei Amphibien," Supp. Afiat. Anz., XVI. 

 23 Montgomery, T. H., Jr., loc. cit. 



39 



that division must be conceived as a 

 true reduction. A number of observa- 

 tions supporting this view will be 

 brought forward in my forthcoming 

 work on Brachystola. 



A^ain, if the normal course of in- 

 heritance depends upon the accurate 

 chromatin-division accomplished by 

 mitosis, it would appear that the inter- 

 jection, into any part of the germ 

 cycle, of the gross processes of amitosis 

 could result only in a radical deviation 

 from that normal course. Such an oc- 

 currence has actually been described 

 by Meves, McGregor and others in the 

 primary spermatogonia of amphibians. 

 In these cases, however, it appears that 

 fission of the cell-body does not neces- 

 sarily follow^ amitotic division of the 

 nucleus. I would suggest, therefore, 

 the possibility that the process may be 

 of no significance in inheritance, since 

 by the disappearance of the nuclear 

 membranes in preparation for the first 

 mitotic division, the original condition 

 is restored, and the chromosomes may 

 enter the equatorial plate as if no 

 amitotic process had intervened.^^ 



There is one observation in connec- 

 tion with the accessory chromosome 

 which deserves mention in any treat- 

 ment of the chromosomes as agents in 

 heredity. This element always divides 

 longitudinally and hence probably 

 equationally. It fails to divide in the 

 first maturation mitosis, in which the 

 ordinary chromosomes are divided 

 equationally, but passes entire to one of 

 the resulting cells. In the second mat- 

 uration division, by which the reduc- 



-^ It is of interest in connection with this 

 question that there occurs regularly in each 

 of the spermatogonial generations in Brachy- 

 stola a condition of the nucleus which sug- 

 gests amitosis but which in reality is nothing 

 more than the enclosure of the different 

 chromosomes in partially separated vesicles. 

 Cf. Sutton, W. S., "The Spermatogonial 

 Divisions in Brachystola Magna," Kans. Univ. 

 Quart., IX., 2. 



