2S6 REDUCTION OF THE CHROMOSOMES 



typically represented by Ascaris and the arthropods, each of these 

 masses divides into four to form a tetrad, thus preparing at once for 

 two rapidly succeeding divisions, which are not separated by a recon- 

 struction of the daughter-nuclei during an intervening resting period. 

 In the other, examples of which are given by the flowering plants and 

 the spermatogenesis of the Amphibia, no true tetrads are formed, the 

 primary chromatin-masses dividing separately for each of the matura- 

 tion-divisions, which are separated by a period in which the nuclei 

 regress toward the resting state, though often not completely return- 

 ing to the reticular condition. These two types differ, however, only 

 in degree, and with few exceptions they agree in the fact that during 

 the prophases of the first division the chromatin-bodies assume the 

 form of rings, the mitosis thus being of the heterotypical form, and 

 each ring having the prospective value of four chromosomes. 



Thus far the phenomena present no difficulty, and they give us a 

 clear view of the process by which the numerical reduction of the 

 chromosomes is effected. The confusion of the subject arises, on the 

 one hand, from its complication with theories regarding the individu- 

 ality of the chromosomes and the functions of chromatin in inheri- 

 tance, on the other through conflicting results of observation on the 

 mode of tetrad-formation and the character of the maturation-divisions. 

 Regarding the latter question nearly all observers are now agreed that 

 one of these divisions, usually the first, is a longitudinal or equation- 

 division, essentially like that occurring in ordinary mitosis. The main 

 question turns upon the other division, which has been shown in some 

 cases to be transverse and not longitudinal, and thus separates what 

 were originally different regions of the spireme-thread or nuclear 

 substance. The evidence in favour of such a division seems at present 

 well-nigh demonstrative in the case of insects and copepods, and 

 hardly less convincing in the turbellarians, annelids, and mollusks. 

 On the other hand, both divisions are regarded as longitudinal by most 

 of those who have investigated the phenomena in Ascaris and in the 

 vertebrates, and by some of the most competent investigators of the 

 flowering plants. 



The evidence as it stands is so evenly balanced that the subject is 

 hardly yet ripe for discussion. The principle for which Weismann 

 contended in his theory of reducing div^ision has received strong 

 support in fact; yet should it be finally estabUshed that numerical 

 reduction may be effected either with or without transverse division, 

 as now seems probable, not only will that theory have to be aban- 

 doned or wholly remodelled, but we shall have to seek a new basis 

 for the interpretation of mitosis in general. Weismann's theory is 

 no doubt of a highly artificial character ; but this should not close our 

 eyes to the great interest of the problem that it attempted to solve. 



