2o6 KINGSBURY. ■ [Vol. IL 



impossible to trace them. The longitudinal splitting may have 

 been complete, and the fusion of the apices been of secondary 

 occurrence. The two chromatin segments so joined in pairs at 

 their middle points show a tendency to diverge from each other, 

 and in most cases, therefore, each pair is formed of four chro- 

 matin threads radiating from the point of fusion and lying in. 

 two planes at right angles to each other, as indicated in Fig. 4, 

 There follow a shortening and thickening of the chromatin 

 threads by which each pair becomes converted into a + or 

 an X. 



These usually tend to lie near the periphery of the nucleus 

 under the nuclear membrane, as do the rings in the correspond- 

 ing stage in the spermatocyte of the first order. Finally, the 

 fusion is dissolved and there are formed from each -f- or X 

 two F's, daughter-chromosomes of the second division. These 

 are generally quite irregularly distributed 

 at first ; soon, however, they take up their 

 position in the center of the cell, to form 

 a somewhat loose equatorial plate in which 

 the daughter-chromosomes are generally 

 quite widely separated from each other. 

 Migration to the two poles to form the 

 nuclei of the spermatids occurs in DcsmognatJuis in the manner 

 already described by Flemming and Meves for Salaniajida. 



The early splitting and early and complete separation of the 

 chromosomes in the spermatocyte of the second order was rec- 

 ognized by Flemming and made a characteristic of the peculiar 

 form of mitosis which he called "homotypic," basing the type 

 on the conditions found in the second division in Salaniandra. 

 Although it seems to me probable that the X's and their 

 dissolution into F-shaped chromosomes is due simply to an 

 incomplete precocious longitudinal splitting of the chromo- 

 somes, which upon its final completion gives an equation 

 division of the chromatin mass represented by the formula 



a-b , a b . . , , .,11 1 r 1 



— 7> thus, — ,> it IS nevertheless possible that the final 

 a-b a b 



separation is represented by y— r» thus, — -, ; and thererore it 



