Studies on Synapsis. 



67 



of the diplotene filaments is clear (fig. 17). But since some of the tetrads at 

 the time when the nuclear membrane disappears, already exhibit the form of 

 practically symmetrical crosses (fig. 18), there is difficulty in referring the 

 plane of division to the axis of the filament from which they are derived. If 

 the long axis of the tetrad when arranged on the achromatic spindle corre- 

 sponds to the line of cleavage in the diplotene stage, the first division is 

 reductive in the modern sense, and must separate univalents united end to 

 end : otherwise it is reductive only in the sense in which Weismanu inter- 

 preted meiotis, dividing each univalent transversely. In either case such a 

 •disposition of the univalents in the metaphase of the heterotype mitosis 

 would be incompatible with the theory of parasynapsis as now understood. 

 This is the arrangement described by Lefevre and McGill for Anax ; but it is 

 clear from their figures that the sequence of the stages on which they base 

 their contention is an inversion of the correct order. Smith adopts the 

 alternative view in the case of Sympetrum and L. basalis ; that is, the tetrad 

 is so placed in metaphase that the transverse division of the tetrad involves 

 the separation of the originally longitudinal halves of the diplotene segments. 

 Though it is impossible to give objective demonstration of the correctness of 

 this description for all the bivalents of Lihellula depressa, at least two of them 

 (cf. figs. 19 and 20) are symmetrical about one axis only, and show subterminal 

 instead of median spindle fibre attachment. These are placed in the 

 metaphase with their single axis of symmetry in the equatorial plane, and 

 there is therefore n ) doubt whatever in regard to these that the heterotype 

 •division separates longitudinal halves of the diplotene bivalents. The 

 similarity of these tetrads to the remainder in all other respects permits the 

 inference that the latter behave in the same way. 



In metaphase each bivalent is eventually draws out, as is Periplaneta, into 

 an elliptical figure ; and as each division completes itself the two arms of the 

 dyads are drawn into closer contact. The homotype mitosis follows quickly 

 on the telophase, and the chromosomes of the second division are actually 

 not more than half the size of those in the first. Consequently it is not an 

 easy task to follow out the prophase of the former, owing to the very small 

 dimensions of the nuclei. A resting phase is not described by Lefevre and 

 McGill in Anax, nor by Smith. But as far as Z. depressa is concerned, there 

 is such a stage, though of very short duration. The telophase chromosomes 

 of the lieterotype mitosis seem to spin out into loops as in the spermatogonia. 

 The prophase chromosomes show cleavage lengthwise. Thus the homotype 

 mitosis is equational, and cannot be referred to the transverse constriction of 

 the dyad. This is exactly as described by Morse and the author in 

 Periplaneta. 



