66 



Mr. L. Hogben. 



Libellula depressa was it possible to find any such stages. As already 

 mentioned the telophase of the spermatogonia! mitoses is characterised by an 

 orientation of the filaments exactly like that which is found in the bouquet 

 stage in Periplaneta ; and in fig. 13 of Schaffer's paper (8) precisely the same 

 occurrence is clearly illustrated. All the available evidence points to the 

 conclusion that the fine polarised leptotene loops of Libellula depressa display 

 frc^ the inception of the meiotic phase the arrangement characteristic of the 

 chromatin filaments in the premeiotic telophase {cf. figs. 4 and 11). As far as 

 could be ascertained the leptotene filaments are from the first present as 

 loops polarised at both extremities (fig. 11). It is very ditticult to count the 

 loops with accuracy ; but their number is clearly more numerous than in the 

 pachytene stage. In the contraction figure, they are drawn away from the 

 nuclear membrane and become considerably abbreviated (fig. 13). Owing to 

 their crowded disposition, it is not possible to observe with certainty actual 

 pairing ; but it is certain that throughout this stage and in that which follows, 

 when it is possible to count with accuracy the haploid number of loops, they 

 retain the original condition of polarisation at both extremities (figs. 13, 14). 



Now if the polarisation of the leptotene loops is consequent upon the 

 character of the last premeiotic telophase, there is no possibility in the case 

 of Libellula that numerical reduction of the loops could be effected by means 

 of an end-to-end union without an entire reorganisation of the chromatin 

 organisation of the nucleus. In the absence of evidence for the intercalation 

 of such a process, those who advocate telosynapsis for Hexapoda fall back on 

 the conception of a continuous spireme ; envisage reduction by the segmenta- 

 tion of the spireme into the haploid number of lengths : and declare them- 

 selves unable to find a previous (leptotene) stage in which the number of 

 loops is more numerous. This contention is based on a fundamental mis- 

 conception as to the nature of the telophase : when this is rightly appreciated, 

 it follows that conjugation can only take place at one point in the sequence 

 of meiotic phenomena — the bouquet contraction, and can only be effected by 

 one means — the lateral approximation of the loops or filaments. 



With respect to the diplotene stage (figs. 14, 15) in which the reduced 

 filaments become extended and display longitudinal cleavage, it is only 

 necessary to state that the nucleus as a whole undergoes considerable increase 

 in bulk during this period (figs. 15, 16). Eventually the haploid number of 

 longitudinally split filaments are detached (fig. 16) ; and their transformation 

 into the definitive tetrads coincides with a progressive shrinking in size of 

 the nucleus (figs. 17, 18). 



The fate and constitution of the tetrads in the Odonata has provoked 

 controversy. That they are formed by the opening out of the line of cleavage 



