Studies on Synapsis. 



73 



loops extend into contact with the nuclear membrane (fig. 32), the longi- 

 tudinal cleavage of the diplotene stage becomes evident. There follows the 

 dissolution of the bouquet ; and finally the abbreviation of the diplotene 

 loops into segments of which the longitudinal halves become drawn apart 

 except at their extremities (fig. 33). These latter condense to form tetrad- 

 like bodies, which soon become increasingly granular in appearance, so that 

 individually they are no longer recognisable, the surface of the nucleus in 

 good preparations being apparently evenly covered with chromatin granules. 

 Thus the diffuse stage supervenes in Libellula as in Periplaneta ; and 

 throughout the entire series of events the plasmosome remains altogether 

 independent of the chromatinic organisation of the nucleus (fig. 35). Thus 

 it is necessary to seek for some alternative interpretation of the double 

 nucleolus of Libellula depressa to that offered by McGill for«Anax and 

 Plathemis ; and when the subsequent history of the oocyte nucleus in 

 Libellula has been stated, there will be found little cause to doubt that 

 the same procedure is in reality to be found in the other two genera. 



Before passing on to an examination of the genesis and significance of the 

 double nucleolus, it is interesting to compare the chromosome cycle of 

 the Odonata with that of other Hexapoda. The early differentiation of 

 the tetrads and the retarded segregation of the X-element in the male germ 

 cells is similar to what has been described by Wilson for several genera 

 of the Hemiptera. On the other hand, the definite polarisation in the 

 synaptic stage conforms to the Orthopteran type. In Libellula, as in 

 Periplaneta and Gryllus, the initial meiotic stages are of longer duration 

 in the female germ cells than in the spermatocytes; also, owing to the 

 larger size of the nuclei, the former are more favourable for investigation ; 

 it is therefore important, in the interpretation of synapsis, that work on 

 the gametogenesis of both sexes should go hand in hand. It is true that, 

 as the parasynaptic view has gained in prestige, some authors, e.g., King (13), 

 have pressed the compromise suggested by Gates (12) to the point of 

 advocating both metasyndetic and parasyndetic reduction not only in closely 

 allied species, but in alternate sexes of the same species. This position 

 may satisfy those whose attention has been focussed exclusively on the 

 frequently very intricate task of elucidating the complicated metamorphosis 

 of the spindle bivalents, the conjugant halves of which in the final stage 

 are often united only by one extremity. But, as Agar has rightly insisted, 

 a survey of the entire sequence of changes reveals such striking similarity, 

 more especially in the universally characteristic diplotene stage, that a 

 satisfactory common ground is not easy to find for cytologists of the 

 conflicting schools of opinion by any such facile means. Furthermore, as 



