A Preliminary Account of the Spermatogenesis of Sphenodon. 351 



showed twisting of the longitudinal halves as described originally 

 by Jannsens ; but whether they become untwisted subsequently, 

 or whether an actual chiasma is effected in Sphenodon, cannot be 

 affirmed. In the majority of these the interspace m the middle of 

 the rods first widens so that a more or less quadrangular element 

 is formed ; this assumes finally the characteristic annular form. 

 In other cases the separation of the two halves is asymmetrical — 

 i.e. the interspace widens first nearer one extremity than the other, 

 so that there is a conspicuous lug to the ring (fig. 10). There appear 

 to be three of these, corresponding it will be noticed to the three 

 pairs of chromosomes in the spermatogonial group which are 

 asymmetrical and display subterniinal fibre attachment ; one of 

 them is the largest element in the diakinesis stage, and it will be 



remembered that the ^ shaped pair are the largest in the sperma- 

 togonial group. It must be added however in qualification of this 

 correlation that asymmetrical heterotype chromosomes occur both 

 in Periplaneta and in Lihellula, genera in which all the premeiotic 

 chromosomes are approximately alike. 



The metamorphosis of the heterotype chromosomes of Hatteria, 

 as in Libelhda, takes place before the dissolution of the nuclear 

 membrane ; consequently it is not easy to refer the axis of division 

 of the symmetrical bivalents which show median fibre attachment 

 to the plane of fission in the diplotene stage. The conspicuous 

 lugs of the asymmetrical ones which display subterniinal attach- 

 ment does, however, make it possible to identify the two in 

 keeping with the parasynaptic view. This is illustrated in text- 

 figs. 10, 13, 14, 15. Text-figs. 11 and 12 show two series, the 



