Cardiff: Synapsis and reduction 279 



Rosenberg ('03, '04a, '04^) in Drosera obtains especially in- 

 teresting results from a hybrid of Drosera longifolia with Z>. rotttn- 

 difolia. The gametophyte of D. longifolia has twenty small chro- 

 mosomes, while the gametophyte of D. rotundifolia has ten large 

 ones. Therefore in the hybrid sporophyte there are thirty, while 

 in the gametophyte there are not fifteen but twenty chromosomes. 

 Rosenberg finds that, in the early prophases of the first mother- 

 cell mitosis of the hybrid, there are twenty chromosomes, ten of 

 which are bivalent, while the other ten are univalent. These ten 

 bivalents are each composed of a larger and a smaller part Rosen- 

 berg believes that the ten large chromosomes from D. rotundifolia 

 fused with ten of the smaller ones of D. longifolia, thus leaving 

 the other ten small chromosomes univalent. The first mitosis 

 separates by a longitudinal fission whole chromosomes. This is 

 true for others than the hybrid. Thus there is a true reduction 

 in the same sense as in the results of Berghs ('04a, '04b, } o$a) and 

 of Allen i^O^a). The second division is also longitudinal. 



Miyake ('05), in a number of monocotyledons, states that the 

 threads are not fully formed previous to synapsis but that masses 

 of chromatin — not chromosomes — fuse and later organize paral- 

 lel threads which separate with the first mitosis. He finds the 

 second mitosis homotypic. 



Overton ('05), in a number of dicotyledons, gets essentially 

 identical results with those of Allen, though he figures masses of 

 chromatin — " protochromosomes " — which he considers the 

 equivalent of the sporophytic chromosomes. These masses pair 

 but do not fuse — during synapsis and separate longitudinally 

 in the first division. Allen ('05^) in a later paper confirms his 

 previous results ('05^). Miss Stevens ('05), in Aphis, finds that 

 reduction in the spermatocytes is effected by a longitudinal pair- 

 ing of the chromosomes. 



Schreiner and Schreiner ('06) in a recent paper on Tomopteris, 

 an annelid, report results essentially similar to those found in the 

 present work. In a number of excellent figures the slender chro- 

 matin threads are shown to arrange themselves parallel and, com- 

 mencing at one pole of the nucleus, conjugate longitudinally. 

 The threads then become shorter and thicker as the first division 

 is approached. After the parallel threads have conjugated the 



