278 Cardiff: Synapsis and reduction 



after its formation. Just previous to the union of the two parallel 

 threads their substance is differentiated into linin and chromatin, 

 the latter aggregated into small granules, chromomeres. With 

 the fusion of the threads the chromomeres fuse in pairs. The 

 chromatin is in synapsis for several days. The threads emerge 

 from synapsis and become distributed throughout the nucleus 

 forming twelve loops, not unlike those figured by Farmer and 

 Moore ('05), which segment transversely at or near the point 

 where they are in contact with the nuclear wall to form the 

 reduced number of chromosomes. Previous to this transverse 

 segmentation there is a longitudinal fission of the thread, the 

 chromomeres, which are still apparent, first dividing. Allen 

 believes that this longitudinal fission represents a separation of 

 the paired threads which fused in the presynaptic stages. He 

 finds a second contraction stage when the chromatin threads are 

 drawn away from the nuclear wall. This contraction, however, 

 is not nearly so marked as in the case of synapsis. The first 

 mitosis is longitudinal and probably separates the two threads 

 which fused in synapsis, thus giving a true reduction. The 

 heterotypic chromosomes are not, as a general thing, V-shaped, 

 but rod-shaped as they pass to the poles, though often remaining 

 attached at their equatorial end, thus forming a V-shaped body. 

 The second division is also longitudinal and divides the daughter- 

 chromosomes of the first mitosis. 



Winiwarter ('oo), suggested a view similar to Allen's as one 

 of three ways in which reduction might be accomplished, and is 

 strongly of the opinion that it is the most plausible explanation of 

 reduction, as it explains the synaptic knot, which he believes to be a 

 constant and important morphological character of the mother-cell. 



Almost simultaneously with Allen's work appeared that of 

 Berghs ('04^, '04^) in Convallaria and Allium, and of Schreiner and 

 Schreiner ('04). They get essentially the same results as Allen, 

 though Berghs believes there is not a complete fusion of the parallel 

 threads in synapsis. Gregoire ('04), in whose laboratory Berghs 

 worked, vouches for the accuracy of the latter' s observations and 

 strongly opposes the idea of an end to end fusion or a reduction 

 by a transverse division. Berghs ('05^) in a later work on Dro- 

 sera y Narthechtm, and HelUboms confirms his former results. 



