106 Prof. J. B. Farmer and Mr. J. E. S. Moore. [May 29, 



spireme contracts, and, as this phase progresses, the threads are pulled 

 into more and more parallel positions (Figs. 2 and 3) ; and, if polarisation 

 is strongly marked, as in most animals, the threads are seen to take the 

 form of loops, or large U-shaped figures. Although still showing longi- 

 tudinal fission, the sides of these loops approximate once more, and 

 eventually nearly all signs of the original split are obliterated. But, 

 even at a later stage, careful search often reveals the original opening 

 running along the sides of the loops. This method of formation 

 of the U-shaped chromatin bands, and the final approximation of their 

 limbs, has given rise to a complete misunderstanding of the ensuing 

 phases ; the approximated limbs of the loop having been regarded as 

 the longitudinally split halves of the spireme thread, and the great 

 increase in their thickness having been supposed to be due to a large 

 contraction in the length of the spireme. 



The full sequence in the production of the loops can only be made 

 out in very carefully prepared material, and as a result of a comparison 

 of sections with uncut cells. The chromatin, indeed, during these 

 stages seems to be in a condition very sensitive to unfavourable 

 manipulation, and we have reason to think that the quite unnecessary 

 length of embedding processes so frequently resorted to may be 

 greatly responsible for the failure to recognise the true sequence of 

 events. Every cytologist who has examined thick sections, or un- 

 injured cells, in which the early heterotype chromosomes are shown, 

 must have been struck with the appearance of forms composed of rods, 

 often twisted on themselves with their limbs open at one end, and 

 forming at the other a conspicuous loop, such loops being in reality 

 nothing more than the bent middle portions of the original U-shaped 

 spireme segments (Fig. 4). The bent portions of the loops, how- 

 ever, mark the place at which, during the metaphase, the heterotype 

 loop will break apart into the two chromosomes, of which it is com- 

 posed. It sometimes happens that, even at this stage, traces of the 

 longitudinal fission can be distinctly seen, and it is the persistence or 

 reappearance of this fission which has been taken for a second longi- 

 tudinal split. The number of the chromatic loops corresponds to the 

 reduced number of the chromosomes, and thus there is direct evidence 

 that each loop is really bivalent in the sense in which the term was 

 used by Hacker, but our view of the actual process of the formation 

 of the loops, and of the subsequent course of events, does not coincide 

 with the conclusions reached by this investigator. 



When the bivalent chromosomes, loops or rings, become arranged on 

 the spindle (Fig. 5), they split asunder in such a way that one limb of 

 the loop passes into each daughter nucleus, and consequently it will be 

 apparent that their division cannot be regarded as the completion of any 

 longitudinal fission, but as a transverse separation of the chromosomes 

 originally united together end to end in the bivalent loop ring or rod . 



