56 THOS. H. MONTGOMERY jr., 



some; or else a break occurs first in the axis of the chromosome, 

 while the chromatin is still continuous at the periphery (Figs. 190 — 

 194). In the latter case, after the chromosomes have become completely- 

 divided, we see a concavity on that surface of each chromosome which 

 lies nearest the other daughter chromosome (Figs. 194, 125). These 

 two modes of division occur with equal frequency, and show that the 

 commencement of the division of the chromosomes is automatic. This 

 active division on the part of the chromosomes must be due either 

 to a migration of the microsomes to opposite poles, or else to some 

 tension or concentration of their linin matrix. 



The several chromosomes do not divide simultaneously, just as 

 they did not all reach the equator at the same time: perhaps those 

 which cume first into position are the first to divide. 



As the daughter chromosomes separate from one another, connecting 

 each two is a linin thread, a connective fibre, derived entirely 

 from the linin within the chromosomes (Figs. 192—202). As in the 

 spermatogonic mitoses, each of these threads appears to be a hollow 

 cylinder of linin, and thus is probably a derivative of a linin sheath 

 of the chromosome. They are the thickest fibres of the chromatic 

 spindle. At the following dyaster stage they are frequently found 

 crossing one another (Fig. 198), due to the chromosomes having 

 shifted theii relative positions during the metakinesis. They persist 

 intact until the separation of the daughter cells (2nd spermatocytes), 

 when they break at the cell membrane (Figs. 201, 202, 206, Cn.F); 

 but I have been unable to determine what becomes of them sub- 

 sequently. 



Up to the monaster stage the centrosomes were paired, with the 

 two of each pair usually lying in close contact with one another 

 (Figs. 151—169, 172—175); about the time of metakinesis the two 

 of each pair begin to separate from one another, in a line which is 

 usually not quite at right angles to the axis of the spindle (Figs. 201, 

 202). The planes of position of the centrosomes do not always coincide — 

 indeed they usually do not, which is necessary to bear in mind in 

 considering those figures where a pair of centrosomes is seen at one 

 pole, and only one centrosome at the other pole. Further, the centro- 

 somes of the two pairs do not separate simultaneously nor with equal 

 rapidity. 



When the two centrosomes of a pair first commence to separate 

 they are connected by a primary centrodesmosis , i. e. a bridge of 

 centrosomal substance {c.ds., Figs. 191, 192). This soon disappears 



