48 THOS. a. MONÏGOMERt jr., 



fibres, and since it could not be positively determined that these fibres 

 did not lie within the nucleus and so represent mantle fibres, I am 

 inclined to conclude that when the centrosomes reach opposite sides 

 of the nucleus , they are no longer connected by fibres , and hence 

 that the original central spindle has disappeared. I studied numerous 

 cells in this stage to determine the possibility of the central spindle 

 lying in a groove or furrow of the surface of the nucleus. In this 

 case one would find evidences of such a furrow on cross sections of 

 the nucleus, i. e. sections transverse to a line connecting the two 

 centrosomes; but no traces of a furrow were to be seen. So we may 

 conclude that the primary central spindle disappears; the more so, 

 indeed, since our present knowledge of the central spindle as an 

 extensible structure , would preclude the possibility of its curving 

 around the surface of the nucleus, and subsequent shortening and 

 straightening in the following monaster stage. The fibres which 

 connect the centrosomes after the nuclear membrane have disappeared, 

 and which shall be described below, would thus appear to have no 

 connection with the primary spindle. Other cases have been described 

 of the disappearance of a primary central spindle , as e. g. in Myzo- 

 stoma (Wheeler, '97), so that Pentatoma is not isolated in this 

 regard. In the spermatogonic mitoses the primary central spindle 

 persisted; and this fact is explained by the earlier disappearance of 

 the nuclear membrane in that generation of cells — the disappearance 

 of the nuclear membrane before the centrosomes had reached opposite 

 sides of the nucleus. 



When the centrosomes have passed to opposite poles of the nucleus, 

 we find each of them doubled , due to a division of each centrosome 

 about this stage (Figs. 150-153, 156, 157, 161—166, 168, 172, 174, 

 175). It is the rule that each is doubled from this stage up to the 

 metaphase ; in a small number of cases they appear single (Figs. 154, 

 155, 158 — 160, 167, 173), but most cases of this kind are probably 

 to be explained by the fact that both centrosomes of a pair do not 

 always lie in the plane of the section. Sometimes a single centrosome 

 at one pole of the nucleus is equal in volume to the two at the 

 opposite pole, and then we can be sure that the former not yet 

 divided, and hence that the division of the centrosomes does not 

 always occur simultaneously. In most cases the two centrosomes of 

 a pair may be seen to be connected by a true centrodesmosis , so 

 that the whole has a dumbbell shape; but often the centrodesmosis 

 is not clearly demonstrable, though it is probable that it is as a rule 



