40 KARYOKINESIS. 



their affinity for nuclear stains- increases until in the resting stage of certain cleavage 

 cells, these centrosomes look like small nuclei filled with amass of minute chromatin 

 granules, figs. 61, 69, 76, text fig. IV, / In other cleavage cells these centrosomes 

 with their contained granules remain much less conspicuous. Of all the early 

 cleavage they are most plainly visible in the macromex'es just before the formation 

 of the first and second quartettes, figs. 69, 74, 75. The cause of this diiference in the 

 appearance of the centrosomes in different cells depends largel}- upon their size and 

 affinity for stains. The size of the centrosome is always proportional to that of the 

 cell in which it lies ; its affinity for stains, during the resting period, increases as it 

 approaches the free surface of the cell, so that although it may stain faintly when a 

 short distance fi^om the surface (figs. 61, 62, 68) it stains deeply when in contact 

 with it (figs. 69, 74). Upon these two factors then depends the relative conspicu- 

 ousness of the centrosome during the resting period. 



In all the cleavages which I have studied, with the exception of the first, the 

 new centrosomes, and probably also the central spindles, arise within the mother cen- 

 trosome, as in the case of the second maturation spindle. The origin of centrosomes 

 for the second cleavage is shown in figs. 62 and 63, though the origin of the central 

 spindle could not be clearly made out in this case. The origin of centrosomes and 

 spindles for the third cleavage is shown in fig. 70, while those for later cleavage? 

 are shown in figs. 74, 75 and 76. In all these cases the centrosomes appear as 

 slightly enlarged granules within the old centrosome. These granules stand at some 

 distance from each other, and in no case in the cleavage have I seen the division of 

 a single granule to form these two ; they are, however, connected by the reticulum 

 of threads and granules which fills the mother centrosome, and when the time arrives 

 for the formation of a new mitotic fiaure the mother centrosome elons;ates, becomina; 

 slightly elliptical in outline, the daughter centrosomes, as two enlarged granules, lie 

 at the extremities of this ellipse, and the reticulum which fills the mother centrosome 

 is drawn out into an irregular spindle shaped body composed of threads and granules, 

 figs. 70, 74, 75, text fig. IV, ^, /z, i. This elongation continues and the spindle 

 shaped body becomes the central spindle (fig. 76), which in this case consists, not 

 of straight fibres running from pole to pole, but of irregular and anastomosing fibres 

 with granules at their nodes. The daughter centrosomes soon become surrounded 

 by a little area free from granules, which is due to a halo of radiating fibres, so 

 fine that few of them can be seen at this stage. This is the first appearance of the 

 sphere ("couche corticale") and it also arises, at least in certain cleavages, within 

 the mother centrosome, figs. 70 and 76, the membrane of which may still persist at 

 this stage. 



At a slightljr later stage these radiating fibres become very evident, and with 

 the formation of the cones or half spindles, as described at the beginning of this 

 account of the centrosome, we have the completion of the cycle of changes under- 

 gone by the centrosome from one prophase to the next. In a word, the most im- 

 portant features of this cycle are (1) the great increase in size of the centrosome and 

 its transformation into a sphere filled with a reticulum of fibres and granules, and 



