48 bulletin: museum of comparative zoology. 



During the migration of the centrosomes the centrosphere has again 

 changed its staining reaction and now shows a stronger affinity for iron- 

 haematoxylin. As in the larger type, however, the stain is removed 

 when the decolorizing is continued for some time, and thus the centro- 

 some itself is made visible. During the late prophase each centrosome 

 is elongated and constricted at the centre (Fig. 61). It is evidently 

 preparing to divide for the second spermatocyte mitosis. 



As the nuclear membrane on the side opposite the centrosomes dis- 

 appears, the mantle fibres quickly form and are seen connecting the 

 centrosomes to the chromosomes. These mantle fibres seem quite plainly 

 to be derived from the linin of the nucleus, although the nuclear mem- 

 brane itself may also take part in their formation. As the membrane 

 disintegrates, the linin, formerly arranged in granular rows, becomes 

 fibrillar, and may be seen connecting the different chromosomes and 

 uniting with the fibres centring in the centrosome (Fig. 63). "Whether 

 or not they form the entire set of mantle fibres, I cannot say, but I 

 believe there can be no doubt whatever that they contribute to it. 

 These conditions are plainly shown in Figures 62, 63. On account of the 

 incomplete migration of the centrosomes, the spindle as first constructed 

 is asymmetrical (Figs. 62, 63). The chromosomes do not at first lie 

 directly between the centrosomes, but are off at one side. Indeed, 

 immediately before the formation of the spindle, these elements seem to 

 be forced toward the side of the vesicle farthest from the centrosomes. 

 However, the mantle fibres soon contract and draw the chromosomes 

 into the space between the centrosomes, and thus a symmetrical, though 

 very short, spindle is produced (Fig. 64). 



The central spindle is very well developed. The astral rays at the 

 sides of the spindle are numerous, stain very distinctly, and can with ease 

 be traced without break through the cytoplasm nearly to the cell mem- 

 brane. Here they divide into several divergent branches, and some of 

 the branches, at least, come in contact with the cell membrane. That 

 they are attached to this structure, there can be no doubt whatever. 

 Such certainly is the case during the metaphase in a large number of 

 cells examined, and such is certainly true of many of these fibres in 

 the anaphase and telophase. Beyond each pole of the spindle, between 

 the centrosome and the cell wall, there is a cone-shaped area through 

 which very few astral rays extend. This region is occupied by a coarse- 

 mieshed network of cytoplasm, in the meshes of which there are no stain- 

 able deposits. The astral rays in this region are not as numerous as 

 •elsewhere in the astral system and furthermore are very short. 



