BLACKMAN : THE SPERMATOGENESIS OF SCOLOPENDRA. 37 



staining black arranged in a circle about tbe periphery of the bundle 

 (Fig. 38). Soon, however, they fuse together and form a ring (Fig. 39) 

 similar to those found iu amphibian material. The spindle remnants of 

 the first spermatocyte in no case persist very long, but gradually dis- 

 integrate and eventually disappear entirely. Very often the destruction 

 of this structure is hastened by the rotation of the cells upon each other 

 which seems always to occur either during the late telophase of the first 

 spermatocyte or during the early stages of the second spermatocyte. 



During the late telophase the astral systems undergo important 

 changes. Up to the time when the cell membrane is entirely con- 

 stricted, they decrease in extent but little, if indeed at all. But as soon 

 as this phenomenon is completed, they degenerate rapidly (Fig. 39). 

 However, they do not wholly disappear. Some of the archoplasm 

 persists in the form of fibres radiating from the centrosphere, while the 

 rest is deposited between the meshes of the cytoplasm. It, however, 

 does not long remain inert, for the early prophase of the second sperma- 

 tocyte is of very short duration, the second spermatocyte division fol- 

 lowing very closely upon that of the first. 



Centrosome and Centrosphere of First Spermatocyte. 



I have said that during the late prophase of the first spermatocyte 

 the centrosomes, while still moving apart along the nuclear membrane, 

 are evidently elongated in a direction parallel to this structure. This 

 elongation is often very pronounced and may even result in the formation 

 of dumbbell forms such as are shown in Figures 21, 22, 47. The sub- 

 stance of the centrosome becomes aggregated at the two ends of the 

 mass into two more or less equal lobes. This is evidently the first sign of 

 the division of the centrosome, which usually is completed in the follow- 

 ing metaphase or anaphase. It should be particularly noted at this 

 point that the body in question is the smallest analyzable portion of 

 the structure at the centre of the asters in the prophase, and is the part 

 toward which the radiations extend and iu which they end, i. e. it is the 

 centrosome as defined by Boveri. 



At this stage (Plate 4, Fig. 47) there is no well-defined structure en- 

 veloping the centrosome, but in a small space around this minute body the 

 cytoplasmic reticulum is absent, thus leaving a clear zone of hyaloplasm 

 (or latent archoplasm) which is traversed by the astral rays only. Later, 

 in the prophase and in the metaphase, however, a distinct centrosphere 

 is present (Figs. 48, 49, 50). All the observations upon the genesis of 

 this structure would seem to justify the conclusion that the centrosphere 



