298 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. 



with which these successive stages can be followed, together with the 

 evidence afforded by the staining reactions of the nucleus, seem to me 

 to show beyond reasonable doubt that the nucleolar shell does actually 

 contribute to the formation of the chromatin reticulum, and thus, indi- 

 rectly, to the chromosomes. The evidence indicates that the chromatin 

 substance which forms the peripheral portion of the nucleolus does not 

 consist of discrete structures, but is rather to be regarded as a granular 

 mass of viscid consistency, which breaks up into irregular portions in the 

 prophase. 



As the chromatin shell of the nucleolus disintegrates, the karyosomes 

 become correspondingly larger and at the same time denser, and more 

 sharply outlined (compare Figures 4 and 7); instead of being situated 

 only at the nodes of the reticulum, they now lie largely along the course 

 of its threads. This is of course further evidence for the view that, 

 with the breaking down of the nucleolar shell, separate masses of chro- 

 matin become disengaged and then migrate outward along the course 

 of the linin strands. I believe that the karyosomes at this stage 

 (Fig. 6) are no more to be regarded as discrete and individual chromatin 

 bodies than are the chromatic masses derived from the nucleolar shell, 

 but that they are merely aggregations of chromatic microsomes which 

 have gathered at different spots on the achromatic strands. This assump- 

 tion is supported by the fact that in different nuclei they vary greatly in 

 size, outline, and composition (being either dense or granular), as well 

 as in number. 



The karyosomes now arrange themselves more and more evenly along 

 the courses of the strands, while the net as a whole becomes looser, ap- 

 parently through the breaking down of some of its meshes (Fig. 6). This 

 process of condensation of the chromatin masses finally results in the 

 formation of definite segments, which, from their composite origin, re- 

 semble roughly a string of beads. These segments are entirely inde- 

 pendent of one another so far as their chromatin is concerned, but they 

 are connected by the persistent linin strands, which form their basis. 



This seems to be the nearest morphologic approach to a spireme that 

 is found in the somatic cells of Gonionemus, for I have not been able to 

 find any evidence that the net ever becomes metamorphosed into a single 

 continuous thread, such as Downing (: 05) observed in the interstitial 

 cells and male germ-cells of Hydra fusca. I have never seen a continu- 

 ous spireme in any prophase in Gonionemus, although I have searched 

 for it diligently, and although my material for the study of the sperma- 

 togonia! and spermatocyte divisions is very abundant. 



