BEHAVIOR OF NUCLEUS AXD CHROMOSOMES. 155 



tially the same features as does Asilns. Hence they \vill be 

 passed over rapidly. 



The pairs of chromosomes appear soon after the final spermato- 

 gonial telophase as granular aggregates (Fig. 2). They are less 

 druse and clear cut than in Asihis, and more nearly resemble 

 those in Dasyllis (Metz, '22). These aggregates soon loosen up, 

 overlap, and spin out into long, granulated threads, intermingled 

 in a net-like structure. The nucleolus, a dense, spherical body, 

 is prominent from the earliest observed stages (see special account 

 of the nucleolus below). 



Soon after the growth period has begun, the chromatin threads 

 become drawn out so finely that they cannot be traced far, indi- 

 vidually. They are heavily granulated, with prominent chro- 

 momeres. From their history and from their subsequent be- 

 havior, as well as from analogy with other forms, they are assumed 

 to be bivalent, although no duality is visible in their structure. 

 Then- is no conspicuous polarization of the threads at this time, 

 nor is there any evidence of a synizesis or dense contraction 

 stage. The nucleus is clear cut in outline and is of the usual more 

 or less spherical shape, as is also the nucleolus. 



At this point changes set in which accompany one another as 

 growth progresses. The nucleolus elongates and eventually be- 

 comes double; the chromatin threads undergo a gradual con- 

 densation; and the smooth contour of the nucleus becomes 

 broken by a series of imaginations and evaginations, at first more 

 or less irregular, but soon conforming to the outline of the thrr.nl- 

 liki- chromosomes. It is difficult to determine what phy-iral 

 changes occur in the cell; but the nucleus acts as if its tension 

 were released, so that the nuclear membrane offered no n-i-t.mce 

 to movements of the cytoplasm or of the chn>m"-'>im--. The 

 former appears to flow in around the chromatin thread-, while 

 tin- latter become extended out into tlir < -\ -topla-m. Si. mi tin- 

 nucleus is converted into a series of lonu l<>U-s or jx.rkrt- rami- 

 fying throughout the cell. Each pocket, .1- .1 rule, encloses only 

 one (bivalent) chromatin thread. These changes are n -|>n -( nted 

 by the cross sections shown in Figs. 2 to 4. The -urf.i-v of the 

 nucleus is tremendously increased by this proce, bin tin- volume 

 is so difficult to estimate that we are unable to determine whether 

 it is affected or not. 



