4IO EDWIN G. CONKLIN. 



pole of the spindle to the other. When chromosomes are merely 

 scattered throughout the cell or along the spindle they usually 

 give rise to chromosomal vesicles wherever they lie, as is shown 

 in Figs. 7, 16, 26, etc., but when in addition they are stretched 

 into elongated threads chromatic connections are left between 

 daughter nuclei. 



In diluted sea-water the chromosomes show a tendency to 

 stick together into masses and to stretch out into long threads 

 instead of dividing and moving to the two poles of the spindle 

 This is probably due to the fact that the linin basis of the chromo- 

 somes is modified so that the latter do not preserve their usual 

 shapes and do not separate normally in division. The nuclear 

 membrane also frequently remains chromatic and in such cases 

 may persist throughout mitosis (Figs. 49-53, 55, 59, etc.). 

 Evidently some of the chromatin which usually enters into the 

 formation of the chromosomes is left in these cases at the peri- 

 phery of the nucleus. Since the nucleus is composed of chromo- 

 somal vesicles more or less completely united this result might 

 conceivably be due to the swelling and bursting of some of these 

 vesicles at the nuclear periphery. 



When mitosis is halted in the prophase of the third cleavage 

 the centrosomes separate and a spindle is formed in the usual 

 manner but the nuclear membrane persists and the entire nucleus 

 becomes pear-shaped (Fig. 49, cell D), or unequally constricted 

 (Figs. 51, 52), the smaller portion corresponding to the micro- 

 meres containing almost all of the deeply staining chromatin 

 in the form of chromatic threads, while the larger portion be- 

 longing to the macromeres contains faintly staining threads and 

 granules. These two portions of the constricted nuclei are 

 approximately proportional in size to the nuclei of the normal 

 micromeres and macromeres, although in cell D, Fig. 49, and 

 cells C and D, Fig. 52, the division wall between the micromeres 

 and macromeres has not formed. The fact that these two 

 portions of the constricted nuclei are proportional in size to the 

 cells to which they belong even when the division wall between 

 those cells has not formed is difficult to explain. Generally 

 the size of a nucleus is proportional to the volume of cytoplasm 

 in which it lies (Conklin, 1912) because the chromosomal vesicles 



