130 BULLETIN OF THE 



for the most part in the form of granules distributed evenly throughout 

 the nucleus. Indications of a reticular or filamentous structure are, 

 however, frequently present. I believe there is a chromatic network 

 throughout the nucleus, but the abundance of granular chromatin pre- 

 vents one from tracing it. Several nucleoli are always present. They 

 are extremely variable in size and shape, and in many cases appear to 

 be only aggregations of granular chromatin. They take a stain with 

 hsematoxylin and carmine in no way different from the rest of the 

 chromatin, except that it is more intense. 



A very large proportion (about four to one) of the cells of the serosa 

 contain two nuclei. These pairs of nuclei have all arisen from single 

 nuclei by amitotic division. It is obvious that division of the cell is not 

 contemporaneous with, aud does not immediately follow, the division of 

 the nucleus. In many cases, especially when the embryo is far ad- 

 vanced, cell division probably does not occur at all. Very few cells out 

 of the thousands I have examined have had more than two nuclei ; but 

 I have found several with three nuclei, and two cells with four. This 

 seems to be the maximum number. These cells of the serosa, therefore, 

 are not to be classed with multinucleate cells in which the nucleus 

 divides into a great number of irregular and unequal fragments. Here 

 the division takes place in an orderly fashion, and division of the cell 

 follows nuclear division in regular sequence, though not immediately. 



In every serosa examined, nuclei were found in process of division. 

 Some preparations furnish many more examples of division than others ; 

 and occasionally three or four adjacent cells will contain dividing nuclei 

 (Fig. 15). Very frequently, however, only one or two dividing nuclei 

 will be found in the whole serosa. It cannot therefore be supposed that 

 nuclear division is frequent ; and I have found that there are more cells 

 with dividing nuclei in the membranes of late stages of the embryo 

 than in the earlier ones. 



The first sign of approaching division is an elongation of the nucleus 

 (Fig. 4), almost always parallel to the long axis of the cell. Naturally, 

 the elongation progresses by insensible gradations from the nearly circu- 

 lar form of the resting nucleus, so that one cannot say positively that 

 the nucleus is going to divide until the elongation has become marked. 

 The absolute amount of elongation varies greatly, and is less in the 

 membranes of young embryos than in those of older ones. The example 

 represented in Figure 4 is from an old membrane, and shows almost the 

 extreme of elongation. This stage, while giving not the slightest evi- 

 dence of ordinary mitosis, is characterized by a longitudinal arrange- 



