106 TWENTY-FIRST REPORT, 



sions. In the later cleavage stages the nuclei are typically, and perhaps 

 always, likewise double, but irregularities occur in these stages and the double 

 nature of the nucleus is not always demonstrable. 



In the fertilization of the egg of Crypto'branchus allegheniensis the egg- 

 iiucieus and the sperm-nucleus do not fuse, but come to lie side by side with 

 nuclear membranes intact. During the long resting stage that precedes the 

 formation of the first cleavage spindle, each germ-nucleus maintains strict 

 individuality ; there is close association, but no actual union, and certainly no 

 mingling of structural contents. The two asters take up positions on opposite 

 sides of the resting germ-nuclei, in the same horizontal plane, and close to the 

 region of contact of the germ-nuclei. The rupture of the nuclear membranes 

 and the formation of distinct chomosomes does not take place in the two germ- 

 nuclei simultaneously, but one of the germ-nuclei becomes active somewhat in 

 advance of the other. 



In the fully-formed mitotic figure the two groups of chromosomes, of 

 maternal and paternal origin, respectively, remain visibly distinct. Further, 

 in the anaphase these two chromosome groups of diverse origin may be traced, 

 in a series of preparations, step by step into the daughter-nuclei, where each 

 group becomes enclosed in a separate nuclear vesicle. Each daughter-nucleus 

 thus comes to consist of two distinct nuclear systems, derived from the egg- 

 nucleus and the sperm-nucleus, respectively. Except for a diminution in size, 

 the appearance of each daughter-nucleus closely simulates that of the original 

 fertilization-nucleus with its two resting germ-nuclei. 



The orientation of the two nuclear vesicles of the newly-formed daughter- 

 nucleus is somewhat variable, but the final position assumed in preparation 

 for the second mitosis is established with great regularity. The two asters 

 take up a position on opposite sides of the daughter-nucleus, such that a line 

 connecting them lies parallel to the newly-formed first cleavage furrow ; if the 

 two nuclear vesicles are not already ranged midway between the two asters, 

 they rotate to this position. This arrangement ensures a segregation of mater- 

 nal and paternal chromatin in the ensuing division, and the second cleavage 

 mitosis is in all essential respects a repetition of the first. 



This same type of cell division persists through the early cleavage stages 

 at least, though there is increasing difficulty in distinguishing the maternal 

 and paternal chromosome groups through the most active stages of mitosis. 

 In the resting stages of the nuclei of the late blastula, the structure is still 

 typically double, but it is often quadruple, with four distinct nuclear vesicles 

 of equal size. In still other cases there are three nuclear vesicles, one of them 

 equal in bulk to the other two combined. It is comparatively rare to find 

 nuclei divided in any other ways than those mentioned, but deeply-lobed nuclei 

 are quite common in these later stages. 



These observations give strong support to the principle that maternal 



