BoT— Vol. I.] CAMPBELL— MAIAS AND ZANNICHELLLA. 27 



acropetal, the terminal cell dividing twice. He says that he 

 has verified this repeatedly, having followed the stages of 

 nuclear division, so that there was no doubt about the accu- 

 racy of his conclusions. Inasmuch as my own investigation 

 of Z annichellia also shows that the basal cell remains per- 

 manently undivided, it seems quite probable that this is 

 regularly the case and that this cell never takes any further 

 part in the formation of embryo. Further investigations on 

 the point are very much to be desired. 



The second division takes place in Naias jiexilis after the 

 terminal cell has elongated, and results in the formation of 

 two cells of unequal size. Whether the next division is due 

 to a second segmentation of the terminal cell, as Schaffner 

 describes for Alisma, or by division of the middle cell, could 

 not be positively determined; but it certainly was not formed 

 by a second division of the basal cell. The latter almost as 

 soon as it is formed begins to enlarge, and its nucleus also 

 increases very much in size, finally reaching enormous 

 dimensions, but never showing any indication of division^ 

 This, as Schaffner has shown, is also true for Alisma; and 

 Scott (1894, PP- ^^7' ^^^) ^^ ^ recent work has reached the 

 same conclusion, although it is not clear from his statement 

 whether he considers the middle of the three cells of the 

 young embryo as arising from the division of the upper or 

 lower of the two primary cells. Following these transverse 

 walls in the young embryo, there is next formed a vertical 

 wall in the terminal cell, which latter develops into the single 

 terminal cotyledon of the older embryo. This is next fol- 

 lowed by a similar wall in the cell below it, and about the 

 same time, by a transverse wall in the cell next the enlarged 

 basal cell. At this stage the embryo consists (fig. 64) of 

 the very much enlarged basal cell, followed by two short 

 secondary suspensor cells, and terminated by a nearly glob- 

 ular body composed of four cells arranged like the quadrants 

 of a sphere. Each of these sphere-quadrants next divides 

 by octant walls, and from the upper four arises the cotyle- 



iThis appears to be much less marked iu N. major (see Magnus, 1870, p. 31). 



