28 CALIFORNIA ACADEMY OF SCIENCES. [Proc. 3D Ser. 



don; from the lower ones, the middle part of the embryo 

 and stem-apex. There is now formed in each octant cell a 

 periclinal wall, separating it into an inner and outer one, the 

 latter constituting the dermatogen of the upper part of the 

 embryo (fig. 65). 



While these changes have been going on in the upper 

 part of the embryo, the cell next below has enlarged later- 

 ally, and the one below this divided again. In the former 

 there are then formed quadrant walls, and each of the quad- 

 rant cells next separates into a peripheral, or epidermal cell, 

 and an inner one, exactly as was the case in the two upper 

 divisions. 



Up to this point all the secondary divisions, except the 

 periclinal ones in the upper section of the embryo, have 

 been vertical, and the limits of the primary segments are 

 still very evident (fig. 66). The first transverse divisions 

 appear in the middle segment, and this determines the sepa- 

 ration of the stem from the root, the bulk of both of these 

 organs being derived from this segment of the embryo. If 

 the accounts of Hanstein and Famintzin are correct, Alisnia 

 differs very essentially from Naias, in that the root of the 

 former is not the product of the middle cell, but of the one 

 next the suspensor, and is derived from it by a secondary 

 division. As the term hypophysis, used by Hanstein for 

 this basal segment, implies a structure not belonging strictly 

 to the embryo itself, but to the suspensor, it would seem 

 best to discard this term, as Schaffner has shown that in 

 Alisma, also, the primary suspensor cell undergoes no 

 further divisions and all the parts of the embryo are deriva- 

 tives of the primary embryo cell. In both Hanstein's and 

 Famintzin's figures, transverse walls are shown in the 

 terminal segment before any periclinal walls appear, but it 

 is quite possible that we really have the same state of things 

 as in Naias, and that the two lower cells of the four terminal 

 sphere-quadrants really belong to the second of the primary 

 segments of the embryo and not to the terminal one. How- 

 ever, as the cotyledon of Alisma is larger at this stage than 

 that of Naias, and the whole embryo longer, it is quite 



