80 



which I have observed points to an origin from a single cell 

 cut off from the apex of the suspensor, I have not seen any 

 specimen of this except one doubtful case. This cell divides 

 by anticlinal walls, of which none appear to be exactly median 

 (fig. 3—7) ^). The pheripherical cells of the group thus formed 

 grow laterally along the surface of the suspensor, and divi- 

 ding further by anticlinal walls form short embryonic tubes 

 comparable to those of Welwitschia though much less developed 

 than these (cf. Strasburger, Angiosp. und Gymnospermen p. 155) ^). 

 The whole embryo now presents the appearance of a single 

 layer of cells covering the apex of the suspensor like a hood 

 (fig. 5). A cell is next cut off from the suspensor by a trans- 

 verse wall (fig. 8) ^). Then follows longitudinal division of the 

 cell thus formed, while in the lateral parts of the peripheral 

 layer cells now begin to divide by periclinal as well as by 

 anticlinal walls perpendicular to the outer surface of the em- 

 bryo, and inclined to one another (fig. 6 — 9) '). This cell ap- 

 pears to be wedge shaped, in fact we have to deal with a 



growth with an apical cell 



By this mode of increase of the peripheral tissue and 



by divisions both longitudinal and transverse of the central 

 group of cells, a large embryonic body is formed in the apical 

 part of which the tissues differentiate in a manner similar to 

 that well known in other member of the group. Internally a 

 root apex is formed while externally there appears at the apex 

 of the embryonic body two cotyledons and a central apical 

 cone (fig. 10). A definite epidermis covers the cotyledons and 

 the hypocotyledonary stem, but, as is the rule in the Gym- 

 nosperms, the root has no such covering, the same is the case 

 with the conical apex of the stem; here the peripheral cells 

 divide by periclinal walls, while the extreme apex is not un- 

 commonly occupied by a single large cell (fig. 13). 



1) Fig. 6 here reproduced as fig. 64. 



2) And his plate XXII, fig. 81—86 (Ref.). 



3) Here reproduced as fig. 65. 



