i 9 2o] DUPLER—TAXUS CANADENSIS 505 



Megasporangium. — In T. baccata Strasburger (36) pointed 

 out the hypodermal origin of the archesporium, describing it also 

 for Larix euro pea. In T. canadensis the sporogenous tissue is 

 also hypodermal in origin, the archesporium becoming differentiated 

 very early in the development of the nucellus while it is yet cone- 

 shaped and the integumentary zone in a rudimentary condition 

 (figs. 38, 61). It may consist of a single cell or a small plate of 

 cells. The periclinal division of the archesporium results in the 

 primary wall cell and the primary sporogenous cell (fig. 62). The 

 wall cell, together with other adjacent cells of the nucellus, divides 

 repeatedly by periclinal divisions, building up a considerable mass 

 of tissue between the sporogenous tissue and the epidermis, the 

 cells of this tissue being in radial rows, at the inner ends of which 

 are the sporogenous cells (figs. 63-65). Morphologically this is 

 the outer portion of the many-layered wall of the megasporangium, 

 and together with the epidermis constitutes the upper portion of the 

 nucellus. The later development results in a considerable mass 

 of sporogenous tissue (fig. 64) , out of which one or more cells func- 

 tion as megaspore mother cells (fig. 65), as pointed out in my 

 previous paper (12). While I have no preparations showing divi- 

 sions of the primary sporogenous cells, the amount of sporogenous 

 tissue present indicates that this takes place, contrasting with the 

 situation in which the primary sporogenous cell functions as the 

 megaspore mother cell, as is probable in most conifers. 



Growth of nucellus. — By the formation of the integument 

 the nucellus becomes limited to a knob, at first conical; but with 

 the development of the megasporangium it soon becomes rounded. 

 From the growth of the wall, as just described, there develops a 

 considerable mass of tissue above the sporogenous tissue. At first 

 this tissue seems to be uniformly meristematic, but later division 

 becomes confined to the inner portions, the outer cells and the 

 epidermis becoming radially elongated. I was not able to find any 

 actual periclinal divisions of the epidermis, but the position of the 

 cells in the layers next to the surface (fig. 65) would indicate such 

 divisions as Strasburger (36) found in the development of the 

 nucellus of T. baccata, giving a several-layered epidermis. The 

 nucellus, therefore, is composed of two morphological entities, 



