420 MORPHOLOGY OF GYMNOSPERMS 



. these details have been obtained is far too small to make them of 

 present service in this connection. 



The general tendency which runs through gymnosperms as a 

 whole, and which reaches its extreme expression among angiosperms, 

 is to mature the eggs earlier and earlier in the ontogeny of the game- 

 tophyte. In the most primitive condition of the gametophyte, the 

 eggs do not appear until the endosperm is nearly full grown; and 

 other gametophytes can be selected and arranged in a series showing 

 the gradual slipping-back of the egg in the ontogeny of the gameto- 

 phyte, until in such a form as Torreya the archegonium initial is 

 differentiated as soon as wall-formation has taken place. A conspicu- 

 ous illustration of the inequality of response to such a general tendency, 

 among related forms is furnished by Torreya and Cephalotaxus, the 

 archegonia not appearing in the latter genus until the gametophyte 

 is well grown. The next stage is illustrated by the situation in Wel- 

 U'itschia, in which eggs are matured before wall-formation is complete, 

 resulting in the elimination of archegonia. The extreme stage in 

 this progressive series of changes among gymnosperms is illustrated 

 by Gnetum, in which eggs are matured at the stage of free nuclear 

 division, the most embryonic stage of the female gametophyte. 



So far as the living forms are concerned, the Cycadales and Gink- 

 goales show little, if any, response to this tendency; and therefore 

 possess the most primitive type of female gametophyte among living 

 gymnosperms. Among the Coniferales, on the other hand, all the 

 stages are found, up to the one just preceding the elimination of 

 archegonia; and this stage is attained by Welwitschia and 

 Gnetum. 



The general tendency of the archegonia among gymnosperms is 

 to eliminate the ventral canal cell. The gymnosperms are distin- 

 guished from the pteridophytes by the complete elimination of neck 

 canal cells, and this tendency to suppress all of the axial row except 

 the egg continues among gymnosperms. Among the living forms 

 a walled ventral canal cell is retained only among the Abietineae and 

 by Ginkgo, so far as known; but it seems to be a safe inference that 

 it was present among the ancient gymnosperms. In the other living 

 groups the wall has disappeared, and the ventral canal cell is repre- 

 sented by a free nucleus. In certain forms even this nucleus may 



