282 F. 0. BOWER. 



was observed (fig. 14). It must be noted, however, that 

 these observations were made on buds dormant in winter. 

 It is possible that at periods of great activity such divisions 

 may occur, and this has been observed by Strasburger during 

 the development of the ovule of G. Gnemon (' Ang. and 

 Gymn.,'Pl. XI— XIV). 



The form of the embryo and the mode of arrangement of 

 the procambium bundles will be seen in fig. 10. The growth 

 from this point onwards is very rapid. Shortly after the 

 embryo has reached the stage represented in fig. 10 a lateral 

 protuberance appears on the hypocotyledonary stem (fig. 11), 

 the position of which is constant as regards the direction of 

 gravity, i. e. it always appears on the under side of the em- 

 bryo, as the seed lies during germination ; its relation to the 

 position of the cotyledons is, however, inconstant, the planes 

 in which they are expanded having no direct relation to the 

 direction of gravity. 



Externally this outgrowth has a rounded apex, and as its 

 development proceeds it assumes an approximately cylin- 

 drical form. It arises through a lateral extension of the 

 tissues of the hypocotyledonary stem, the epidermis, cortex, 

 vascular bundles, and pith all taking part in its formation. 

 It is the morphological equivalent of the structure already 

 described under the name of "Feeder," as appearing late- 

 rally at the base of the hypocotyledonary stem of Wel- 

 witschia {' Quart. Journ. Micr. Sci.,' vol. xxi, new ser.) ; 

 there is, however, this difierence between them : that in 

 Gnetum the cortex and epidermis do not form the main bulk 

 of the organ, as is the case in Welwitschia, but rather the 

 pith and vascular bundles. The latter may be seen at an 

 early stage curving outwards from the hypocotyledonary 

 stem into the organ (figs. 11, 12), Avhile in the fully 

 developed feeder they extend almost to its extreme apex. 



As in Welwitschia, the growth of the organ is rapid ; it 

 soon overtops the hypocotyledonary stem, and finally attains 

 a length of about half an inch (fig. 15). As its length in- 

 creases, its apex being fixed in the body of the endosperm, 

 the weaker point at the mycropylar end of the seed gives 

 way; first the radicle and then the body of the embryo are 

 gradually pushed out through the ruptured testa. Direct 

 evidence of the pressure exerted on the apex of the feeder is 

 afforded by cases such as that in fig. 16, where the end of 

 the feeder is folded upon itself. Immediately on protruding 

 from the seed the radicle curves downwards, while the sus- 

 pensors and the apex of the nucellus remain attached late- 

 rally to it at a point near the level of the seed. Shortly 



