286 



F. O. BOWER. 



form the two primary phloem masses of the root. The 

 other bundles pass downwards in their original position, 

 their phloem merging with the phloem of the root, while 

 their xylem ends ^blind in the cambium of the root (figs. 

 24, 25). 



The position of the feeder relatively to the plane, in which 

 the two larger bundles and the primary xylem groups of the 

 root run, is not constant. In 24, a — c, d — -f, are repre- 

 sented two extreme cases ; in the one case {d — -f) the plane 

 in which these bundles run coincides with the median 

 plane of the feeder ; in the other (a — c) it is at right angles 

 with the median plane of the feeder ; in other cases the 

 planes were found to cut one another obliquely. Whether 

 the relative position of the cotyledons and of the primary- 

 tissues of the root is constant I have not been able to 

 determine, but have no evidence to the contrary, and it 

 must be remembered that in allied plants (Welwitschia and 

 Ephedra) it is constant. 



The inconstancy of the position of the feeder relatively, 

 on the one hand to that of the cotyledons, and on the 

 other to that of the tissues of the root, coupled with the 

 constancy of its position on the under side of the embryo 

 in the germinating seed, shoAvs us (1) that the direc- 

 tion of gravity has a direct influence upon the point at 

 which it appears, while the relative position of the parts of 

 the embryo has not ; and (2) that the direction of gravity 

 has in this case no determining influence upon the position 

 of the parts of the embryo. 



The bundle sheath appears first about the same level as 

 the lower side of the feeder, and on the opposite side of the 

 axis to that structure (fig. 24, h) it encloses a pericam- 

 bium several layers of cells thick. The primary structure 

 of the root resembles that of Welwitschia, &c., and need not 

 be described in detail. The primary xylem masses fuse 

 centrally at some distance below the feeder, but paren- 

 chyma is interspersed between the elements. 



As in Welwitschia, the cuticularisation of the bundle 

 sheath begins on the radial walls, and spreads later to the 

 tangential walls. At points opposite the primary xylem 

 masses, however, the change in the tangential walls is de- 

 ferred, and these cells of the bundle sheath, together with 

 some cells of the cortical tissue, take part with the peri- 

 cambium in the formation of the lateral roots ^ (cf. ' Conif. 

 und Gnet.,' p. 348). 



1 This observation led me to study afresh the development of the lateral 

 roots of Welwitschia. I find that my former opinion (* Quart. Journ. 



