194 Johnson . — On the Nursing of the Embryo 
its wall stretches and forms the outer wall of the superficial 
layer of the endosperm. It is necessary to distinguish between 
the two directions of prolongation of the embryo-sac. The 
position of the unfertilised embryo-sac is represented in the 
seed by scarcely more than the meeting-point of the pla- 
cental embryo-sac tube and the upper endosperm-cells, the seed 
proper lying below this meeting-point. Such is the increase 
in size of the seed that the thin-walled cells surrounding the 
central strand of cells continuous with the placental pillar are 
torn through ; this strand with the placental column (the two 
containing the placental embryo-sac tube) is pushed to one 
side, and a single cavity extending from top to bottom of the 
fruit comes into existence, and is in the end as much filled by 
the seed as was the ovary by the ovules (Figs. 16, 17). The 
ripe seed is not terete but ‘ ovoideo-triquetrum,’ 1 endosperm 
and embryo being, as seen in transverse sections, trigonous 
(Fig. 17). A comparison of Fig. 15 with Fig. 16 will show 
that the relative positions of the parts has remained unaltered, 
and will explain how the embryo is placed with radicle 
superior. The embryo is described in the Flora Antarctica 2 , 
and previously by Robert Brown 3 , as completely enclosed in a 
‘funicular membrane.’ Grisebach and Hofmeister 4 were un- 
able to find any trace of this structure, a result agreeing with 
the present examination. From the figures given in the Flora 
Antarctica it is, I think, possible to explain the cause of the 
contradiction of the two statements. The exact relation of 
the embryo to the endosperm and to the placenta has not, up 
to the present, been shown. It will be seen from Fig. 17 that 
the embryo lies in an apico-lateral lobe of the endosperm in 
the loop formed by the placental embryo-sac tube, and that 
the radicle is not exserted. The endosperm-cells enveloping 
1 Guillemin in Delessert’s leones Selectae, III, f. 47, t. 80, 1837* Guillemin, 
however, described the seed as pendulous from apex of locule, and saw no sign of 
an embryo, in M. oblongifolium. 
2 J. D. Hooker, op. cit. p. 301, figs. 19-21 in plate 104. 
3 R. Brown, in Trans. Linn. Soc. xix, p. 232 (note). 
* Abhandl. Gesellsch. Gotting. 1854, vols. 5 and 6. Grisebach, p. 109. 
Hofmeister, p. T33 (Appendix). 
