in Myzodendron punctulatum, Banks et Sol. 185 
posed, naked, but is partially covered by a single integument, 
formed of a single layer of large cells, leaving the apical part 
of the ovule naked, in which region the single layer of 
nucellus cells covering the embryo-sac is for the most part 
shrivelled (Fig. 7). The cells in the immediate neighbour- 
hood of the egg-apparatus are well developed, and are very 
easily mistaken for the first endosperm cells of a young seed. 
I was not able to find any young ovules, and saw the embryo- 
sac in its mature condition only when ready for fertilisation. 
At this stage its cavity is full of granular protoplasm, in which 
the egg-apparatus and secondary nucleus are as in typical 
angiosperms. The antipodal cells, which were often not seen, 
possibly by reason of the granularity of the general contents 
and narrowness of the embryo-sac, appeared to vary in 
position, being sometimes at the end of the narrow part of 
the sac, at other times in the general body of the sac. The 
trigonous sessile flower is characterised by the presence at each 
of its three angles of a pair of longitudinal cellular lamellae, 
which form a groove, extending from near the apex to the 
base of the flower and enclosing a filamentous barbed appen- 
dage which has received amongst other names that of 4 seta/ 
The surface of the groove is formed by a true epidermis, for 
on it are numerous simple stomata. The nature of the seta 
has such an important bearing on the right understanding of 
the homologies of the flower that I have placed before readers 
the following description of them from the Flora Antarctica 1 . 
‘ The seta ascends from the base of the ovarium, and gradually 
elongating finally escapes from the cavity where it was 
lodged (Plate CIV. Fig. 10). . . . The plumose pappi of the 
achenium [the setae of the flower] afford one of the great 
peculiarities of this genus ; of their function there can be 
no doubt, though their origin and true nature are not quite 
so evident. De Candolle, from an examination of very im- 
perfect specimens, described them sufficiently accurately as 
scales contained in the walls of the pericarp. Guillemin 
1 J. D. Hooker, FI. Antarct. p. 292. 
