Page:The Botany of the Antarctic Voyage.djvu/326

294 all bear to Olacinea, as was also first indicated by Mr. Brown and afterwards well illustrated by M. Decaisne, after a comparison of the ovules of Viscum album, with those of Thedum. The ripe fruit of M. punctulatum forms an achenium, which generally dehisces longitudinally and allows of the partial or complete exsertion of the seed; but I am not aware whether germination takes place by the embryo becoming thus excluded, or whether, as in the following species, the radicle protrudes at the apex of the fruit, pushing the disc and style before it. The feathery filaments (hereafter to be described) are not so long in this species as in the following, or probably as in any of its congeners, except the M. imbricatum, Pœpp., of South Chili; they are plumose with long hairs, which are capitate at the apex. The walls of the pericarp are more membranous here than in the other species, and, when fully ripe, the ealycine portion looks like three plates, attached longitudinally by part of their surface to the endocarp; the filaments being lodged in the spaces formed by their contiguous non-adherent portion (Plate CIV.f. 16.). The stout central column of the ovarium is elongated in the fruit into a slender chord, pressed between the seed and walls of the cavity of the fruit, and resembles a funiculus ; its edges are ragged from the rapid elongation of its substance. The true funiculus is extremely short, and bears at its base the two very minute unimpregnated ovules (Plate CIV.f. 18 and 19). The seed is linear-oblong, narrow and obtuse at both extremities. The albumen is copious, formed of utricles that at first are readily separable; and though the outer ones adhere closely, they do not seem covered with any distinct testa; a cavity in the upper part contains the greater portion of the embryo, which is of very highly organized tissue, and the radicle which is not immersed in the albumen is covered with a membrane apparently continuous from the funiculus, which at an early period may be traced downwards, lining the cavity of the albumen (Plate CIV. f. 19, where the membrane is represented as too thick in texture). At no time can I detect the lower portion of this membrane (first observed by Mr. Brown), except whilst the cells of the albumen are loosely held together and may be scraped by the knife from its surface, and then it appears homogenous and of a different texture from what covers the radicle, which is cellular (Plate CIV. f 20). The embryo is very small, the radicular extremity capitate, with a depression at the top; the cotyledonary terete, abrupt, slightly curved and fistulose at the apex: the cotyledons are consolidated and present no trace of any line of union. The tissue of the radicle differs materially from that of the cotyledons, which may be seen even before, but still more remarkably after, dissection (Plate CIV. f 20). The upper portion of the capitulum, above the upper margin of the cavity of the albumen, is composed of delicate filiform cells of considerable length, enclosed in a cellular cuticle of great tenuity. The lower half and terete cotyledonary portion consists of closely-packed oblong cells, projecting in the form of a cone towards the radicular end. This structure, somewhat modified, exists in M. bracJiystachi/um (Plate CV. f. 20 and 21), the tissue of the radicle being much more lax than that of the cotyledons. The peculiar functions of the radicle doubtless demand this highly organized structure, both for rapid elongation and for the sudden spread of the membrane by which the following, and probably all the species, are first attached to the bark whereon they grow. In the present, the true radicle which pierces the bark is probably the conical continuation of the cotyledonary portion.

The plumose pappi of the achenium 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 imperfect specimens, described them sufficiently accurately, as scales contained in the walls of the pericarp. Guillemin also considers them to be pappiform appendices, contained in fissures of the achenia. Neither of these