76 PROFESSOR HENFREY ON THE DEVELOPMENT OF THE OVULE 
it appears to me can be urged in favour of the idea that the central body is not a placenta, 
but a compound nucleus with three embryo-sacs, are as follows d : 
The central body is stated by Griffith to form at first, in Osyris, but a slightly -elevated 
cone with three (or four) papillæ (or nuclei) at its sides. In the youngest specimens of 
Santalum I examined, the placenta was highly developed as a pillar in the centre of the 
ovary (Tab. XVII. figs. 1 & 2). The papillæ might be either three points of one ovule, 
or three nuclei with three chalazal ends blended in the centre, and the growing-up of the 
central column, which goes on for a certain time after the embryo-sacs are distinguish- 
able, might be regarded as a conversion of the compound ovule, or the three conjoined 
ovules, into an anatropous structure, the chalazal end being at the apex of the conical 
mass, the micropyles free and turned down next the “ funiculus.” The subsequent retro- 
version of the embryo-saes would still be anomalous, connected however in some degree, 
through Osyris, Avicennia* and Myzodendron+ (in which the embryo-sac is not extended 
unless fertilized) with the ordinary conditions. Then the growth of the posterior ends of 
the embryo-sacs would seem to represent a kind of chalazal structure, connecting them 
below in a kind of central raphe with the peduncular placenta. The relative positions of 
the remains of the nuclei and the central body, and of the remains of the peduncle of the 
placenta, in the ripe fruit, would then indicate a rupture at the umbilicus, which was 
carried up to near the top of one side of the seed, while the funiculus was driven outwards 
on the same side, and almost obliterated f. 
Some degree of likelihood attaches to the explanation just given when the Santalacee 
are compared with the Loranthacee, with which they appear to be nearly connected 
through Myzodendron (which genus is nearer the former than the latter family as regards 
ovulary structure$). In the Loranthaceæ the supposed compound, triple ovule, being 
atropous instead of anatropous, would naturally be still more completely combined into 
one piece; it would here be almost a question of words whether there were confluent 
nuclei or one nucleus with three parallel embryo-sacs. Then the phenomena observed in 
Loranthus might afford a still farther confirmation, as indicating analogy with the Coniferæ, 
in which the existence of a number of (secondary) embryo-sacs is the rule. 
These points deserve further attention, and can only be elucidated by the study of the 
development in more of the genera of these remarkable families. I should not omit to 
mention here the observations of DeCaisne on Thesium ||, which I have not yet repeated, 
but which agree essentially with the statements respecting Santalum made by Griffith 
and myself. 
The reason which perhaps most of all induced me to present these observations to the 
—- of the Society, is the remarkable analogy which is shown to exist between the 
. phenomena of fecundation above described, and those which have recently been demon- 
isti ished in the conical plac > aan, ine to Griffith, 
(I. e.) the vascular cord eme mecal placenta ; in Avicennia, according to 
reaches nearly to the base of the embryo-sac, so that the chalazal end of the ovule is marked. 
