302 PROF. OLIVER ON A NEW GENUS OF BEGONIACEZÆ. 
We have no information as to the size attained by this plant, nor as to its general 
habit, which, however, must entirely agree with that of very many true Begonias. The 
dried specimens sent home consist of leaf-bearing succulent branches, with well-deve. 
loped inflorescence and fruit; and these abundantly suffice to settle all important points 
in its structure, which has appeared to me sufficiently remarkable to deserve illustration 
in the Society's Transactions. 
Hillebrandia differs from all Begoniaceæ hitherto described in having an ovary with 
the upper one-third free and open above in the exserted portion, owing to the non- 
cohesion and divergence of the constituent carpels at the apex, as in Datisca and Reseda, 
A further difference of considerable interest, though perhaps not one to which equal 
importance ought to be attached, consists in the existence in the flower of petaloid 
organs which I cannot but regard as true petals, corresponding to the petals of most 
dichlamydeous flowers. These are the principal grounds upon which Hillebrandia is 
entitled to generic distinction. In other respects it is very closely allied to Begoma, 
and especially to Mezierea of Gaudichaud, which latter, however, presents a closer 
affinity with Begonia than with Hillebrandia. The special point of agreement between 
Mezierea and Hillebrandia consists in the placentation, which is parietal in both, the 
ovary being subunilocular, though with more or less cohesion below of the placentary 
plates. "The symmetry of Hillebrandia is pentamerous in both sexes, that of Mezierea 
dimerous in the male, tetramerous or trimerous in the female flower. But this is 
unimportant, since the symmetry in this small order is very variable, one sex in the 
same species sometimes having the sepals quincuncial, while in the other sex the same 
organs are in decussating pairs. 
The petaloid organs referred to above occur both in the male and female flower, and 
in precise alternation with the sepals. From this circumstance, together with their form 
and texture, I see no alternative but to regard them as petals. If this view be correct, 
the outer leaves of the perianth in this species must be regarded as sepals; and perhaps | | 
the inference may be extended so as to embrace the perianth-segments of the rest of 
the Begoniaceæ, which have been differently designated by different botanists—as sepals 
(Lindley), as sepals and petals (A. De Candolle), or as petals only (Klotzsch). The prin- 
cipal hesitation which I should feel in regarding the perianth-leaves of all Begonia as 
strictly homologous (using the term in a subordinate sense) with the sepals of dichlamy- 
deous plants generally, arises from the various kinds of symmetry and evolution of parts 
which obtain in the order; so that, until Similar petaloid organs shall have been found 
in flowers with a decussate development of their outer whorls, it may be the safer course 
to suspend judgment as to the general application of the terms sepal and petal in 
Begoniace. 
The remarkable character of an open ovary in a true Begoniacea is of peculiar interest 
in connexion with the question of the affinity of this isolated little Order. Ithink it must 
be sufficiently apparent that in this particular we have an important confirmation of ps 
view, first definitely advanced by Dr. Lindley, that Begoniaceæ are nearly allied to Datis- 
caceæ (Veg. Kingdom, p. 316). It is perhaps in this very character of the gaping ovary 
(so far as floral structure is concerned) that the difference between the families has bee? 
