COUKESrONDENCE. 127 
which the flo\\Trs are situated, but whether the writers intend to say the bracts 
of different flowers are so close as to he almost connate, or whether thej intend 
to say that each flo^ver is subtended by some two, three, or four bracts united 
together, I cannot tell; if the latter is intended, you will see my opinion 
differs. 
The flower is slightly sunt in the stem, and this naturally raises a kind of 
edge round it, which looks like rudimentary connate bracts ; the writer, how- 
ever, may not mean this, but only that the bracts of the different flowers are 
crowded close on each other. 
It seems to me certain that the perianth is a calyx, so I suppose those who 
can make up their minds that it is so, will be able satisfactorily to fii the genus 
in its proper place. I see you referred it (Eot. Herald, p. 98) to Samt/decB, and 
I now think that it belongs to an apetalous section of BixacecBj in the position 
of the raphe it differs from the Violacem I hare examined. 
I have written the following description of the flower : 
TEXfiATHTLACiTTM, PtKj:?^. —Fiower sessilc in the axil of a bract, the margin 
of which extends round its base, or else the flower is imbedded in a shallow 
cavity ; the margin of the cavity being continuous with the edges of the bract, 
so that it cannot be seen where one ends and tlie other begins. Bracts 2, right 
and left of the axis, very short but nearly meeting before and behind.* Sepals 
4, united into a tube, much swollen and indurated at the basej two anterior 
and posterior relatively to the axis, and two right and left, so that the latter 
are opposite the two bracts. Stamens 4, altemating with the sepals, insei-ted 
into the lower third of the calyx ; filaments broad at the base ; anthers introrse 
•with a rather thickened connective, which ia forked below. Ovary attached 
"by a broad basis, densely covered with short hairs, ^hich extend in a slight 
degree on to the otherwise smooth calyx, 1-celled with 4 broad parietal pla- 
centae ; ovules numerous, anatropal, mostly liorniontal or obliquely ascending 
with the raphe lateral, the uj>per in part ascending with the raphe next the 
placenta. 
I have seen Baillon's ' Adansonia' (1862-1863), and find he has very clearly 
figuvedi Sin apis arvensis, having ovaries with naked ovaries occupying tlie place 
of ovules, and even small fully- formed flowers (pL xii. figs. 8, 9). I take this 
as a final proof that my view of the structure of the ovary of Nelumhiiun 
(Ann. Nat. Hist, 2nd ser. vol. xii. p. 12) is correct, i. e. it consists of a com- 
pound inflorescence, the disk in which the carpels are immersed being a mass 
of barren carpels, each producing one flower at its base. Each carpel is tJiere- 
fore a naked flower, and it stands with its back to the axis and its ventral 
suture directly towards the stffmens. I consider the French botanists tqvj 
clever, and their views as regards the Crucifer<B and other famihcs of doubtful 
structure are, I believe, well worthy of attention. 
The flowers of Nehimlimn are so beautiful that they are seldom cut, which 
to some extent accounts for the singular structure of the ovary not having 
atti'acted attention. 
* If I rightly understand the descriptions given of this genus, Muse are 
regarded as a calyx. 
