234 
corolla but from the margin of the sinuses, so that the corolla might be de- 
scribed as 10-cleft, five of the divisions being broad and coloured, and the 
other five much narrower, and shorter, colourless, and having anthers ; and 
secondly, in the embiyo being a filiform body, slightly 2-lobed at one end. 
But both Diapensia and Pyxidanthera disagree with Hydroleaceae in having a 
calyx consisting of five unequal sepals forming a broken whorl as in the caly- 
cose group of Polypetalous Dicotyledons ; secondly, in having the anthers 
bursting transversely, and with a very broad connective ; thirdly, in having 
only one style instead of two ; fourthly, in being destitute of an hypogynous 
disk ; and finally, in the embryo lying in the midst of fleshy albumen across 
the hilum, so as to be what modern Botanists call heterotropous. At least 
this is certainly the case in Pyxidanthera, and I have no reason to doubt its 
being equally the case with Diapensia. In Hydrolea certainly, and I believe in 
Wigandia also, the embryo is orthotropous, and the quantity of albumen so in- 
considerable as to be of no moment. It is plain, therefore, that Diapensia and 
Pyxidanthera must be separated from Hydi^oleacese. With Polemoniacese I have 
already stated that their atfinity appears to me to be much greater. In the 
first place they are not very difi’erent in habit ; compare Pyxidanthera barbu- 
lata with Phlox setacea. Secondly, they agree in the trimerous structure of 
the ovary, and in the single style. But on the other hand DiapensiaceaB have 
a totally different calyx and embryo, their corolla is not the corolla of Polemo- 
niaceae, and they have no trace of an hypogynous disk. I therefore think it 
absolutely necessary to give them a distinct station. The characters of the 
two genera, usually, but incorrectly, combined, will be found in the Ap- 
pendix. 
Let me add, that although the name of Diapensiaceae originated with Link, 
yet that author in placing it in Convolvulaceae was obviously unacquainted with 
its real structure, and in assigning it for a character “ semina membrana in- 
clusa,” seems to have assumed that in this respect it agrees with Hydroleaceae, 
which is not the fact. 
Geography. Mountain plants of the North of Europe and North Ame- 
rica. 
Properties. Unknown. 
GENERA. 
Diapensia, L. (35) 
' Pyxidanthera, Michx. (36) 
Order CXXVIII. HYDROLEACE^. 
R. Brown Prodr. 482. (1810) without a name; Id. in Congo (1818). — Hydroleace.e, 
Kunth in Humb. N. G. et Sp. 3. 125. (1818) ; Synops. 2. 234. (1823) ; Bartl. Ord. 
Nat. 189. (1830) ; Choisy Descr. des Hydroleacies (no date). 
Essential Character. — Calyx 5-parted, inferior, persistent, with slightly imbricated 
aestivation. Corolla hypogynous, monopetalous, regular, not always agreeing with the calyx 
in the number of its divisions. Stamens arising from between the lobes of the corolla, regu - 
lar, agreeing in number with the segments of the calyx ; anthers deeply lobed at the base, 
opening longitudinally, with a very narrow connective; filaments petaloid at the base. 
Ovary superior, surrounded by a short annular disk, 2- or 3-celled; styles 2 or 3 ; stigmas 
thickened. Fruit capsular, papery, enclosed in the calyx, 2- rarely 3-celled, sometimes 
splitting through the middle of the cells ; placentce either single and fungous, or double 
and thin, sometimes central and loose. Seeds indefinite, very small, with a testa more or 
less loose ; albumen fleshy, in the axis of which lies a taper, straight embryo, which is 
orthotropous and slightly 2-lobed at one end. — Herbaceous plants or under-shrubs, some- 
