26 
PRELIMINARY TREATISE. 
ing by a vertical slit from top to bottom. They seem divisi- 
ble however into Operculous and Nonoperculous. By oper- 
culous I mean having the base of the style which is enlarged, 
persistent after the decay of the rest, and forming a hollow 
top to the ovary, more or less prominent, and becoming part 
of the seed-vessel. The operculous may be called Hypoxideae, 
and divided into 1. Hypoxidiform, 2. Lanariaeform, covering 
the hexandrous part of the present Haetnodoraceae, 3. 
Alstroemeriaeform ; or those portions may form three subor- 
ders. The nonoperculous may be divided into Dioscoreae 
and Agaveae, the latter into Ixiaeform and Agaviform. The 
scapaceous are divisible intoSchistandrous, containing the true 
Amaryllideae and Narcisseae, and Porandrous of which the an- 
thers open only by the upper part, which I call Galantheae. 
The hypogynous division consists of Liliaceae, that is 
Asphodeleae, Hemerocallideae, Liliaceae or Tulipaceae accord- 
ing to the separations of different writers, Smilaceae and Me- 
lanthaceae. A portion of Melanthaceae is clearly separable 
by posterior anthers, those with anterior anthers being 
improperly joined with them. The separation of He- 
merocallideae cannot be possibly maintained, the main dis- 
tinction of the seminal integument being contrary to the fact, 
and the difference of a longer tube quite trivial for the cha- 
racter of an order. Smilaceae, according to Dr. Brown’s ad- 
mission, are distinguishable from those which have been 
called berried amongst Asphodeleae, only by the stjde being 
oftener tripartible, which, not being absolute, is no distinction. 
This hypogynous division, Liliaceae, maybe properly divided 
into suborders as the Amaryllidaceae, and perhaps in this man- 
ner, Branching, Caulescent, and Scapaceous. Caulescent 
erect, Asphodeleae. Caulescent twining, Smilaceae, inclu- 
ding asparagus medeola, &c. Scapaceous ; anthers anterior, 
Allieae ; anthers posterior, Melanthieae; the termination aceae 
being reserved for the higher division. I can lay no great 
stress, except as a generic distinction, on the difference of 
seminal integument, when I look to the seeds of Leucojum 
and Galanthus, Pancratium and Hymenocallis, nor of cap- 
sule valved or valveless, which latter has been erroneously 
called a berry, when I look to Haemanthus and Buphane, nor 
of erect and twining stem when I look to Bomarea and Al- 
stroemeria, all respectively confounded heretofore in genus ; 
nor on the dioecious character, when I see one species of vitis 
dioecious and another not so ; but they are features which 
