AMAR YLLIDACE/E. 
121 
summer time, and will endure the winter if planted pretty 
deep in light soil and covered over with leaves in the cold 
season, especially if any sloped heading be laid on to throw 
off the wet. Even Acutifolia, which in the greenhouse 
keeps its leaves through the winter, will succeed with that 
treatment. 
The genus Bomarea was named by Mirbel, who divided 
it from Alstroemeria, founding the separation on edulis, 
erroneously named salsilla, ovata, and multiflora, a species 
which cannot be identified and must be altogether erased, 
having been described only from a drawing. Poiret (suppl. 
encyl. meth. p. bot. t. 1. p. 677) states that Mirbel's charac- 
ter of Bomarea was perianth and filaments straight, capsule 
rounded and flattened from top to bottom; of Altroemeria sepals 
reflex ? (renver sees) filaments curved, capsule lengthened, which 
appeared to Poiret insufficient. Mirbel was on the right 
scent, but he had not laid hold of the substantial differences. 
The filaments of B. salsilla are much bent. The elongation 
of the capsule is a weak feature, and varies in the different 
species of Alstroemeria. 
Suborder 3. Agaves. — Schistandrous, not operculate. 
§. 1. Dioscoreasformes. — Inflorescence axillary, spicate. 
Plants tuberous, climbing, dioecious, with the 
veins of the leaves usually, but not always, con- 
fluent. (Petiole articulate ?) 
Dr. Lindley relies on the articulation of the footstalk of 
the leaf in Dioscoreae, and a small portion of Smilaceae, in- 
cluding only Smilax and Ripogonum, of which I believe 
but one species is known, for the character of an order, and 
even of a group of orders ; but I doubt there being such a 
decisive difference between Dioscorea and Bomarea ; and 
if it were established, I do not see how such weight can be 
attributed to that feature, while articulations of the seed-pod 
are not considered of sufficient importance to effect a like 
separation. I believe that the axillary inflorescence will be 
found a constant feature in the plants I have characterized 
as Dioscoreaeformes. Dioscoreae may *be taken as a third 
suborder, and Agaveae confined to Ixiaeformes and Agavi- 
formes, if it be thought preferable. It is immaterial, except 
with a view to consistency, when the other orders shall have 
been rectified and arranged. 
