14 
PRELIMINARY TREATISE. 
me to reform the errors which I find in the existing character 
of Amaryllidese, I will try so to trace the peculiarities which 
separate them from other vegetables, as to chalk out the 
scheme by which that whole subdivision should be arranged. 
It is incumbent on me, however, first to enumerate some at 
least of the inaccuracies which I observe in Dr. Brown’s 
character of Amaryllidese and the cognate orders, for the 
purpose of shewing that they cannot stand as at present con- 
stituted; their consideration being forced upon me by the 
subject of my undertaking. It must however be premised, 
that as Dr. Brown did not enumerate all the genera of which 
the several orders are now reputed to consist, he is by no 
means answerable for the disagreement of some of those 
genera with the characters he has defined, though further 
enquiry may have shewn that they cannot be excluded 
without inconsistency. To commence with Amaryllidese: 
1. Perianth in six divisions (sex-divisum) is incorrect, for it 
excludes the tube which is very common in the order; and 
if not meant to exclude the tube or annular junction of the 
segments, it seems incorrectly worded, and at all events 
nugatory, since all the cognate orders are equally six-cleft 
at the apex. 2. Regular is inaccurate, if I rightly under- 
stand the expression, that it is intended to indicate that the 
sepals and petals are respectively uniform, witness the genus 
Alstroemeria. 3. Stamina inserted at the base of the seg- 
ments is inapplicable, for they are frequently inserted in the 
tube, agreeing thereby with his character of Hemerocallidese, 
and sometimes on the disk of the germen. 4. Anthers 
anterior, cannot be assumed as the distinction of one order, 
unless the same distinction is carried through the kindred 
orders, yet in Melanthaceae we find anthers both anterior and 
posterior. It is however an important feature and seems 
there improperly confounded. 5. Style one is not correct; 
the style is triple or tricomposite, often tripartible. I have 
even seen it tripartite; but it is equally triple in all ttie cog- 
nate orders. 6. Stigma 3-lobed is not correct; the stigma 
in many genera is not 3-lobed, as it is stated to be, and the 
angles are even obsolete, nor is it distinguishable from that 
of Hemerocallideae, which he terms 3-lobed or simple. 
7. Pericarp 3-valved or a berry, is not an accurate fact. In 
Crinum it is not valved, yet it is not wliat Dr. Brown himself 
calls bacca or berry, but something intermediate between 
that and a capsule. 8. I demur to its ever being a berry, 
