PRELIMINARY TREATISE. 
11 
knowledge ; and I would earnestly excite him to perfect 
that system upon which he has thrown a strong, but as yet 
insufficient, light. I apprehend no ultimate difference of 
opinion between me and him, or any other person of clear 
understanding and unprejudiced mind, because I build only 
on fact, and by that test I wish every thing I advance to be 
tried, and should instantly correct any thing I found incon- 
sistent with it. It is a main object of this Treatise to reduce 
the divisions which rest on opinion, to their proper insig- 
nificancy. 
The principal merit of the Linnaean division was, that 
each separation rested on a single fact of pretty easy access, 
although it might remove to too great a distance genera, 
associated by other peculiarities. The system now in vogue, 
will be found to rest in many respects on features quite as 
artificial, which separate kindred genera as objectionally as 
its predecessor. I am at a loss to conceive, in what 
manner it can possibly be substantiated, that the position of 
the stamens adopted by Jussieu is a more natural feature 
for classification than their number, by which Linnaeus was 
guided in most of his classes. They are evidently facts of 
like nature and deserve about equal weight ; but the position 
of the stamens, instead of ranking high amongst natural sub- 
divisions as assumed by Jussieu, ought to occupy a very 
subordinate place. To illustrate this I may state, that there 
are plants amongst Amaryllideae, which but for the difference 
of having the perianth and stamens superior instead of in- 
ferior to the ovary, would be almost identical with others 
amongst Asphodeleae; for instance, if the scentless Alliums 
of the latter, (an occidental race forming, I believe, a separate 
genus, which might be called Pseudoscordum), had the 
ovary inferior, it would require nice discrimination to sepa- 
rate them from Lapiedra of the former, and their general 
aspect would touch very close upon Strumaria, though there 
would be points of distinction; yet plants which are separated 
from actual identity of genus by little more than that feature 
and some difference of seed, which escape the observation 
of an uninstructed observer, are not merely removed by Dr. 
Lindley’s arrangement into a different alliance, but two 
whole groups, ten alliances, and twenty-five natural orders 
intervene between them. Such incongruities are found in 
every mode by which the orders have been arranged. In 
Sweet’s Hortus Britannicus we find the water-lilies close to 
