PRELIMINARY TREATISE. 
21 
tually exclude some race, to which his attention has been 
more closely directed, let him not blame me lor the want of 
more accurate knowledge, which I do not pretend to have 
acquired, but let him substitute the term which his own more 
perfect investigation shall have ascertained to be more effec- 
tual for their exclusion. 
The monocotyledonous plants were divided by Jussieu into 
three classes: — 1. Having the stamens inserted above the 
germen or future seed-vessel. 2. Inserted round it. 3. In- 
serted below it. This is a distinction of easy access ; but to 
judge what rank it should hold of itself, we must see whe- 
ther all the plants in each of those classes are more akin to 
each other, than to any which are not included in it ; and it 
is evident that Asphodelese, Hemerocallideae, and Melantha- 
cese, are more nearly allied to Amaryllideae in another class, 
than to Palms and Rushes in their own. It is therefore bad 
on the face of it as a natural division, that is, a limitation 
of natural groups ; it is as much an artificial division as the 
number of stamens, nor is it more convenient than the 
Linnaean arrangements, though it might be combined with 
them. Another mode of classification suggested itself from 
the structure of seeds, and Dr. Brown founded the charac- 
ters of several orders mainly on the colour, hardness, and 
separability of their external integument, and the direction 
of the embryo. These characters were assumed on a partial 
view of the seeds of some of the plants they were intended 
to include, or at least of those which they are now held to 
comprehend, and the result is, that they are inconsistent 
with the fact. Hardness and inseparability, softness and 
separability, of the seed-coat occur in plants so closely 
allied, that although I consider it a proof of generic differ- 
ence, their generic distinction has not yet been universally 
admitted by botanists ; as, for instance, Pancratium and 
Hymenocallis. I have already noticed the seeming vague- 
ness of the radicle in some plants where Dr. Brown has 
assumed its direction to be certain, and if correct, it is too 
secret and concealed a feature for general application. The 
hardness and softness of the integument is so uncertain, that 
it is found soft in plants in which, according to Dr. Brown’s 
character, it should be crustaceous and fragile. It is there- 
fore evident, that the hardness and separability of the seed- 
coat, and the direction of the embryo, furnish no satisfactory 
characters for the definition of natural groups amongst mo- 
nocotyledonous plants. It is certainly possible that the in- 
