PRELIMINARY TREATISE. 
29 
manent varieties ; and that in some genera even cultivated 
varieties preserve themselves distinct in their generations 
more easily than natural species in others when approximated. 
2 1 . That in some genera intermediate diversities from different 
localities so confound the limits of species, that it is waste of 
words to argue whether a plant is a species or a permanent 
local variety; though it is a matter of convenience to assign 
some specific features, and to arrange the subordinate forms 
as varieties, but that the distinction of closely allied speci6s 
is of very trivial importance ; botanical species being merely 
the long established varieties of genus or kind, perhaps con- 
sequent on the dispersion of seeds by the deluge and refera- 
ble to the unknown period when the different races of man- 
kind assumed their peculiar appearance. 22. That the dis- 
crimination of an immutable generic feature is the discrimi- 
nation of a fact, the limits of species and varieties are the 
offspring of human opinion. 23. That a mixed offspring 
from the mutual contact of any two plants, if not absolutely 
conclusive as to the identity of their origin, gives strong rea- 
son to presume it, and to doubt the validity of the features by 
which they may have been distinguished. 24. That the 
fecundity of such mixed offspring establishes the identity of 
kind. 
In framing the character of an order or genus, for the 
sake of perspicuity and correctness no feature should pro- 
perly be included but such as are necessary to distinguish 
the order or genus from all others, and from which a depar- 
ture would decidedly exclude a plant that should contravene 
it ; and all important points that appear to coincide with the 
character should be added as observations subsidiary to it ; 
nor will the system be perfected till each genus is separated 
from its nearest kin by a single primary and immutable 
feature, its subordinate peculiarities being subjoined. Per- 
fection of the system cannot, however, be effected without a 
complete knowledge of the peculiarities of all the indivi- 
duals which each order contains, and our ignorance of im- 
portant points, even concerning those which are considered 
as having been described, presents an obstacle which can 
only be surmounted by time and assiduity. I doubt the 
correctness of any generic character which rests upon cumu- 
lative features, of which no one has sufficient weight by 
itself to distinguish and uphold it ; and I think that no 
botanist has sufficiently considered what are the features 
