ON THE DISTINCTION OF SPECIES. 621 
ing free from other herbage, this latter species forms dense 
cushion-like masses, such as we see formed by Saxifraga 
hypnoides or Sedum reflexum under the like circumstances. 
But if a starved or half-smothered plant of C. alpinum should 
be. compared with a healthy and free-growing plant of 
C. latifolium, doubtless the terms * branched” and “ simple” 
might be found applicable enough. 
Sixthly, characters should as little as possible be taken 
from conditions which are known to be very variable in other 
plants, and more particularly, if known to be inconstant in 
species nearly allied to those under consideration. Appen- 
‘dages of the cuticle, for example, as hairs and ramenta, are 
very uncertain. Bromus mollis and B. racemosus may be 
regarded as book species, separated almost solely by the 
presence or absence of pubescence, which varies much in 
these and allied species: even B. commutatus and B. secalinus, 
usually described as being glabrous, are not invariably 
so in nature. We have another example of inconstancy, in 
the awns of grasses. Lolium multiflorum has been distin- 
guished from L. perenne by its awned flowers. But the 
analogy of Lolium temulentum and L. arvense suggests the 
weakness of that character; and finding the other alleged 
distinctions unsound, I must now look upon L. multiflorum 
as a book species only, properly reduced to the natural spe- 
cles L. perenne. Again, the length of the internodes is a very 
changeable character in plants. Polygonum maritimum and 
P. Roberti (Lois. Br. Fl) are in part distinguished by the 
relative length of their stipules and internodes, acknowledged 
to be variable in these plants, and particularly variable in the 
allied species, P. aviculare. I suspect P. Roberti to be only 
a book species, which should be received rather as a variety 
of P. maritimum. So, also, I would say that a little more or 
less of membrane in the bracts and sepals of the genus 
Cerastium is a very insufficient character for specific distinc- 
tions. The half-membranous bracts, said to distinguish 
C. semidecandrum from C. atrovirens and other alleged species, 
would go to disunite plants which grow intermixed, and are in 
VOL. II 2v 
