INTRODUCTION. xx 
ployed at any time. Where their situation is infrasti- 
pulary, asin the greater part of Cinnamomee, they pre- 
sent an important character: but one of which we can 
judge only from a living plant; because most species 
with scattered prickles occasionally produce specimens 
in which they are placed beneath the stipule. The 
fruit of a few kinds is prickly and as it seems con- 
stantly so. 
Glands, which are perhaps better distinguished 
from setz by their scent than any thing else, are for 
the most part attached to the leaves on the under sur- 
face. They are employed to divide Rubiginose from 
Canine, myriacantha from spinosissima ; and, among 
other things, Brunoniit from moschata. R. pulverulenta 
is the only Rose having them on the upper surface of 
the leaves. The curious substance known by the name 
of moss, which makes its appearance on centifolia and 
rubiginosa, may be considered glands under another 
form. 
Pubescence on the branches, peduncles, or tube of 
the calyx is the only invariable character I have disco- 
vered in Roses. Distinctions drawn from it I have 
every reason to consider absolute. It is either persis- 
tent or deciduous. When persistent it becomes an im- 
portant criterion of a section, and characterizes Sim- 
plicifolia, Feroces, and Bracteate; the two latter divi- 
sions being formed by its absence from the fruit of the 
former. When it is deciduous it becomes of specific 
importance only; in that state it distinguishes R. abys- 
sinica from sempervirens, glutinosa from rubiginosa, 
and contributes to the separation of microcarpa from 
Banksice. Just the reverse is the case with pubescence 
upon the leaves. There it is usually of no consequence 
