XXil INTRODUCTION. 
the rootshoots, where they are quickly changed into 
aculei by losing their gland. In general they are de- 
ciduous after the first year. On their presence on the 
branches depend some of my most natural divisions. 
Spinosissime are divided from Canine by that character 
among others. Turbinata is essentially different from 
its nearest allies in the want of them. R. hystria is 
easily known from the rest of its division by their pre- 
sence on its branchlets. ‘They cause nevertheless some 
unaccountable anomalies. R. canina produces them 
now and then, in a division where they are not other- 
wise found. Such a disposition also exists in R. arven- 
sis hybrida; nor is R. rubiginosa by any means free 
from them. These however ought to be considered 
unimportant exceptions, which cannot materially af- 
fect the general utility of setze in characterizing species. 
It is remarkable that an organ, which on the branches 
is of so much value, should be of all others the most 
variable on the fruit and peduncle. R. rubiginosa and 
tomentosa produce setigerous and naked fruit ; in some 
instances indiscriminately on the same bush. R. ca- 
nina is not unfrequently furnished with setz on its 
peduncle, rarely on the fruit. Spinosissima, carolina 
and others, are equally disposed to vary with setige- 
rous and naked flowerstalks and calyx. Yet there are 
species which are not subject to such inconstancy in 
the surface of their fruit. For instance R. involuta, 
Sabini, villosa, levigata, sinica, &c. have never been 
seen without setigerous heps. 
By the form of the prickles Canine are tolerably 
distinguished from Villose; and their inequality di- 
vides Rubiginose from the former. Their presence or 
absence on the petioles is much too variable to be em- 
