4.42 ON THE GENUS RUBUS. 
sends out branches along its whole length; those nearest the 
original root are generally barren, like itself, but these are fol- 
lowed by erect flowering branches from one to two or even three 
feet long, but gradually becoming shorter as we approach the end 
of the shoot. The first and the last panicles are apt to be im- 
perfect ; those which best characterize the plant rise from the 
middle of the shoot. ‘ 
Though the tendency to arch downwards is common to all the 
Brambles, yet it is so much more strongly marked im some than 
in others that-one subdivision of the tribe has been called erect, 
and is believed never to curve so much as again to reach the 
ground, much less to take new root there. This peculiarity is 
important, but unfortunately cannot be shown in the herbarium ; 
even with the growing plant it requires some caution. The 
flowering branches of the second year, arising from the assurgent 
shoot, are erect, and this is also the case with the lower branches, 
which do not produce flowers. When therefore this branch sur- 
vives the year, and brings forth flowers on the third year, the 
plant has an appearance of being more erect than it truly is, and 
may be erroneously attributed to the division with upright shoots. 
The shoots of the year are sometimes round, and sometimes 
marked more or less decidedly by five angles. This character 
varies a good deal in the same species, and even in the same indi- 
vidual, the angular form of the young shoots filling up mto a 
cylinder as it gets older. Yet some species show much more 
strongly than others a tendency to the prismatic form, and some 
are entirely without it; it cannot therefore be neglected im the 
formation of specific characters. 
The leaves of the barren shoots are received as the typical form 
of those parts of the plant; they are quinate or ternate. In the 
former case very often pedate ; not exactly however with the pre- 
cise arrangement usually found where that term is applied ; for in 
the pedate leaf we find the additional leaflet or leaflets rismg on 
the inside of the first lateral leafstalks, but in Rubus it is always 
external. The central stalk 1s much longer than those on each 
side of it, and these again invariably longer than those of the 
outer leaflets. The typical shape is that of the middle leafit. 
No Rubus has a septeno-digitate leaf, but when there are seven 
leafits it is always because the middle leaflet is divided into 
three, 
