952 THE JOURNAL OF BOTANY 
really, but also apparently, in comparison with the length of the 
fruit, small fruits often having the effect of making the Sapo 
look longer than they actually are. We believe length to be of 
ch importance, though aa emt is tolerated by Yc in 
botanists. In many cases it is a primary group charact aes 
example, one of the features by which the #. Sherardi geotiy 
distingnished from that of A. tomentosa is the short peduncles of 
R. s and shor af W nk the 
lngth of the peaenill lative is the ful y developed fruit is often 
a helpful characteristic, though actual measurement is better 
the sepals and their persistence on the ripening fruit, some also 
to their degree of pinnation, but not ish | 0 their clothing with 
hairs or glands. ‘Taking the = feature first, it is usual for sepals 
to be tomentose within, and the tomentum to extend round 
the edges so as to form a row or wide border on the back, and 
ess numerous gia ese on the e = _especi- 
ally on the tips of pina, are found i many species, and 
much stress has sometimes been laid on thei elr presence or absence 
as a specific ote cter by which two allied species may be distin. 
ju 
on edges 
presence or absence is very variable. Glandular development o 
the backs is of greater heedaher ss but care must be taken to dis- 
tinguish glands from a warty or scabrous surface, Throughout 
1 
so among the Donk except in the ssiorsee leaved 
members of the i: coriifolia 
The degree of pinnation, Sdiolibh liable to much variation in 
the same species, is important. In the section Pimpinellifolia, 
e 
sepals are the rule, while in the R. tomentella group and in most 
of that of f. tomentosa they are much divided. In the rest there 
be treated as a group rather than a specific character. 
The direction of the sepals on the maturing fruit, and still 
Q 
hough, gener speaking, a group rather than a_ specific 
character, the LE a of the sepals must not be too rigidly 
