954 THE JOURNAL OF BOTANY 
section by means of the stigmas being in a flattish or hemi- 
spherical head, not in a more or less elongate cone, as in the 
co 
62] 
a>) 
e 
ce 
fe) 
Ss 
ie) 
i} 
ct 
4 
° 
Lec 
Qu 
M 
Qu 
fa») 
mM 
fe) 
ss 
rs 
er 
mae 
a 
© 
io) 
eb 
ot 
a 
far) 
fio) 
co] 
2 
— 
a 
= 
=e 
mn 
cot 
2 
n 
4 
o>) 
2 
S 
=) 
of the flower, and except as an inefficient substitute for the shape 
of the fruit is of doubtful value. Generally speaking, the calyx- 
tube becomes more inflated as the fruit develops, so that a 
globose calyx-tube may safely be taken to indicate a globose fruit, 
but a very elongate calyx-tube is very unlikely to produce even a 
t 
subglobose fruit. 
The d 
has not been much studied in Britain, but 
value, also the development and especially the degree of elevation. 
of the disc or space between the styles and the stamens. 
example, the disc is remarkably conical throughout the section 
Stylos@, and is narrow in the RB. pomifera group. 
conclusion, we would remind students of the genus that no 
single character taken by itself must be held to be conclusive, so 
that analytical keys must be used with caution. Thus varieties 
of the subsections Villose and Rubiginose are known which have 
smooth peduncles, or, in ormer group, showing glabrous 
leaflets, or eglandular ones in the latter. The amount of hair, 
