240 MESSRS. T. A. SPRAGUE AND J. HUTCHINSON ON 
Rehmannii, laxiflora, Descampsii, Mastersii, heliocarpa, are exceedingly critical. 
It is possible that when more ample material comes to hand they will be found 
to form a complete chain in the order mentioned ; in the meantime the first 
three, which have narrow leaves, are grouped together under a single species, 
Welwitschii, and the last three, with broader leaves, under Masters. 
All the species of Lappula with the exception of T. Dekindtiana are charac- 
terized by а bur-like fruit. The fruit of 7. Dekindtiana outwardly resembles 
that of Lasiothriz, being covered with rather weak plumose bristles, but it is 
4-celled, with 2 seeds in each cell, just as in mierantha and setulosa, and the 
inflorescence is of the Lappula type. Observations in the field as to the 
dispersal of its fruits would be of great interest, as it appears to be the only 
species of Lappula in which the fruits are adapted for distribution by the wind. 
For practical purposes, three series of species may be distinguished in 
section Lappula. The Stellatæ have the prickles terminated by several 
spinules stellately arranged; the Geniculate by a single straight spinule 
inserted at an angle and thus forming a knee with the prickle; and the 
Uncinatæ by а hooked spinule. 
These series, though very useful for the purposes of a clavis, do not coincide 
with natural groups, but appear to represent three o£ the phylogenetie stages 
through whieh many species of Lappula have passed, the stellate arrange- 
ment being the most primitive and the uncinate the most highly modified. 
The following considerations are in favour of this hypothesis :— 
1. In the sections which are apparently the more primitive, namely 
Lepidocalyx, Porpa, and Lasiothrix, only the stellate and geniculate conditions 
are found. 
2. Some species appear to be in a state of transition. The most striking 
case is that of heterocarpa, in which some of the prickles may be of the stellate 
type, others geniculate and others uncinate. In flavescens the stellate and 
geniculate conditions occur, either on separate fruits or on the same fruit. 
In setulosa and most of the allied species the prickles are usually of the 
stellate type, but a few geniculate prickles commonly occur among the others. 
In tomentosa the prickles are usually geniculate, but those of Angola specimens 
are subuneinate. Finally, in trichocarpa (type) the prickles are uncinate, but 
a specimen collected in the Chari Region has some of the prickles stellate 
and others geniculate, none of them being uncinate. 
3. Certain closely allied species have prickles of different type. Thus 
abyssinica, with geniculate prickles, is allied to micrantha, buettneriacea, and 
dubia, which have stellate prickles; and Antunesii and delicatula with 
uncinate prickles are undoubtedly closely related to Dekindtiana and paradoxa 
respectively, in which the prickles are stellate. 
4. As to the sequence of the types of prickle, the reduction from stellate 
to geniculate is much more probable than the reverse, and it is highly 
improbable that the uncinate type, which is so efficient in bringing about the 
