458 MR. G. BENTHAM ON COMPOSITAE. 
Gynoxys, or retained by all botanists in Senecio itself. These ap- 
pendages, however, are very short and often obscure, and scarcely 
more than rounded instead of truncate tips, or, as observed by 
Weddell, deceptions caused by the inequalities in the length of 
the hairs which form the terminal tuft. I had also long retained 
the North-Asiatie genus Ligularia, in which the collecting-hairs 
or papille descend more or less down the back of the style- 
branches, accompanied by a peculiar habit, with large subracemose 
radiating capitula. But here, again, I had ultimately to abandon 
the separation; sometimes the characters, sometimes the habit 
and geographical relations were at fault; and at present we are 
obliged to follow De Candolle in making our primary divisions of 
the genus purely geographical, and in each country subdivide them 
according to characters which have locally acquired relative im- 
portance. I think, however, that if any experienced monogra- 
phist were carefully to study the 900 odd species of Senecio, and 
especially to compare the various forms the ripe achenes assume, 
with the characters derivable from the styles, the anthers, the 
vegetative organs, and the apparent geographical origins, he might 
succeed in bringing out sectional combinations which have escaped 
me, and might even reestablish as independent genera some of 
those Cacalioid or Ligularioid groups which in the present state 
of our knowledge I have felt compelled to unite with Senecio. 
A. number of small genera, more or less divergent from Senecio, 
have a much more local character. One only, Erechthites, a genus 
of about a dozen species in a great measure tropical, distin- 
guished chiefly by the filiform female florets, has a wide range. 
Its great centre is South America; but it is found northwards 
as far as Carolina; and in Australia and New Zealand it has 
established several endemic species. In tropical Asia the single 
species observed is probably a recent introduction from America. 
In Africa it is, I believe, unknown; the Euseneciones which 
there assume the above-mentioned main character of Erechthites 
are connected with Senecio through different channels. 
The other American genera closely connected with Senecio are 
Culcitium, about 14 species, and Gynoxys (from which, following 
Weddell, we exclude the alternate-leaved scandent species), about 
12 species, both genera Andine. Culcitium is very near Senecio, 
differing from some of the genuine species of that genus from the 
same country only in the involucre; and even in that respect 
there are intermediate species which have been alternately referred 
