480 
47. COMPOSITACEiE. 
Bot. i. 35 (not Linn., or Lam., or Willd.). Gn. tomentosum 
Bowd. Exc. in Mad. p. 63 (not Hoffrn.). Antennaria leuco- 
phyttci Beiclib. in Hull’s List in Hook. J. of Bot. (2nd ser.) i. 19. 
— Ski*, per. Mad. reg*. 1. cc. Bare and rugged sea-rocks and 
cliffs everywhere all round the coast of Mad., hut not found in 
either PS. or the Desertas. The Loo Rock (Illieo) at Funchal 
is almost clothed with it. It prefers the hardest and blackest 
basalt entirely devoid of other vegetation, springing out of cre- 
vices or fissures and contrasting conspicuously by its whiteness 
with its native rocks. It is never found away from the imme- 
diate vicinity of the sea. At most seasons, but chiefly June- 
Nov. — A low thickly bushy proliferously branched shr. 1-2 ft. 
high with stout erect greyish or almost snow-w. tomentose 
corymbose branches naked below, thickly leafy upwards, and 
though stiff hard and woody somewhat fragile. L. attenuatelv 
stalked crowded 1^-lf in. long (including the petiole), about 
T 7 in. broad, very entire subacute or often a little obtuse, of a 
firm soft thickish leathery substance and covered on both sides 
with a close short dense snow-w. tomentum. Cyme single 
and simple terminal snbsessile or only a little overtopping the 
1., 1-2 in. in diam., the divisions and ped. very short stout w. 
tomentose without 1. or bractlets, the whole in fl. dull-y. FI. 
like the whole pi. slightly odoriferous. Heads 3-4 lines long, 
2|-3 broad, shortly turbinate and slightly ventricose some- 
what like the fr. of a Medindla, shining golden-brown or 
greenish ; the lower scales continued down the ped., regu- 
larly imbricate and gradually becoming larger upwards. Flo- 
rets smooth hoary or whitish in the bud, then dull y., all 5- 
cleft and fertile, the 2 outer rows female very slender or fili- 
form, the rest all perfect and altogether much larger. Recept. 
convex. Pappus in all the fl. rather longer than the fl., smooth 
downwards or very shortly and minutely puberulous or rough ; 
bearded or fringed towards the tip with longer hairs. Ach. 
subturbinately oblong dark chestnut-brown angular and strio- 
late minutely puberulous or rough with glittering erect points. 
Judging from the description and from a single head of II. 
hemisphcericum 1)C. (Prodr. vi. 181) most kindly communicated 
by Prof. Alphonse Be Candolle from his father’s original spec., 
I strongly suspect that it will prove to be not really distinct 
from II. obconicum , in which the 1. have very frequently only the 
middle nerve discernible (the shape also varying from subacute to 
subobtuse) and the heads become not unfrequently more rounded 
at the base or hemispherical as they advance after flowering. 
The original head of II. hemispluericum before me is indeed 
rather obconical than hemispherical, and in its colour, its short 
