SEHPERVIVUM. 
331 
slender pedic. large and numerous in proportion to tlie size of the 
pi., about \ in. in diam., full golden-y., yet somewhat paler 
though certainly rather larger than in 8. dumosum, forming a 
conspicuous handsome more or less corymbose leafy panicle. 
Pet. mostly 8, often 7 or 9, sometimes 6, lanceolate, sometimes 
in the first-opening fl. ovate, finely hair-pointed spreading or 
recurved, 2^ lines or 5 mill, long, 1-1^ line broad. Ilypog. 
glands orange mostly somewhat distinctly bipartite or bilobed, 
one lobe or the other unequally 2-3- or even 4-fid, rarely simple, 
perhaps more clavate than palmate i. e. higher or longer than 
broad, yet flattened with a distinct narrow-oblong rather than 
wedge-shaped stalk. Stam. 12-18 much shorter than the pet. 
Styles and ov. mostly as many as and y. like the pet., smooth 
and shining. — On immersion in hot water this sp. gives out a pe- 
culiar fishy disagreeable smell a little like scalded codlin apples. 
S. villosum Haw., Buch Canar. and DC. Prodr. iii. 411, is sup- 
posed to be S. ( Aichryson ) radicescens Webb Phyt. Can. i. 183. 
And S. villosum Lindl. Bot. Beg. 18. t. 1553 is S. (. JEonium ) 
Lindleyi Webb Phyt. Can. i. 189. t. 33. 
S. villosum Ait. has never been found in the Canaries, unless 
the imperfectly known S. pygmceum Chr. Sm. ( Aichryson WB. 
i. 184) may be supposed to be a starved or depauperated state 
of it. But a very similar and closely allied sp., intermediate 
between the true Mad. pi. and the Can. S. punctatum Chr. Sm., 
represents it plentifully in Gomera (S. subvillosum Lowe), dif- 
fering prima facie in its gr. merely subpubescent not hoary vil- 
lous aspect. It is Bourgeau’s “PI. Can. no. 443, Aiclir. villo- 
sum Webb, Ins. Gomera, Degollada de San Sebastian Apr. 1845,” 
and also probably his “ no. 1277, Aiclir. immaculatum Webb 
MSS. Teneriffse, Baxamar ; ex itin. 2 do 1855,” in BH. and HU. 
I found it abundantly on rocky banks up the valley of S. Se- 
bastian in Gomera, 2 or 3 miles above the town, in Febr. 1858 
and Apr. 1861. It seems to have been mixed up by Webb 
formerly with 8. punctatum, to which it approaches indeed very 
closely. In the Banksian Herb. I have also found two spec, 
of the Gomeran S. subvillosum on a sheet marked u 1 Ma- 
dera Fr. Masson 1776.” The sheet contains 5 spec. ; the three 
middle ones being true S. villosum Ait., and the 2 outer on the 
right and left S. subvillosum. Thus the latter may possibly 
prove to be also a Mad. pi. ; though from another instance of 
probably similar confusion of Mad. and Can. spec, in the case of 
