SEMPERYIYUM. 
337 
tt6. S. arboreum L. u Saiao ” Brot. 
Smooth, not viscid, fruticose and even tree-like, with stout 
suhumbellate erect straight very smooth and even regularly 
quincuncially scarred branches ; 1. in disk-like terminal rosettes 
narrow lanceolate-cuneate or acutely spathulate, elongate and 
much attenuated downwards, shortly and minutely but closely ci- 
liate upwards; cymes in terminal compact close half-oval or ob- 
long obtuse thyrsoidal heads or panicles ; pet. 9-12 linear-lanceo- 
late ; hypog. glands shortly and broadly wedgeshaped notched or 
bilobed and minutely toothed. — Desf. i. 389; Brot. ii. 378; Spr. 
ii. 468; DC. PI. Grass, t. 125 and 125* ; DC. iii. 411 ; Bot. Reg. 
2. t. 99 ; Haw. Syn. 164, Rev. 63 ; FI. Gr. t. 473 (not Buch’s 
Mad. List). S', paniculatum Sol. MS. in BH. P — Shr. per. Mad. 
reg. 1, cult, or half-naturalized, r. On the tops of walls and 
roofs of houses here and there in Funchal, growing spontane- 
ously, but merely as a straggler or outcast from gardens, and 
certainly not indig. in Mad. Apr. -June. — A small sparingly 
but proliferously branched erect dwarf tree-like shr., 2 or 3 ft. 
high, with a distinct st. and stout stiff naked very round and even 
upright branches as thick as the little finger, fleshy and some- 
what thickened upwards, pale brown very smooth but regularly 
scarred quincuncially below the terminal rosettes or disks of fiat 
and rather thin but fleshy very shining bright gr. 1., which are 
2-3 1 in. long and only 6-8 lines broad towards their end. FI. 
bright clear y. in thick close terminal oval or oblong thyrse-like 
bunches 6-12 in. long, at first short and pyramidal but lengthen- 
ing out as the fl. expand. Branches of thyrse, pedic., br. and cal. 
very minutely glanduloso- or furfuraceo-puberulous. Upper 1. 
and br. lanceolate deciduous. Sep. ovate-lanceolate. Pet. mostly 
12 ; stam. twice as many ; ov. and styles 9 or 10. Hypog. glands 
short truncate broadly wedgeshaped notched or bilobed, the 
lobes very minutely eroso-denticulate. 
A spec, in BH., marked “ Semperv. cirboreum Linn. Sp. PI. 
664 Madera,’ 1 collected by Banks and Solander, is assuredly not 
S. arboreum L., but a vigorous young pi., first beginning to 
branch before flowering, of S. divaricatum (Ait.). This was 
however doubtless the sole original authority for S', arboreum 
of both Solander’s and Buch’s Mad. Lists. 
Though unrecorded as indigenous in the Phytogr. Can., and 
merely referred to by Webb as a northern precursor of the M<\- 
caronesian sp. (WB. i. 185), I have met with Se?np. arboreum L. 
abundantly and apparently quite wild in two or three islands of 
the Can. Archipelago : viz. Tenerife up the Barranco de Mar- 
tianez about a mile above the Montana del Oreo near Orotava, 
