Hebenstreitia. | SELAGINER (Rolfe). 103 
‘“Sourn Arnica: without precise locality, Forster! Forsyth! Hutton! 
Masson ! 
Coast Reeion: Clanwilliam Div. ; Modder Fontein, Whitehead / Riversdale 
Div. ; Riversdale, Rust, 89! Mossel Bay Div.; on sandy hills near Fish Bay, 
below 200 ft., Drége! Kieiu Berg, 800 ft., Galpin, 4398! Knysna Div. ; on: 
sand-hills at the western end of Groene Valley, Burchell, 5659! Uniondale Div. ; 
Long Kloof at Wagenbooms River, Bolus, 2427! Uitenhage Div.; in sandy 
places near the Zwartskop River, Zeyher, 958! 3583! Port Elizabeth Div. ; Algoa 
Bay, Cooper, 3076! Port Elizabeth, Cape Herb. ! Albany Div.; near Grahams- 
town, Bolton! Burke / and without precise locality, Cooper, 1558! Atherstone, 
91! Miss Bowker! Williamson! Stockenstrom Div.; Kat Berg, 3000-5000 ft., 
Drége! Stutterheim Div, ; Stutterheim, Cape Herb.! Bathurst Div.; at Riet 
Fontein, between Theopolis and Port Alfred, Burchell, 3937! British Kaffraria, 
Cooper, 11%! 
~Crentrat Region: Aliwal. North or Wodehouse Div., Zeyher ! 
Pali hoe ie Region; Vanrhyusdorp Div. ; near Mieren Kasteel, below 100 ft., 
rege ! 
. Katanart Reeion: Bechuanaland; Eastern Bamanguato Territory ? between 
Shoshong and Molopolole, Holub! Transvaal; near Lydenberg, Atherstone ! 
Var. 8, Transvaal; without precise lvcality, Mrs. Stainbank in Wood Herb., 
3645! ° 
Eastern Recion: Natal; on a sandy flat near Durban, 50 ft., Wood, 1720!- 
100 ft., Wood, 330! . 
A very variable species, which seems to include all the different forms above 
cited, though Choisy keeps H. aurea, Andr., H. scabra, Thunb., and H. virgata, 
E. Mever, distinct, while making H. tenwifolia, Schrad., synonymous with H.: 
integrifolia, Linn., which he enumerates as H. dentata, Linn., var. B, integrifolia. 
He also refers the Abyssinian specimens to this, in which I cannot follow him. 
The species was based on H. foliis integerrimis, Linn. Hort. Cliff. 497, and the 
characters given us separating it from H. dentata, Linn., point distinctly to the 
present one. No specimen appears to have been preserved. It is easily 
separated by its nearly filiform leaves and long narrow flower-spikes. From this 
there seems to be a regular gradation down to the more stunted forms repre= 
sented by H. virgata, E. Meyer and H. scabra, Thunb. 
10. H. Watsoni (Rolfe) ; perennial (?), somewhat branched, 
over 1 ft. high; branches glabrous or very minutely puberulous ; 
leaves numerous, filiform-linear, subacute, entire or minutely denticu- 
late, minutely canescent, }—2} in. long ; spikes elongate (young in the 
specimens seen) ; bracts ovate-lanceolate, long acuminate, 2—4 lin, 
long; calyx ovate-oblong, obtuse, 1j-l)_ lin. long, with 2 very 
slender or obscure nerves; corolla white, with or without an orange 
blotch in the centre of the limb, 4-5 lin. long; tube slender; lobes 
oblong, inner pair longer and narrower than the outer. 
Coast Recion: East London Div.; East London, on the sea coast, 50 ft., 
Galpin, 1569! and a cultivated specimen from the same locality, Watson / 
Nearly allied to H. integrifolia, Linn., but differing in its more lanceolate and 
acuminate bracts. It is apparently a coast plant. 
- 11. H. Rehmanni (Rolfe); annual (?), much branched, more or 
less diffuse, 6-10 in. or more high ; branches rather slender, puberu- 
lous ; leaves numerous, narrowly lincar, subobtuse, entire, 8-12 lin. 
long; bracts spreading or somewhat reflexed, ovate or ovate-oblong, 
shortly acumiuate, 2-24 lin. long ;. calyx ovate-oblong, obtuse, 1¢ lin. 
long, with 2 herbaceous nerves ; corolla 33-4} lin. long; . tube 
' 
