56 ERICACEH (Guthrie & Bolus). [Hrica. 
Stellenbosch Div.; Lowrys Pass, Mund, 8! Caledon Div.; on mountains, 
Burchell, 7719 ! 8104! Zeyher, 3176! Drége, 7707a! Bolus, 5343 4 Schlechter, 
7583! George Div.; Keurbooms River, Burchell, 5130! Knysna Div. ; various 
localities, Burchell, 5197 ! 5854! Bolus, 2886! Humansdorp Div.; near Kromme 
River, Drége, 7707b! and cultivated specimens! Var. B: Caledon Div.? top of 
hills (Houw Hoek?), Niven, 17! Herd. Salisbury! Var. y: Div. ? Guthrie, 
3795! Var. 6: Tulbagh Div.; Witsen Berg, Burchell, 8652! Roode Zand, Niven, 
208! Herb. Salisbury (at Kew)! 
We adopt the older name of Linnzus the younger, and have little hesitation in 
grouping the several forms above described. The type and var. ‘y almost pass into 
each other by gradations ; var. 8 is more distinct, but it differs chiefly in its broader 
sepals, It appears to have been described by Salisbury, and figured by Andrews, 
from cultivated plants ; it is scarcely possible in a group where the sepals vary so 
greatly to rely upon them for specific differences. We have dissected eight 
gradations of form as indicated in the text above, without finding any other 
uniformly correlated differences of any weight. ‘The mere greater or less width 
of the corolla is of even less value. 
17. E. filipendula (Benth. in DC. Prodr. vii. 663); glabrous, 
1-2 ft. high; branches erect, subvirgate, bearing in their upper 
portion a somewhat lax pseudo-raceme of flowers; leaves 4-nate, 
erect-spreading, not densely crowded, linear, acute, somewhat round- 
backed, 3-5 lin. long; pedicels 3-4 lin. long ; bracts remote, nearly 
basal, small; sepals lanceolate or linear-lanceolate, keeled, acute, 
scarious, 13-2 lin. long, whitish; corolla ovoid to ovoid-urceolate, 
canary-yellow, 4 lin. long, mouth somewhat contracted, texture 
papery, glabrous, dry, finally brownish; limb short, erect; anthers 
included, oblong, bilobed at the base, prognathous, and sometimes 
sharp-pointed, about % lin. long, aristate ; awns nearly as long as the 
cells; ovary on a short stipe. 
Var. 8, major (Bolus); sepals broadly ovate; corolla tubular-urceolate to 
inflated-tubular, 5-9 lin. long; ovary somewhat long-stipitate. E. broadleyana, 
Benth. in DC. Prodr. vii. 637, not of Andr. ; 
Var, y, minor (Bolus) ; like the preceding but the corolla smaller, hardly 
3 lin. long, white, becoming pink near the apex; anthers slightly or scarcely 
prognathous at the base; ovary substipitate or contracted at the base. 
SourH Arrica;: without locality, Bowie! Var. 8: Thom, 1094! and cultivated 
specimen ! 
Coast RuGion: Bredasdorp Div.; hills near Elim and Ratel River, 300- 
400 ft., Guthrie, 3785! Bolus, 8450! Var. B: Bredasdorp Div. ; fairly abundant 
on the downs between Elim and Ratel River, 300-600 ft., Guthrie, 3786! Bolus, 
8452! Schlechter, 7618! 7726! and in MacOwan, Herb. Aust.-Afr., 1920! 
Var. y: Bredasdorp Div.; between Elim and Ratel River, 200-800 ft., Guthrie, 
3784! Bolus, 8451! Schlechter, 10472; here also apparently belongs Zeyher, 
1090, of which the locality is unknown to us, 
These three forms, unlike as they appear at first sight, can hardly be separated 
specifically. Beyond the size and colour of the corolla there is scarcely anything 
to distinguish them excepting a small difference in the length of the stipes of the 
ovary, a very variable character, where it exists, throughout the genus. They 
afford a fresh example of the cases, which a larger supply of material has brought 
to light, of the connection and gradation between supposed species with long- 
and others with short-flowered corollas which have hitherto been placed in 
different sections. In this instance the affinity is shown with § Hermes, and 
_ other cases occur in E, casta and E. coccinea, in this section ; besides several in 
_§ Evanthe, There is a considerable resemblance between our var. 8 and E. gilva, — 
and E, mammosa, and the almost identical shape of anthers in all three, as well — 
