Hypericum glandulosum (Nomenclature)
Shrub 0.25-c. 1.4 m tall, straggling to spreading below, with branches rigid above, the whole plant glabrous. Stems pale yellowish to reddish, shallowly 4-lined but not ancipitous when young, soon terete; internodes shorter than leaves; cortex exfoliating in strips; bark straw-brown, finely striate. Leaves sessile; lamina (20-)30-60 x (6-)10-20 mm, narrowly elliptic to elliptic-oblanceolate, some- what paler beneath, chartaceous, not glaucous, plane, spreading; apex acute or subapiculate to obtuse or rounded, margin entire or undulate to serrulate, with prominent but sessile round- or flat- topped black glands or proximally or rarely wholly entire with glands immersed, base narrowly cuneate to attenuate; venation: (2)3 pairs of laterals curved-ascending from lower 0.35-0.4 of midrib, tertiary reticulation obscure; laminar glands pale, dense, unequal, prominent; marginal and/or intramarginal glands black, dense. Inflorescence c. 10-45-flowered from up to 3 nodes, curved-corymbiform to hemispherical, dense; pedicels 3-4.5 mm; bracteoles (and bracts at uppermost 2 nodes) reduced, narrowly elliptic, with prominent marginal black glands. Flowers c. 15-20 mm in diam.; buds narrowly ellipsoid, acute. Sepals 4.5-7 x 1.5-2(-2.5) mm, slightly unequal, free, elliptic-oblong to elliptic or (the smaller) lanceolate-elliptic, acute, with sessile marginal glands or distally glandular-denticulate; veins 5, branching; laminar glands pale or a few black, linear or interrupted; marginal glands black, flat-topped, prominent, dense. Petals rather pale dull yellow, tinged or lined bright red dorsally, 10-14 x 3-4 mm, c. 2 x sepals, narrowly elliptic-oblong, acute (i.e. apiculus acute, apical); laminar glands pale, striiform to punctiform; marginal glands all black or some pale, all prominent or some immersed distally. Stamens c. 25-30, longest 7-9 mm, c. 0.7 x petals; anther gland amber. Ovary 2.5-3 x 1.5 mm, ellipsoid; styles 5-6 mm, 2 x ovary, spreading-incurved. Capsule 4-5.5 x 2.5-4 mm, ovoid, shorter than sepals, enclosed by petals twisting together. Seeds pale yellow-brown, 0.5-0.6 mm long; testa very shallowly linear-foveolate, almost smooth.
2n=18 (van Loon & de Jong, 1978; Reynaud, 1986), 40 (Dalgaard, 1986).
Open rocky hillsides or cliffs in Laurus forest or among Erica thickets; 300-900 m or sometimes down to maritime region in barrancos (Canary Is.), 200-700 m (Madeira).
Canary Islands (Tenerife, Gomera, La Palma, Gran Canaria), Madeira.
H. glandulosum has as its nearest relative in sect. Campylosporus the east African H. quartinianum, its flower and fruit differing essentially from those of the latter in having '3' stamen fascicles, a 3-merous ovary with spreading styles, fading petals that twist around the developing ovary, and almost smooth seeds. The inflorescence branching is initially dichasial, as it often is in the closely related H. roeperianum, which has a more widespread African distribution than H. quartinianum (see Robson, 1985: 194-201).
H. glandulosum is rather distantly related to the other Canary Island shrub in sect. Adenosepalum, H. reflexum, and differs in its glandular-margined leaves (which have laxer venation and are broadest about the middle, not below), and glandular-ciliate sepals and bracteoles. In addition, H. glandulosum includes tetraploids on the base 10 (2n=40) and diploids on the base 9 (2n=l 8), whereas H. reflexum is apparently diploid only (2n=18). H. glandulosum is usually completely glabrous, whereas H. reflexum normally has a pubescent stem. Plants intermediate in leaf form between the elliptic typical of H. glandulosum and the triangular-lanceolate of typical H. reflexum have been found. Those like H. reflexum but with ovoid or ovoid-lanceolate leaves have been named H. reflexum var. myrtillifolium Bornm., whereas those similar to H. glandulosum in leaf shape but with villous-tomentose stems and villous-pilose leaves have been described as a distinct species, H. joerstadii Lid.
It should be noted that: i) H. glandulosum and H. reflexum are typically not only distinct but very different in stem, leaf and sepal characters; ii) H. glandulosum in Madeira (in the absence of H. reflexum) shows no tendency to vary towards that species; and iii) each species has apparently given rise to quite distinct species groups on the African mainland. For these reasons it seems best to regard H. glandulosum and H. reflexum as 'good' species and all or most of the variation observed in the Canaries as due to hybridisation between them.
Dryander (the actual author of Hortus kewensis) cites only ‘ Introduced 1777, Fr. Masson. but there is no Masson specimen in the BM collection and Masson s introduction may have been of living material only. At any rate, Dryander's description was almost certainly derived from the Banks & Solander specimen, which therefore should be regarded as the holotype.