Hypericum canariense (Nomenclature)
Shrub or tree 1-4 m tall, erect, bushy, with branches erect or ascending. Stems green to pale reddish brown, 4-lined when young, soon 2-lined, eventually terete, internodes shorter than leaves; bark becoming whitish then pale grey. Leaves sessile; lamina 20-70 x 5-15 mm, narrowly elliptic to narrowly elliptic-oblong, the upper often broader, plane, paler beneath with midrib prominent, not glaucous, chartaceous, deciduous shortly before and during growth of new shoots; apex acute to apiculate-obtuse or rarely rounded, base narrowly cuneate to subangustate; venation: c. 8-1 2 laterals forming looped intramarginal vein, sometimes with subsidiary laterals, densely reticulate towards margins, tertiary venation dense and obscure; laminar glands dense. Inflorescence up to c. 30-flowered from up to 5 nodes, sometimes with flowering branches from up to 7 lower nodes immediately below or separated by sterile zone, the whole broadly rounded-pyramidal to broadly cylindric; pedicels 4-10 mm; bracteoles reduced foliar to triangular-subulate. Flowers 20-25(-37) mm in diam.; buds narrowly ovoid to narrowly ellip- soid, acute to subacuminate. Sepals 3-4.5 x 1-2.2 mm, unequal, varying from lanceolate, acute and basally united to oblong or oblong-spathulate, rounded and c. 0.5 united, veins branched distally, laminar glands basally linear, distally punctiform. Petals bright 135 yellow, not tinged red, 12-17 x 5-7 mm, c. 4 x sepals, oblanceolate- unguiculate, cochleariform, rounded. Stamens 10-13 mm long, c. 0.7-0.8 x petals. Ovary 3-4 x 1 .5-2 mm, ellipsoid; styles 8-14 mm, 2.7-4.7 x ovary, basally separated and widely spreading-incurved. Capsule (9-)10-12 x 7-8 mm, pyramidal-ovoid to ovoid-ellipsoid, truncate to retuse, with horn-like persistent style bases, exceeding sepals. Seeds yellowish brown, 1.5-2 mm long; testa linear-reticulate to linear-foveolate (cf. Reynaud, 1991, f. 1, 3-6). 2n = 40 (Larsen, 1962; Borgen, 1969; Reynaud, 1986; Dalgaard, 1991).
Open rocky slopes, cliffs and ravines, disturbed ground, upper part of litoral zone, relict Laurus forest; (20-)180-900(-1200) m.
Canary Islands (all western islands and possibly Fuerteventura), Madeira. The only author to record its presence in Fuerteventura is Voggenreiter (1974: 688, map). Naturalised in the Hawaiian Islands (Maui) and southern California.
H. canariense is widely isolated from its nearest relatives in sect. 1. Campylosporus on the African mainland, both morphologically and geographically. Its inflorescence is most similar to that of H. roeperianum, which occurs in West Africa and has densely reticulate leaf venation; and for those reasons I at first regarded it as the nearest relative of H. canariense (Robson, 1981: 68). Hagemann (1989: 242) has pointed out, however, that the growth form of H. canariense is much nearer that of the mainly East African H. revolutum, which also occurs in the Cameroon mountains and Fernando Poo. I agree and now regard the Canary Island plant as most nearly related to the broader-leaved form of H. revolutum in Ethiopia. H. canariense differs from it inter alia by its broader leaves, more branched inflorescence, smaller flowers with relatively small, pale-gland-fringed sepals, narrower and not orange-tinged petals, fewer stamens (the fascicles grouped 2+2+1), trimerous ovary with relatively longer, spreading and basally distinct styles, and relatively narrower capsule with linear-foveolate rather than linear-reticulate seeds.
The type specimen of H. canariense has broad, rounded sepals, whereas in that of H. floribundum they are narrow and acute. Spach (1836, b), immediately followed by Webb & Berthelot (1836), recognised an intermediate state; and, as was his wont, regarded the difference between this group and the rest of Hypericum as worthy of generic rank. Thus, in order of increasing sepalline acuity, Spach and Webb & Berthelot, between them, described Webbia canariensis, W. platysepala, W. heterophylla and W. floribunda. Later workers found W. platysepala to be indistinguishable from W. canariensis, and the intermediate 'species' W. heterophylla was also soon ignored; so H. canariense and H. floribundum assumed a rather spurious distinctness. This view was encouraged by the geographical distribution of the extremes: H. canariense in Tenerife, Gomera and Hierro, H. floribundum in Gran Canaria, La Palma and Madeira. Spach's intermediates do exist, however; and because of their existence and rather irregular distribution, it is not possible to recognise more than one variable species, H. canariense L. [But see Part 9: 32. Insert discussion in later paper]