Hypericum sect. Thornea (Breedlove & McClintock) N.Robson

Robson, Norman K. B., 2016, And then came molecular phylogenetics-Reactions to a monographic study of Hypericum (Hypericaceae), Phytotaxa 255 (3), pp. 181-198 : 192-193

Hypericum sect. Thornea (Breedlove & McClintock) N.Robson , comb. & stat. nov.

Basionym:— Thornea Breedlove & McClintock (1976: 369) . Type: Thornea matudae (Lundell) Breedlove & McClintock (≡ Hypericum matudae Lundell ).

Morphologically and geographically, sect. 1. Campylosporus seems to be the ‘basal’ section of Hypericum , i.e. the section most closely related to all other parts of the Clusioid clade and the section to which all other sections of the genus are related directly or indirectly. But none of the molecular or cladal treatments gives it this position. Among the Old World sections with regularly pentamerous flowers it is linked morphologically to sect. 3. Ascyreia and most of the rest of the Old World group, as well as to the whole of the New World group, through the plesiomorphic species without dark glands, and to 26. Humifusoideum through H. madagascariense , which sometimes has dark glands. Its most plesiomorphic species, in my view ( H. bequaertii De Wildeman (1920 : B4) and H. keniense Schweinfurth (1892: 15) , which have the ‘parallel’ leaf venation also found in the ‘basal’ New World species H. terrae-firmae Sprague & Riley (1924: 12) . It seems possible that this basic type of venation indicates a relationship between these large-flowered Hypericum species and the large-flowered genera of Calophyllaceae (e.g. Kayea Wall. and Mesua L.).

The grouping together of all or most of the herbs with 3+3 stamen fascicles, thereby resurrecting Keller’s (1925) section ‘ Euhypericum ’ (Nürk & Blattner 2011, a superfluous name as it includes the type of the genus and thus subsequently named ‘core Hypericum ’ in Nürk et al. 2013), is a puzzle. This grouping produces a hotch-potch of species that share nothing except the above two characters. Having shown that these species belong to three groups of sections, I have no explanation to offer for this ‘core Hypericum ’ grouping beyond suggesting that it could be due to the operation of a gene that selects the 3+3 fascicle floral structure.

Nürk et al. (2013) and Meseguer et al. (2013) both, in effect, have queried the distinctness of sects. Brathys and Trigynobrathys , having found some species that I placed in Trigynobrathys among their Brathys species. There are two main points at issue: (i) Did these sections arise separately directly from African Campylosporus stock or did Trigynobrathys originate from one of the basal species of Brathys ? If the latter hypothesis is correct, then these taxa should be treated as subsections of Brathys . (ii) Do some of the species that I allocated to Trigynobrathys really belong in Brathys ?

(i) The leaves of H. rigidum subsp. rigidum Saint-Hilaire (1825: 336 ; the ‘basic’ taxon of Trigynobrathys ) are narrowly elliptic to linear-oblong with 3–4 pairs of basal or near-basal veins, whereas only in subsect. Styphelioides , the basic subsection of Brathys ( H. terrae-firmae in Belize and H. styphelioides Richard, 1845: 237 , in Cuba) and sometimes in H. phellos Gleason (1929: 106 ; subsect. Phellotes) are there any ‘primitively basal’ leaf veins. They reappear in species with a broad leaf base, e.g. H. mutilum Linnaeus (1753: 787) .

The inflorescence in Brathys is single-flowered with what I have called pseudodichotomous branching (i.e. with paired single- to numerous-noded branches from the node below the flower); whereas the inflorescence in Trigynobrathys is dichasial (i.e. with a terminal flower and axillary branches), except in H. rigidum subsp. rigidum , where it is sometimes partly pseudodichotomous. It is clear, therefore, that H. terrae-firmae and H. rigidum are related; but the overlap in characters is relatively small and I have preferred to keep them in separate sections.

(ii) Most of the appearances in sect. Brathys of species that I have classified in Trigynobrathys can be treated as casual and unexplained (e.g. the African H. scioanum beside the African Humifusoideum species H. peplidifolium, Meseguer et al. 2013 ), especially as, in the following year ( Meseguer et al. 2014), H. peplidifolium appears in a group of ‘ Euhypericum ’ species.

However, one such grouping does deserve more detailed examination. Nürk et al. (2013) placed the North American H. denticulatum Walter (1788: 190) subsp. acutifolium (Elliott) Robson (1990: 66) in Brathys , whereas I classified it (along with the rest of the H. denticulatum group from eastern North America) in Trigynobrathys near the southeastern Brazilian H. rigidum subsp. sellowianum Robson (1990: 54) . The inflorescence of this group, when fully developed, is of axillary pairs of repeated dichasial branches, i.e. typical of Trigynobrathys and not pseudodichotomous as in Brathys .

(iii) It seems clear, then, that Brathys and Trigynobrathys are best treated as distinct sections, as in the monograph, with their respective contents as I have indicated.