Hirtella (Nomenclature)
Perennial herbs up to 0.8 m. tall, glabrous to pubescent or scabrid or rarely with branched glandular emergences (20. thymopsis), often glaucous, with stems erect or decumbent at the base from taproot, rarely rooting (7. hyssopifolium), usually with procumbent to ascending basal sterile shoots and axillary branches (often condensed), with dark (black) glands on petals, usually on sepals, sometimes on stems and leaves, but not on anthers or capsules. Stems 2-lined, usually glandiferous. Leaves opposite, decussate, sessile or shortly petiolate, free or very rarely united, persistent, subcoriaceous to herbaceous; lamina entire with venation pinnate or 1-nerved; laminar glands pale and rarely dark; intramarginal glands pale or very rarely some dark (11. apiculatum, 23. asperulum); ventral glands absent. Inflorescence (1) few- to ∞-flowered, with branching monochasial after 1st (or rarely 2nd) flower, from 5–15 nodes, often with flowering branches from up to 4 nodes below; buds erect. Flowers stellate, homostylous. Sepals 5, unequal to equal, free or up to 0.6 united, imbricate or not, persistent, erect in fruit, with margin entire to irregularly or regularly gland-fringed or denticulate; veins 3–5, not branched; laminar glands pale or very rarely black, linear to punctiform; marginal glands black or absent. Petals 5, persistent, unguiculate, spreading or deflexed after flowering, without apiculus, occasionally red-veined or red-tinged or (27b. capitatum var. capitatum) wholly suffused red; margin with sessile glands or glandular-ciliate; laminar glands pale or very rarely dark (23. asperulum), all punctiform or some shortly striiform or very rarely elongate-striiform (19. scabrum) or linear (20. thymopsis); marginal glands black, globose to ellipsoid or obconic. Stamen fascicles ‘3’ (i.e. 2+2+1), persistent, with stamens totalling 25–60, basally united in each fascicle; anther gland amber; pollen types X–XI, grains often partly irregular. Ovary with 3(4) axile placentae, each ∞–6-ovulate; styles 3(4), divergent from discrete bases; stigmas small or slightly capitate. Capsule 3(4)-valved, coriaceous to papyraceous, longitudinally vittate or rarely lateral vittae swollen (19. scabrum and 22. hirtellum in part), very rarely dehiscing as few- to 1-seeded cocci (26. olivieri). Seeds cylindric, not carinate or appendiculate; testa papillose.
BASIC CHROMOSOME NUMBER (X). 12, 10, secondary number 14; ploidy 2, 4.
Open woodland, meadows, steppe, dry stony areas, on acidic or basic substrates; 20–3900 m.
Eastern Kazakhstan, northwest China (Xinjiang) and Kashmir to Turkey and Jordan; The Krym, Serbia, Bulgaria, Greece, southern France, northwest Italy, southern Spain, Morocco.
30 species.
Section Hirtella is derived from Sect. 3. Ascyreia, its nearest relative in that section apparently being the western Himalayan H. oblongifolium Choisy. A form of this species (e.g. Polunin, Sykes & Williams 721 (BM) from near Culpani, Nepal) has a narrowly thyrsoid inflorescence similar to the basic inflorescence of Sect. Hirtella. The leaves of H. oblongifolium, however, are never as narrow as those of H. elongatum Ledeb., which, on several grounds, has a good claim to be the basic species of Sect. Hirtella. Nevertheless, this Himalayan species would seem to be the sister species of H. elongatum and thus of the whole Sect. Hirtella. Taking the combination of narrow leaves and a narrow inflorescence in H. elongatum as basic to the section has made the evolutionary trends in it much clearer.
Hypericum elongatum has a wide disjunct distribution from western Turkey to the Altai region of Kazakhstan and has given rise to five distinct lines of descent (Fig. 1, i–v). But this variation is not always accompanied by clear morphological disjunctions, a situation that has resulted in widely differing attempts at classifying the H. elongatum group. There are intermediates between H. elongatum and respectively the H. hyssopifolium group (Species 4–9), 10. H. apricum and 11. H. apiculatum, which have been treated variously as infraspecific relatives of H. elongatum itself or of H. hyssopifolium. In particular, the plants with free and regularly gland-fringed sepals that otherwise resemble H. elongatum have proved to be difficult to classify satisfactorily. I have therefore treated such intermediates as varieties, those leading to the H. hyssopifolium group (Species 4–9) as 1b. H. elongatum var. lythrifolium (central to western Anatolia), those from Turkmenistan linking H. elongatum with 10. H. apricum as 1c. var. racemosum, and those from eastern Anatolia and southern Transcaucasia linking H. elongatum with 11. apiculatum as 1d. H. elongatum var. antasiaticum.
In a previous treatment of Sect. Hirtella (Robson, 1986), I confused two species, both with elongate, narrowly cylindric to spiciform inflorescences, under the name H. apricum. The true H. apricum Kar. & Kir. (Sp. 10) has ovoid capsules and ranges from north-eastern Iran and adjacent Turkmenistan to the Alatau region of Kazakhstan, a few transitional forms with H. elongatum occurring in the Ashgabad region of Turkmenistan (elongatum var. racemosum). The plant with a similar inflorescence that ranges from western Anatolia to southern Transcaucasia has a globose capsule, and transitional forms to H. elongatum var. lythrifolium are found in central Anatolia. The correct name for this species (Sp. 4) would appear to be H. karjaginii Rzazade.
When making a detailed survey of the whole section for this revision, I also realised that some of the relationships discussed in the above-mentioned paper (Robson, 1986: 256) cannot be maintained. H. scabrum and its relatives (Species 18–20) can be differentiated from the rest of the group that I distinguished as subsect. Platyadenum by the capsule (ovoid-acuminate to ovoid-ellipsoid, not broadly ovoid-rostrate to depressed globose); so the corymbiform inflorescence of 18. H. scabroides and 19. H. scabrum, on the one hand, and the subcorymbiform one of 27. H. capitatum, on the other are the result of evolutionary convergence. In contrast, the branched glandular emergences on the stem of 20. H. thymopsis do seem to indicate an origin of that species from H. scabrum, not, as I previously thought, a relationship with 16. H. lydium.
The five evolutionary lines of Sect. Hirtella can be divided into two main groups (Fig. 1): Elongatum (Species 1–12) and Hirtellum (Species 13–30), which differ according to whether the bracts and bracteoles are broad or narrow (7. H. hyssopifolium, with rather narrow bracts, is an exception in the Elongatum group, cf. Ramos, 1985). Two other groups were originally distinguished respectively as subsects. Stenadenum and Platyadenum (Robson, 1986), depending on whether the glands fringing the sepals were ellipsoid to globose and rounded (Species 1–17) or flat-topped (Species 18–30). On further investigation, however, it was found that flat-topped glands are confined to species 21–30; so the Elongatum and Hirtellum groups mentioned above have been named Subsect. Stenadenum and Subsect. Platyadenum respectively, although these epithets are not now wholly appropriate.
Subsect. Stenadenum comprises four distinct derivative lines from 1a. H. elongatum var. elongatum (Fig. 1, i–iv), of which two species (2. tymphresteum and 3. callithyrsum) are directly derived and three species or species groups (given informal names below) are linked to it by intermediate varieties: (ii) Hyssopifolium (Species 4–9), linked by 1b. elongatum var. lythrifolium, (iii) Apricum (Sp. 10), linked by 1c. elongatum var. racemosum and (iv) Apiculatum (Species 11–12), linked by 1d. elongatum var. antasiaticum. None of these groups has been formally distinguished taxonomically here. In contrast, Subsect. Platyadenum is divided into three morphologically distinct subgroups (i) Lydium (Species 13–17, Fig. 1 line vi), (ii) Scabrum (Species 18– 20, Fig. 1 line vii) and (iii) Abbreviatum (Species 21–30, Fig. 1 line viii), all being derived directly from 1a. H. elongatum var. elongatum (Fig. 1, line v). These three subgroups are all on one evolutionary line (v), which, as has been stated above, can be distinguished by having linear (not broader) bracts and bracteoles. As they are morphologically distinct, they have been recognised here as series.
It is therefore possible to divide the four main groups described above into two subsections: Subsects. Stenadenum (Species 1–12) and Platyadenum (Species 13–30), the latter comprising three Series: Lydia (Species 13–17), Scabra (Species 18–20) and Hirtella (Species 21–30).