A taxonomic revision of the small Cape genus Steirodiscus ( Asteraceae : Senecioneae )

The genus Steirodiscus Less. is revised, with full descriptions, nomenclature, illustrations and distribution data. Five species are recognised in two sections: sect. Steirodiscus with free involucral bracts and sect. Psilothonna with bracts connate into a smooth cup. S. linearilobus DC. is established as an earlier name for S. schlechteri Bolus, which is relegated to synonomy.


INTRODUCTION
Steirodiscus Less. is a minor genus of Senecioneae comprising fi ve species of small, wiry annuals, all endemic to the Greater Cape Floristic Region (Goldblatt & Manning 2000;Herman et al. 2000).Currently included in subtribe Senecioninae (Pelser et al. 2010), its relationships within the subtribe are still unclear (Nordenstam et al. 2009).Although treated as one of the 'othonnoid' genera by Jeffrey (1992) on account of its ebracteolate and sometimes gamophyllous involucre, molecular data place it among the group of genera that includes Bolandia Cron, Cineraria L., Emilia Cass., Mesogramma DC. and Stilpnogyne DC., rather than with the Othonninae (Nordenstam et al. 2009;Pelser et al. 2010).A recent phylogenetic analysis of plastid and nuclear sequence data identifi es Steirodiscus as one of several lineages with strongly incongruent positions in the two analyses (Pelser et al. 2010).Hybridization and incomplete lineage sorting were suggested as explanations for the incongruence.
Steirodiscus was established by Lessing (1832) to accommodate the species described by Linnaeus fi l. (1782) as Cineraria capillacea, and was originally characterised by a uniseriate involucre of free bracts without associated bracteoles, female ray fl orets with pubescent cypselas, and male-fertile disc fl orets.At the same time Lessing (1832) also recognised the genus Gamolepis for four shrubby Cape species now treated in Euryops plus the solitary annual species Othonna tagetes L. (under the illegitimate superfl uous name Gamolepis annua Less.), which were linked by their smooth, cup-shaped involucres, female ray and bisexual disc fl orets, and epappose cypselas.These generic circumscriptions were followed by Candolle (1838), who added a second species, Steirodiscus linearilobus DC.The sole annual species of Gamolepis, G. tagetes (L.) DC., which had by then been selected as the lectotype of the genus (Pfeiffer 1874), was subsequently transferred to Steirodiscus by Schlechter (1899), who was struck by its marked similarity to that genus in habit, fi nely dissected leaves and epappose cypselas, despite the marked difference in their involucres.At this time two further species were described, S. gamolepis Bolus and S. schlechteri Bolus (Schlechter 1899), bringing the total in the genus to fi ve.A sixth species [now S. speciosus (Pillans) B.Nord.] was later named by Pillans (1931) under the generic name Gamolepis.
Steirodiscus is defi ned by its annual habit, pinnatisect or bipinnatisect leaves with subsecund, linear-fi liform segments, and solitary or laxly corymbose, radiate capitula with unseriate, ebracteolate involucre, and all fl orets lacking a pappus.The involucral bracts may be free or connate into a smooth cup.The generic name is derived from the sterility of the disc fl orets in S. capillaceus (Lessing 1832), but this condition is not easily assessed in fl owering capitula as maturation of the central fl orets is tardy.We have confi rmed that the central disc fl orets in S. tagetes fail to develop fruit but whether this is due to congenital sterility or to nutritional constraints we cannot say.This will need to be assessed through careful anatomical investigation for the presence or absence of an embryo in the central disc fl orets.
A brief review of the group under the illegitimate superfl uous generic name Psilothonna (E.Mey.ex DC.) E.P.Phillips (Nordenstam 1968) was presented by Phillips (1950), who also provided a key to the fi ve species recognised.This review, although based on very few collections, provided some much-needed clarity in the genus but Phillips (1950) inexplicably overlooked Candolle's (1838) S. linearilobus.This name was applied (incorrectly as we show below) by Hutchinson (1946) to plants from the Klein Roggeveld and the species has remained poorly understood until now.
Five names in Steirodiscus are included in the Red List of South African Plants (Raimondo 2009), four as either EN or VU and S. linearilobus as DD signifying its uncertain identity.At the moment therefore, although Bothalia 43,1: 109-119 (2013) A taxonomic revision of the small Cape genus Steirodiscus (Asteraceae: Senecioneae) Bothalia 43,1 (2013) six names are validly published in the genus, only fi ve species are accepted (Goldblatt & Manning 2000).The genus is urgently in need of revision, not only to establish the number of recognised species and the application of the available names, but also the distribution and thus conservation status of the taxa.Three of the species (S. gamolepis, S. linearilobus and S. speciosus) are highly localised endemics and another (S.tagetes) is restricted to the West Coast, where it is evidently already extinct from part of its range and threatened by coastal development elsewhere.The fi fth species (S. capillaceus), although relatively widespread, has not been recorded from most of its historical range for over 50 years.
We provide a complete taxonomic revision of Steirodiscus, including full descriptions and synonymy, illustrations and distribution data for all species.The species fall into two groups depending on whether the involucral bracts are free to the base or connate into a smooth cup, which we treat as the sections Steirodiscus and Psilothonna respectively.

Additional specimens seen
Steirodiscus linearilobus was described from a collection made by Drège at the base of the Knakkiesberg, an inselberg west of Klawer (Candolle 1838) but the name was inexplicably overlooked by Phillips (1950) in his review of the genus (as Psilothonna).A later collection from just east of Klawer formed the basis for Steirodiscus schlechteri (Bolus 1899), separated from S. linearilobus essentially by its glabrous ovaries.The vestiture of the ovaries in S. linearilobus is not mentioned in the protologue and its identity has remained uncertain until now.It is only through the kindness of Mark Newman at the Royal Botanic Garden, Edinburgh, who dissected the Edinburgh isotype and confi rmed that the ovaries of Drège's collection are glabrous (Newman pers.com.), that we are able to establish fi rmly that the two names are conspecifi c.Bolus (1899) distinguished S. schlechteri from S. linearilobus (and S. capillaceus) by its stouter habit, smaller leaves, larger heads, and glabrous achenes, but this is evidently just a general diagnosis of the species rather than an implication that the ovaries in S. linearilobus are pubescent.This implied pubescence of the ovaries in S. linearilobus led Hutchinson (1946) to misapply the name to his collection of S. capillaceus from the Klein Roggeveld.

Additional specimens seen
Distribution and ecology: Steirodiscus tagetes is relatively widely distributed along the West Coast of Western Cape, from St Helena Bay to False Bay (Figure 5).There are historical records from the Cape Peninsula but the species is probably now extinct there.It occurs on sand dunes near the coast, typically in lightly disturbed situations in strandveld vegetation, especially Cape Flats Dune Strandveld, where it may be locally abundant.
Diagnosis and relationships: the most common member of sect.Psilothonna, Steirodiscus tagetes is distinguished by the characteristic compound trichomes on the ray fl orets.These palmate structures (Figure 4E) form a fl eshy, collar-like fringe around the top of the corolla tube at the base of the limb (Figure 4C, D), and are readily visible with a hand lens.The capitula are typically larger than in S. gamolepis and S. speciosus, with a campanulate involucre 4-5 mm diam.It is superfi cially very similar to S. speciosus, but in this species the trichomes on the ray fl orets are not coalescent into compound structures and the involucre is generally smaller, 2-3 mm diam.The fl owers in both species are typically bright yellow but may also be tinged orange.The species is variable in fl ower size with the largest and most attractive forms occurring west of Atlantis.
Distribution and ecology: a highly local endemic of the West Coast between Darling and Atlantis (Figure 5), Steirodiscus speciosus is very poorly known.Described from plants collected between Darling and the coast Diagnosis and relationships: Steirodiscus speciosus is very similar to S. tagetes and was diagnosed against it by Pillans (1931) (as G. annua), and as having larger (± 18 × 3.5-4.0mm), more numerous (14-23) orange rays lacking a distinct annulus of papillae around the mouth of the tube and with more strongly exserted styles.Most of these characters do not hold, even in the type of S. speciosus, which has up to 15 rays that are no larger than are found in most populations of S. tagetes and also does not differ in the stylar characters.Although the specifi c status of the taxon was questioned in 1933 by R.A. Dyer (letter on K isotype), Phillips (1950) accepted the species in his synopsis of the genus (as Psilothonna), but keyed it against S. tagetes solely on the absence (vs.presence) of 'glands' at the base of the rays.Both taxa, like all species in the genus, have multicellular trichomes ('glands') at the top of the ray corolla tube on the abaxial (outer) side and the critical difference between them is in the nature of the trichomes themselves.In S. tagetes the trichomes coalesce into compound, palmate structures (Figure 4E) that form a fi mbriate collar around the corolla tube (Figure 4C, D) whereas in S. speciosus the individual trichomes remain discrete (Figure 6C,D).S. speciosus is also distinguished by its smaller involucre, 2-4 mm diam., which constricts markedly above the ovaries as it dries, giving it a characteristic urceolate shape in herbarium specimens (Figure 6B).This is not evident in fresh material, at which time the distinction between the two species is more cryptic, requiring careful examination of the ray fl oret vestiture with a hand lens.Ecologically S. speciosus appears to favour sand plain fynbos (this remains to be confi rmed for additional populations) whereas S. tagetes is defi nitely restricted to strandveld vegetation along the coast.

Diagnosis and relationships:
The smallest species in the genus, Steirodiscus gamolepis is distinctive in its short, few-lobed leaves and diminutive, markedly pedunculate capitula with the peduncle grading into the narrowly campanulate or obconic involucre.The rays are relatively few, up to half as many as the involucral lobes, and the ovaries and cypselas are pubescent in the basal two thirds but glabrous above.The corollas of the disc fl orets are glabrous and unique in the genus in the Distribution and ecology: Steirodiscus capillaceus is the most widely distributed species in the genus, with a curiously scattered range along the West Coast from Hopefi eld to the foot of Piekenierskloof Pass, in the northern Cedarberg and Bokkeveld Mountains, and further inland on the Klein Roggeveld at the foot of the Komsberg Pass (Figure2).It favours sandy loam soils, typically in renosterveld shrubland.Diagnosis and relationships: Steirodiscus capillaceus shares free involucral bracts with S. linearilobus, but is distinguished by the puberulous ovaries and cypselas, and by the sparsely puberulous corolla lobes of the disc fl orets.The uppermost leaves are invariably fi liform.
Distribution and ecology: only rarely collected, Steirodiscus linearilobus is a local endemic of the lower reaches of the Olifants River and adjacent fl ats south and east of Vanrhynsdorp, recorded from the Knakkiesberg west of Klawer, northeastwards to the foot of the Kobee Mountains (Figure2).The species occurs on sandy fl ats in open shrubland, including Klawer Sandy Shrubland(Mucina & Rutherford 2006).