New synonyms and combinations in Drimia Jacq . ( Hyacinthaceae ) in southern Africa

Zingela N.R.Crouch et al. (Crouch et al. 2018) and Austronea Mart.-Azorín et al. (Martínez-Azorín et al. 2018a) are two recently described southern African genera of Hyacinthaceae subfamily Urgineoideae (Angiosperm Phylogeny Group [APG] 2003), alternatively treated as Asparagaceae subfamily Scilloideae tribe Urgineeae (APG 2009, 2016). They were named as part of an ongoing generic reclassification of the subfamily that derives from an as-yet only partially published molecular phylogenetic analysis of the group. This classification is increasingly divergent from the conservative interpretation proposed by Manning, Goldblatt and Fay (2004) that adopted a larger, inclusive circumscription of the genus Drimia Jacq. The latter classification is the preferred option among southern African botanists, where it has been adopted by the major herbaria in the country (e.g. BOL, NBG, PRE) and in recent publications on the flora (e.g. Manning & Goldblatt 2012, 2018; Snijman 2013; Van Jaarsveld & Eggli 2016). The two genera Austronea and Zingela accord, respectively, with Drimia sects. Marginatae and Thuranthos (= Vera-duthiea Speta) in the recent revision of the southern African species of Drimia (Manning & Goldblatt 2018).


Introduction
Zingela N.R. Crouch et al. (Crouch et al. 2018) and Austronea Mart.-Azorín et al. (Martínez-Azorín et al. 2018a) are two recently described southern African genera of Hyacinthaceae subfamily Urgineoideae (Angiosperm Phylogeny Group [APG] 2003), alternatively treated as Asparagaceae subfamily Scilloideae tribe Urgineeae (APG 2009(APG , 2016)).They were named as part of an ongoing generic reclassification of the subfamily that derives from an as-yet only partially published molecular phylogenetic analysis of the group.This classification is increasingly divergent from the conservative interpretation proposed by Manning, Goldblatt and Fay (2004) that adopted a larger, inclusive circumscription of the genus Drimia Jacq.The latter classification is the preferred option among southern African botanists, where it has been adopted by the major herbaria in the country (e.g.BOL, NBG, PRE) and in recent publications on the flora (e.g.Manning & Goldblatt 2012, 2018;Snijman 2013;Van Jaarsveld & Eggli 2016).The two genera Austronea and Zingela accord, respectively, with Drimia sects.Marginatae and Thuranthos (= Vera-duthiea Speta) in the recent revision of the southern African species of Drimia (Manning & Goldblatt 2018).
The application of molecular phylogenetic techniques in the family provided testable hypotheses of relationships among the species for the first time.Stimulated by the initial findings (Pfosser & Speta 1999), the Austrian taxonomist Franz Speta (1998) rejected the concept of a more broadly circumscribed Drimia that had been gaining acceptance among taxonomists working actively on the African elements (e.g.Jessop 1977;Stedje & Thulin 1995) in favour of narrowly circumscribed genera as determined by the clades resolved in the first phylogeny generated for the family.This analysis included just 15 samples of Urgineoideae, representing as many taxa out of a current estimated total of ± 120 species (Manning & Goldblatt 2018).Until this time, genera in the subfamily had been recognised based on differences in floral morphology, primarily the ramification of the inflorescence, the degree of fusion of the perianth, and the attachment and mode of dehiscence of the anthers (e.g.Baker 1897;Jessop 1977).The progressive discovery of new species that blurred many of these differences highlighted the inadequacy of the conventional taxonomy of the time, prompting botanists to broaden the circumscription of Drimia to include the segregates Tenicroa Raf., Thuranthos C.H.Wright and Urginea Steinh, but retaining the morphologically more distinct genera Litanthus Harv., Schizobasis Baker and Rhadamanthus Salisb (Jessop 1977;Stedje & Thulin 1995).
The emerging phylogenies appeared to provide adequate justification for the splitting of an enlarged Drimia into smaller, evidently monophyletic genera, and on this basis Speta (1998) recognised 13 genera in a subfamily in which contemporary authors recognised just five.Increased sampling, however, continues to reveal that some of these smaller segregates are themselves not monophyletic, demanding the description of additional, smaller segregates in order to maintain generic monophyly (see Manning & Goldblatt 2018 for a summary).This applies to even the smallest possible genera: just a few years after establishing the ditypic genus Sagittanthera Mart.-Azorín et al. (Martínez-Azorín et al. 2013), its authors discovered that the two member species were not in fact sister taxa, despite the superficial morphological similarity between them, and perforce transferred one species to a new monotypic genus Auolstemon Mart.-Azorín et al. (Martínez-Azorín et al. 2017).As current sampling stands, the number of genera in Urgineoideae recognised by Martínez-Azorín and coworkers is more than 20 and increasing.The justification for this plethora of genera is solely the application of the principle of monophyly as driven by the initial decision of Speta (1998) to implement a taxonomy based on an incomplete phylogeny that included little more than 10% of the species known at the time.Additional considerations like utility, nomenclatural stability and phylogenetic support, as identified by Backlund and Bremer (1998), have played no role in informing what is becoming an increasingly fragmented taxonomy.Indeed, it seems likely that the initial decision to recognise smaller genera may have been a premature one that has driven subsequent events.It is possibly significant that two of the authors involved in this process have also proposed splitting the well-established genus Iris L. (Iridaceae) (260-300 spp.) into at least 23 smaller segregates (Mavrodiev et al. 2014).This proposal has, unsurprisingly, been ignored.
In contrast, a more broadly circumscribed Drimia is readily diagnosed and identified, and it permits the recognition of numerous infrageneric groups as required to accommodate the morphological complexity in the group without affecting nomenclatural stability (Manning & Goldblatt 2018).Claims that this option is 'based on disputable phylogenetic data' (Crouch et al. 2018) are inexplicable and patently unjustified (Ali et al. 2013;Buerki et al. 2012).
Here we provide new combinations and other nomenclatural adjustments necessitated by the appearance of three recent publications describing yet another two genera in the subfamily (Crouch et al. 2018;Martínez-Azorín et al. 2018a, b).In addition, the examination of paratypes and more recent collections of Ornithogalum toxicarium C.Archer & R.H.Archer (Archer & Archer 1999) revealed that this species is misplaced in subfamily Ornithogaloideae and that it is actually a species of Drimia, and we accordingly transfer the name to that genus.

Taxonomic treatment
Drimia Jacq., Collectanea Suppl.: 38 (1797).Type species: Drimia elata Jacq.Note: The new genus Austronea coincides exactly with the circumscription of Drimia sect.Capitatae as described by Manning and Goldblatt (2018), but includes several newly described species for which we provide the necessary combinations or synonymy in Drimia.Note: Drimia densiflora is morphologically anomalous in the section in its very short pedicels, forming a congested, subspicate raceme rather than the corymbose-capitate raceme that characterises the other species.Its association with other species in the group is impossible to determine on morphology alone, and we assume that it was described in Austronea based on phylogenetic data.On the basis of its morphology, Austronea densiflora keys to Drimia sect.Ledebouriopsis in the recent revision of the genus by Manning and Goldblatt (2018).Here it resembles both D. salteri (Compton) J.C.Manning & Goldblatt from the south-western mountains of the Western Cape, from which it is distinguished by its shorter inflorescence and shorter, thicker leaf, and D. barbata J.C.Manning & Goldblatt from the Richtersveld, which it closely resembles in the size and shape of its inflorescence but which has a basally puberulous scape, flowers with a pubescent cup and a longer, more slender leaf.Note: Manning and Goldblatt (2018), in their recent monograph of Drimia in Southern Africa, included in their circumscription of D. vermiformis the material segregated as Austronea linearis by Martínez-Azorín et al. (2018a) on the basis of its slightly narrower leaves, as well as that from the Roggeveld that they recognised as A. grandiflora on the basis of its slightly larger flowers and less flattened leaves.We have re-examined the relevant collections in light of this but remain unconvinced that there is adequate reason to segregate them from D. vermiformis.We accordingly relegate these names to synonymy.In contrast, collections with unusually short, succulent leaves that were identified as the new species A. pinguis by Martínez-Azorín et al. ( 2018) in our opinion represent a distinct species.We had excluded the single collection of this taxon that was available to us from our circumscription of D. vermiformis but were unable to formally describe it as it lacked flowers.

Drimia barkerae
Drimia virens (Schltr.)J.C. Manning and Goldblatt in Goldblatt and Manning in Strelitzia 9: 712 (2000).Urginea virens Schltr. in J. Bot., British and Foreign 35: 433 (1897) Note: Austronea papillosa was segregated from D. virens on the basis of the papillate peduncle and pedicels; however, this condition is also found in populations of D. virens from Vanrhynsdorp and the Cederberg, as described by Manning and Goldblatt (2018), who included paratype material of A. papillosa under that species.We are not able to differentiate two taxa on this basis and regard them as one.
We, however, follow Martínez-Azorín et al. (2018a), who concluded that we had erred in our identification of the type of D. virens as conspecific with Urginea minor (sect.Physodia).The differences between the two species are not always clear in herbarium material and a similar error had earlier been made by Speta (1998).From their examination of isotype material at Edinburgh, Martínez-Azorín et al. (2018a) concluded that the flowers lack the fusiform filaments that are diagnostic for sect.Physodia.The name D. virens thus becomes the earliest name for the species in sect.Capitatae that we treated as D. pygmaea, and Drimia minor remains the earliest name for the species in sect.Physodia that we treated there under the name D. virens.
The very small bracts are, however, anomalous here.The species was transferred to Albuca subg.
with other species of Ornithogalum with similar floral and fruit morphology, by Manning et al. (2004) in their revised generic classification of subfamily Ornithogaloideae.Recent collections of the species that we have seen, however, confirm the initial conclusions of A.A.Obermeyer that it is in fact a species of Drimia.This is clear from the presence of a short but evident spur on the lowermost bracts of some and other specimens, the uni-nerved tepals in the type and all other collections, and the winged seeds with loose testa.Examination of the type, which comprises several plants, shows it to be a mixed gathering, with the flowering and sterile plants conforming to the protologue description and accompanying illustration but the fruiting material representing a species of Albuca, as is evident from the large bracts, multi-nerved tepals and prismatic style.We accordingly exclude the fruiting material and lectotypify the name against the flowering and sterile material only (Turland et al. 2018: Art. 9.14).The relationships of the species in Drimia are not immediately apparent, but the congested, relatively longlived flowers without bracteoles and the moderately large capsules with perianth persisting below the developing fruit for somewhile suggest a possible relationship with D. sanguinea (Schinz) Jessop (sect.Macrocentrae).(2018) have concluded that D. indica is strictly an Indian species based on an as-yet unpublished phylogeny.These authors accordingly described the new genus and species Zingela pooleyorum for the plants from KwaZulu-Natal, stating that the taxon is likely also to occur in southern Mozambique.And indeed Manning and Goldblatt (2018) considered that Urginea zambesiaca from Expedition Island in Mozambique was conspecific with the KwaZulu-Natal plants, in which case it is the earliest available name for Z. pooleyorum, which thus becomes a synonym.Zingela pooleyorum is a perfect match for the illustration of Urginea zambesiaca provided by Kativu and Drummond (1994), which shows the spathulate bulb scales, relatively short pedicels, and suberect, filiform filaments that are diagnostic for the species.
Note: see under D. vermiformis below.