New combinations in Decalobanthus (Convolvulaceae)

The recent revisionary work of “Merremieae” has resulted in the segregation of the c. 100 species of the pantropical genus Merremia into six genera. Thus, the formerly monotypic genus Decalobanthus was expanded, aggregating 13 species of Merremia s.l. The genus delimitation is coherent, with strong molecular phylogenetic and morphological support. Decalobanthus, hence, consists of woody climbers with broadly cordate leaves, large yellow or white flowers, four-valved chartaceous capsules, and is mostly distributed in Asia and the Pacific. Four species have remained classified in Merremia s.l., although molecular and morphological evidence unequivocally suggests their placement in Decalobanthus. In the present work, these are formally transferred to the genus, which now extends to 17 species.

Decalobanthus, in particular, results from the expansion of a previously monotypic genus endemic to Sumatra (Indonesia), with the type species D. sumatranus Ooststr. This genus was initially described by Van Ooststroom (1936) for the presence of a small tubular corolla, the lobes of which are reflexed, with each lobe further parted into two. As the flowers of Convolvulaceae are pentamerous, this creates a particular "10-lobed" arrangement of the corolla fauce, that inspired the generic name Decalobanthus. As morphological and molecular studies recently brought to light that this species is nested within a clade of other woody climbers from SE Asia (Simões et al. 2015;Simões & Staples 2017), the generic delimitation was expanded, accommodating 12 additional species previously placed in Merremia s.l. The genus now comprises 13 species, the monophyly of which is strongly supported in molecular phylogenetic analyses based on ITS and three plastid markers (Simões et al. 2015;Fig. 1).
Decalobanthus is not characterised by unequivocal synapomorphies, viz. characters that uniquely evolved in this genus. However, species of Decalobanthus are robust woody climbers or lianas, for which they are very easily separated from the herbaceous or slender twining herbs of Merremia s.s. (Fig. 2). A combination of other morphological characters helps to differentiate them from closely related genera: leaves simple, entire, commonly large and broadly cordate; inflorescences paniculate or corymbiform, the lowermost bract often foliaceous; sepals strongly convex (boat-shaped); corolla bright yellow to white, usually glabrous outside; anthers spirally twisting at dehiscence; fruits valvate capsules with the exocarp delaminating above the middle (lower half is dark brown, upper half is strawcoloured); seeds always pubescent, often with long golden hairs either covering the entire surface or concentrated along the edges of the seeds (Simões & Staples 2017;Fig. 2). The genus is distributed mostly across SE Asia and the Pacific (Map 1), with one species being widespread to Eastern Africa and Madagascar (D. peltatus (L.) A.R. Simões & Staples) and Santa Cruz Island of North America (D. bracteatus (P.S.Bacon) A.R.Simões & Staples) (Map 1). As a new expanded circumscription of Decalobanthus is adopted and contributes to a better understanding of the relationships and biogeographical patterns of this group of species, several challenges come in the way of a full revisionary study of this genus. Unlike many Convolvulaceae taxa, Decalobanthus are woody vines that grow in primary forests and require targeted fieldwork in undisturbed and well-conserved patches of forest. Many collections are available in Asian and European herbaria. Still, a significant proportion of the specimens is relatively old, preserved in alcohol, or have been treated with pest-control chemicals, which has impeded adequate molecular sampling of the genus. Also, the wide geographic range, with several narrow endemic species from hardly accessible areas, as the islands in the Pacific, further complicates the mission of sampling and morphologically comparing the taxa. Taxonomic studies are underway with the aim of improving our knowledge of the genus, namely by Pisuttimarn, who has conducted targeted fieldwork for specimen collection, and fresh samples for DNA analysis. However, it could be some time until we have the deep knowledge of the full morphological characterisation of the genus, a good understanding of the interspecific relationships, and a finer-tuned delimitation of the species.
The current expanded delimitation of Decalobanthus encompasses in the most part the species of Merremia previously placed in two sections: 1) section Hailale Hallier f., emended by Van Ooststroom (1939)  In the light of molecular phylogenetic results (Simões et al. 2015;Fig. 1), and considering that these species present the diagnostic characters of Decalobanthus as expanded by Simões & Staples (2017;Fig. 2), they have, for the most part, been combined under this name (Simões & Staples 2017). However, two species in section Hailale (Merremia clemensiana Ooststr. and M. crassinervia Ooststr.), one in section Wavula (Merremia calyculata Ooststr.) and one without an assigned section (M. gracilis E.J.F.Campb. & Argent) have not been, up until now, formally transferred to Decalobanthus, although they confidently belong in this genus. Merremia gracilis was even included in the molecular phylogenetic studies of Simões et al. (2015), and resolved as a member of the Decalobanthus clade. However, these names were never combined under Decalobanthus in Simões & Staples (2017), for uncertainty about the species delimitation.
Three of the species (Merremia clemensiana, M. crassinervia and M. gracilis) are morphologically similar to Decalobanthus korthalsianus (Ooststr.) A.R.Simões & Staples and overlap in geographical distribution (all documented as endemic to Borneo). Merremia calyculata is known only from the type specimen, collected on the Fiji Island of Taveuni (Smith 1991). The morphological and distributional overlap with D. pacificus (Ooststr.) A.R.Simões & Staples is considerable, for which its status as a distinct species is also dubious and should deserve attention in future studies (Staples 2009(Staples , 2010.
In the context of this study, we propose to combine the names of these four species under Decalobanthus, given the confidence in their generic placement, to avoid ambiguity in systematic and floristic studies of the region.
Hence, in total, 17 taxa are now considered to belong in Decalobanthus, based on genetic markers and conspicuous morphology (Table 1). It is hoped that ongoing systematic, morphological, palynological and anatomical studies will contribute to further elucidation of the relationships between these species and the remainder of the genus.