Revision of the Malesian species of Lobelia section Rhynchopetalum ( Campanulaceae : Lobelioideae )

You are free to share to copy, distribute and transmit the work, under the following conditions: Attribution: You must attribute the work in the manner specified by the author or licensor (but not in any way that suggests that they endorse you or your use of the work). Non-commercial: You may not use this work for commercial purposes. No derivative works: You may not alter, transform, or build upon this work. For any reuse or distribution, you must make clear to others the license terms of this work, which can be found at http://creativecommons.org/licenses/by-nc-nd/3.0/legalcode. Any of the above conditions can be waived if you get permission from the copyright holder. Nothing in this license impairs or restricts the author’s moral rights. Blumea 56, 2011: 218–224 www.ingentaconnect.com/content/nhn/blumea http://dx.doi.org/10.3767/000651911X604151 RESEARCH ARTICLE


INTRODUCTION
Lobelia L. sect. Rhynchopetalum (Fresen.) Benth. comprises over 60 species found from south-eastern Brazil across the Paleotropics to Malesia. These plants are robust tetraploids (2n = 28), often woody or pachycaul and up to 9 m tall, with sessile often apically rosulate leaves, terminal racemes or panicles of large unilabiate or sub-bilabiate flowers, and capsules producing numerous lenticular and commonly winged seeds with a striate-reticulate testa (Lammers 2011). Monophyly of the section is supported by several molecular phylogenies , 2006, Knox & Palmer 1998, Antonelli 2008. Although the African species have been the subject of numerous taxonomic studies (e.g., Bruce 1934, Hauman 1934, Hedberg 1957, Mabberley 1974, 1975a, b, Knox 1993, Knox & Kowal 1993, those from the remainder of the range have been accorded far less attention. In particular, the identification of the plants in Malesia has varied dramatically among the few authors who have addressed the question.
When the first Malesian representatives of L. sect. Rhyn chopetalum were discovered in the Philippines, they were identified (Merrill & Merritt 1910, Merrill 1923, Elmer 1934 as L. nicotianifolia Roth ex Schult., a species originally described from peninsular India. Skottsberg (1928), however, restricted that name to certain populations in India and Sri Lanka, and segregated the Philippine populations as L. philippinensis. This approach was accepted by Wimmer (1953Wimmer ( , 1968, who recognized in Malesia five additional taxa referable to this section as currently circumscribed: L. sumatrana on Sumatera; L. eryliae, L. epilobioides var. epilobioides, and L. epilobioides var. sarasinorum on Sulawesi; and L. epilobioides var. luzonica in the Philippines. All these names except L. sumatrana, as well as many others based on specimens from the Asian mainland (e.g., L. colorata Wall., L. pyramidalis Wall., L. rosea Wall.), were treated as synonyms of L. nicotianifolia in the influential Flora Malesiana. In that treatment, Moeliono (1960: 123 -125) considered these synonyms to apply to "microspecies" and "local forms" of a "widely distributed and very polymorphous species," which did not merit formal recognition. Subsequent authors (e.g., Cramer 1983, Lian 1983, Haridasan & Mukherjee 1988, Hong & Zhang 1992, Lammers 1998, 2007a have disagreed with Moeliono's expansive circumscription of L. nicotianifolia, and most of the names based on mainland materials have been reinstated. However, the status of the Malesian plants has not been re-examined and their identity remains problematic, a fact highlighted by Argent et al. (2007) in their discussion of possible relatives of the recently discovered Philippine species, L. proctorii.

MATERIALS AND METHODS
The objective of the present study is to re-examine the Malesian material of L. sect. Rhynchopetalum and to determine its best classification. As in my previous studies of insular taxa (e.g., Lammers 1991Lammers , 2005Lammers , 2007bLammers , 2009, I placed special emphasis on discerning geographically correlated patterns of morphological variation. Plants referable to L. sect. Rhynchopetalum have been collected on six islands in Malesia: Luzon, Negros, Biliran, and Sibuyan in the Philippines; and Sulawesi and Sumatera in Indonesia. The initial hypothesis was that each island would harbour a single morphologically defined taxon unique to it. If this hypothesis were correct, it would be possible to predict the provenance of a given specimen from its morphology, or its taxonomic identity from its provenance. To test the hypothesis, I gathered morphological data from a representative sample of over 90 specimens drawn from nearly a dozen herbaria (see Acknowledgements). From these data, I then attempted to answer two questions. First, were all specimens from a given island relatively homogeneous, with most characters evincing a continuous pattern of variation? If so, the hypothesis would be supported; if instead, several characters consistently showed correlated gaps in their patterns of variation within an island, the hypothesis would be refuted. Second, were the plants with a given suite of morphological features restricted to a single island? If so, the hypothesis would be supported; if not, it would be refuted.
Once taxa were discerned in this fashion, they were compared to nomenclatural type specimens to determine the correct name of each. Decisions on rank for the taxa were made in light of the definitions of species and subspecies I have employed previously (Lammers 1991(Lammers , 2005(Lammers , 2007a(Lammers , 2009.

RESULTS AND DISCUSSION
The material studied fell into eight taxa on the basis of comparative morphology. Negros, Biliran, Sibuyan, and Sumatera each have a unique taxon. Two taxa were discerned on Luzon, but they are allopatric, with one growing at elevations of 1600 m or less, while the other occurs only at 2100 m. Sulawesi likewise harbours two taxa. They are largely allopatric, with one restricted to South Sulawesi, and the other found primarily in Central Sulawesi; their sole area of apparent sympatry is the vicinity of Rantelemo in the Latimojong Mountains of South Sulawesi. These results support the hypothesis in large part, though not completely. The provenance of a given specimen can be predicted with confidence from its morphology, while the taxonomic identity of a specimen can be predicted from its provenance more often than not. None of the eight taxa matched material of L. nicotianifolia from India or Sri Lanka, nor did they match any other members of L. sect. Rhynchopetalum from the Asian mainland. Details of this broader comparative study will appear in a forthcoming monograph.
Two of the taxa are accorded specific rank on the basis of their highly distinctive habits. The Sumatera taxon, recognized universally as L. sumatrana (Wimmer 1953, Moeliono 1960, Lammers 2007a, is a rhizomatous slender-stemmed herb just 20-40 cm tall, with nearly all of the leaves confined to a basal rosette. The Sibuyan taxon (recently described as L. proctorii) is an apically rosulate treelet 40 -60 cm tall.
The six remaining taxa all are robust suffruticose herbs or shrubs, 1.3-4 m tall, with thick unbranched erect stems that are leafy throughout; these were the plants called L. nicotianifolia in Flora Malesiana. The South Sulawesi and high-elevation Luzon taxa differ from the remaining four by their distinctly smaller flowers, with the corolla only 12 -19 (vs 22 -39) mm long, the filament tube 5.5 -10 (vs 12 -18) mm long, and the dorsal anthers 2.5 -4.5 (vs 4 -6) mm long. Moeliono (1960: 125) had in fact noted that the specimens of his L. nicotiani folia fell into two broad groups based on floral size, but did not ascribe any importance to this difference because the two "do not show a geographical replacement". However, differences in size of corolla and staminal column typically are correlated with pollinator differences in the Lobelioideae, and thus may serve as a mechanism for reproductive isolation (Wood 1961, Young 1982, Lammers & Freeman 1986, Lammers 1991, 2000, 2009, Thompson & Lammers 1997. For this reason, the South Sulawesi and high-elevation Luzon taxa are judged to be specifically distinct from the remaining four. The South Sulawesi and high-elevation Luzon taxa can be distinguished from each other on the basis of morphology, but these differences are less substantive, involving quantitative variation (often contiguous or partly overlapping) in features that are not likely to contribute to reproductive isolation. Hence, they are considered to constitute a pair of allopatric conspecific subspecies. The sole name referable to the Philippine subspecies is L. epilobioides var. luzonica, while L. eryliae, L. epilobioides, and L. epilobioides var. sarasinorum refer to the Indonesian subspecies. The small-flowered species thus becomes L. eryliae, as that name has priority; its Indonesian subspecies takes the autonym, while a new combination at subspecific rank is effected here for the Philippine subspecies.
Among the four larger-flowered taxa, there is some difference in floral size between the Central Sulawesi taxon and those in the Philippines: the corolla of the former is 22 -25 (vs 26-39) mm long and the filament tube 12-13 (vs 13-18) mm long. In addition, the calyx lobes of the former are 4-6 mm long, equalling to only half again as long as hypanthium, while those of the latter are 7-14 mm long and typically two to three times longer. Although the gap between the two sets of taxa is less pronounced that that separating them from L. eryliae, it is nonetheless considered commensurate with recognition at species rank. No name is available for the Central Sulawesi species, so it is described here as new. The only names referable to the Philippine species are L. philippinensis and L. nicotianifolia var. mollis, so it takes the former.
As was the case in L. eryliae, the taxa comprising L. philippi nensis can be distinguished only by less substantive sorts of characters. Consequently, the islands of Luzon, Biliran, and Negros are each considered to harbour an endemic subspecies of L. philippinensis. The types of both L. philippinensis and L. nicotianifolia var. mollis represent the Luzon subspecies and so it takes the autonym. No names are available for the Biliran and Negros subspecies and they are described here as new.
Distribution -Endemic to northern Luzon in the Philippines, from Ilocos Norte to Aurora and Zambales.
Distribution -Endemic to Central and South Sulawesi in Indonesia.
Habitat & Ecology -Evergreen forest and meadows at 800-1620 m. Flowering and fruiting: April through November.
Habitat & Ecology -Evergreen forest and meadows at 600-2300 m. Flowering: January through June; fruiting: April through October.
Distribution -Endemic to Mt Paoay and Mt Santo Tomas in Benguet Province on Luzon in the Philippines.
Habitat & Ecology -Mossy forest among shrubs at 2100 m. Flowering: January through July.
Distribution -Endemic to Aceh Province in northern Sumatera.