Taxonomy, Phylogeny, and Biogeography of Baccaurea, Distichirhops, and Nothobaccaurea (Euphorbiaceae)

This study comprises a taxonomic revision and phylogenetic and biogeographic analyses of the genera Baccaurea, Distichirhops, and Nothobaccaurea (Euphorbiaceae). Baccaurea was chosen as a model genus for study in the framework of the research program ‘Pathways from Asia to new Guinea’ of the Dutch Science Foundation because of interesting problems in the areas of phylogeny (Chapter 1), biogeography (Chapter 2), and species delimitation (Chapter 3). Moreover, Baccaurea species are also of economic and ecological importance. During the study of Baccaurea it became obvious that two species from Pacific islands, formerly recognised as Baccaurea species, constitute in fact a new genus: Nothobaccaurea. Furthermore, it became obvious that three unidentified species do not belong to any known genus; they are placed in an also newly described genus, called Distichirhops. Chapter 1 contains the phylogenetic analyses of Baccaurea and Nothobaccaurea. The aim was to find the best hypothesis for the genealogical relationships of the species in Baccaurea, Distichirhops, and Nothobaccaurea. The phylogenies of these genera are a necessary prerequisite for the biogeographic analyses. The morphological species concept is used. Hybridisation is not thought to be a major speciation mechanism in Baccaurea. A study of the subtribe Scepinae showed the monophyletic status of Baccaurea, Distichirhops, and Nothobaccaurea. Aporosa is chosen as outgroup, because it is the sistergroup of the clade containing the three revised genera. Distichirhops and B. microcarpa are excluded from further analyses, because of our incomplete knowledge of these taxa. A total of 102 characters have been defined and scored, most of which have distinct states. Fifteen characters were later omitted for several reasons. The final analyses are, therefore, based on 87 characters, of which 23 concern vegetative, 22 staminate, 23 pistillate, 7 fruit, 2 ecological and 10 leaf anatomical characters. For most of the parsimony analyses the computer program PAUP 3.1.1. is used. A parsimony analysis with equally weighted characters resulted in 12 Most Parsimonious Cladograms (MPCs), with a fairly well resolved Strict Consensus Cladogram (SCC). The confidence indices for cladistic analyses like Consistency Index and Retention Index are within the expected range for a data matrix of this size. Even though the SCC is well resolved most internal branches of the 12 MPCs are not corroborated, because they have no Bremer support. A new method of analysis, called Iterative Taxon Reduction (ITR), is applied to find the relatively stable parts of the MPCs, and to make a possible selection among the 12 MPCs. This new method reduces the number of taxa by replacing stable clades (= those with Bremer support) in the data matrix by the character states of their common internal node. A new analysis is then performed with the reduced data matrix. The procedure is repeated till no further reduction of the data matrix is possible or till all clades have obtained Bremer support. The result is a SCC-like cladogram with relative Bremer support per clade. Other methods suitable for selecting cladograms are weighted character analyses. The weighted cladograms are dismissed, because they did not select among the 12 MPCs and preferred other, less homoplasious equal weight cladograms. Therefore, all 12 MPCs are chosen as equally good hypothesis for the phylogenetic relationships in Baccaurea and Nothobaccaurea. Important conclusions drawn from the Most Parsimonious Cladograms are that the New Guinean species of Baccaurea are not closely related to each other except for B. papuana and B. carinata. Till recently the phylogenetic connections of several species were unclear, like the close relationships among B. angulata, B. lanceolata, B. dulcis, and the newly described B. purpurea. Other relationships, already expected by former taxonomists, like those of the Pacific species B. nesophila and B. taitensis could be confirmed. The consequences of the phylogenetic analyses for formal classification are limited to the confirmation of Baccaurea and Nothobaccaurea as monophyletic entities. Chapter 2 contains the biogeographic analyses. The most important question is how, in time and space Baccaurea did reach New Guinea. Baccaurea occurs on both sides of Wallace’s line (a division through the central part of the Malay Archipelago), which means that Baccaurea can only have reached one of the sides through dispersal. Taxon biogeography is based on the phylogenetic relationships of the taxa and the knowledge of the geology of the areas where the taxa occur. The geological framework is derived from the literature. The areas used in the analyses are rather large. The first method used is substituting the species in the 12 MPCs by their distribution areas. The split up of the genera Baccaurea, Maesobotrya, and Nothobaccaurea corresponds with the geological split up of Asia, Africa, and the Pacific. Three or even four dispersal moments between the Sunda Shelf and New Guinea can explain the occurrence of Baccaurea in New Guinea. Brooks Parsimony Analysis (BPA) leads to a single scenario with one unexpected feature: Baccaurea first arrived in the Pacific from the mainland of South East Asia, and only later dispersed from the Pacific to New Guinea, the Philippines and the Sunda Shelf. Because of this unexpected result the method was adjusted to the situation in Baccaurea. Ancestral areas in BPA are simply defined as the combined distributions of the descendant taxa. This is implausible in Baccaurea, because geological history shows that the genus largely occurs in areas, which were only reached by dispersal. It is, therefore, assumed that the ancestral distribution areas roughly correspond to the distribution areas of the modern species. An analysis based on this idea showed only little resolution. One of the possible interpretations of this result could be that Baccaurea reached most areas at roughly the same time. Hovenkamp Analysis (HA) defines the borders where certain vicariance events occurred, and also the time frame in which the events took place. HA also indicates a dispersal route for Baccaurea from India, via the Southeast Asian mainland and the Sunda Shelf to New Guinea. Finally, the distribution patterns in Baccaurea are compared with a general area cladogram based on the analysis of many zoological and botanical taxa. This also strongly indicates that the most basal branch in Baccaurea occurred in India. The combined information of all above mentioned methods are combined in one dispersal/vicariance scenario (Fig. 2.12) for Baccaurea and allied genera. Conclusions drawn from this scenario are that Baccaurea, Maesobotrya, and Nothobaccaurea are of Gondwanan origin. The common ancestor of the clade containing these three genera speciated as reaction on the vicariance event coursed by the split off of Africa, India, and Australia from Gondwana. Nothobaccaurea went extinct or was primitively absent on Australia, but survived on Fiji. Baccaurea dispersed from greater India via the Southeast Asian mainland into the Sunda Shelf. From the Sunda Shelf dispersal back to India and the Southeast Asian mainland took place as well as dispersal to the Philippines, New Guinea and the Pacific. Chapter 3 contains the revisions of the three genera. Baccaurea, Nothobaccaurea and Distichirhops all belonging to subtribe Scepinae of subfamily Phyllanthoideae. A key to the genera of this subtribe is presented. The taxonomic revision of Baccaurea Lour (1790) comprises a brief history of the genus, a genus description, a general and several regional identification keys to the species, and descriptions of all 43 species. In comparison with the studies by Airy Shaw, who published most recently about Baccaurea, my treatment differs in the following perspectives. The subspecies B. nesophila var. microcarpa is raised to the species level. Thirteen names are synonymised. Four species are newly described, viz. B. carinata, B. mollis, B. purpurea, and B. simaloerensis. Two species, formerly belonging to Baccaurea, are transferred to Nothobaccaurea, viz. B. stylaris and B. pulvinata. Finally, B. dolichobotrys, which was formerly interpreted as a synonym of Baccaurea polyneura, received the species status again. Four names could not be interpreted and they are enumerated as species dubiae. Three names are excluded, because they do not belong to Baccaurea.