Monopetalanthus exit. A systematic study of Aphanocalyx, Bikinia, Icuria, Michelsonia and Tetraberlinia (Leguminosae,Caesalpinioideae)

<img src="/wda/abstracts/i2686_01.jpg" ALIGN="Right" HSPACE="6" VSPACE="4" WIDTH="148" HEIGHT="196" ALT="© Photo: Jan Wieringa mounting Tetraberlinia moreliana" BORDER="0"/>This dissertation deals with the ecology and the taxonomy of a group of dominant forest trees of the African high forest. Their size and inaccessibility has accounted for their comparative rarity in herbarium collections. Recently, due to various original collection methods, particularly special tree mounting techniques, quite a number of new collections from the canopy were obtained which provided a better insight into these taxa.The core question at the start of this study was the generic and specific delimitation of Monopetalanthus , one of the most troublesome genera of African Caesalpinioideae. The genera Aphanocalyx and Tetraberlinia were supposedly closely related to or even congeneric with Monopetalanthus , so I have taken them into account as well. Since Aubréville had united Michelsonia with Tetraberlinia , that monotypic genus is dealt with as well. These genera are presently classified in the tribe Macrolobieae.All available material belonging to these four genera has been studied in the major herbaria or was received on loan. A large number of macromorphological characters were studied and evaluated for their taxonomic potential. During this study nine new species and one new subspecies have been discovered. Of the original taxa all have been maintained, although one of them has been reduced to subspecific level, resulting in a total of 33 species. For all species concerned extensive descriptions have been made with distribution maps and most of them have been illustrated. Keys to the genera and to the species have been added.<CENTER><img src="/wda/abstracts/i2686_02.gif" VSPACE="6" WIDTH="428" HEIGHT="412" ALT="Inline: Bikinia media Wieringa, type of the new genus" BORDER="0"/>Bikinia media Wieringa, type of the new genus</CENTER>In order to ascertain the phylogenetic history of these taxa and to see which characters are really indicating natural groups, a cladistic analysis was undertaken. To enhance resolution near the specific level some "minor" characters have been included as well. Characters that did not easily separate into discrete entities have also been used. They have been coded using a method similar to gap weighting. The resulting phylogeny proved to be strong on a number of points, but remains inconclusive on others. Since the stronger and most stable clades are those that are supported by several characters, these clades are the best to use for generic delimitation. When we treat the largest stable clades as genera, the species of the four old genera reshuffle into four new aggregates. Michelsonia remains a monospecific genus and Tetraberlinia hardly changes, only Monopetalanthus longiracemosus is now contained in it. Aphanocalyx engulfs the major part of Monopetalanthus , while for the remaining part of Monopetalanthus the name Bikinia is proposed. A new species from Mozambique, which seems to be related to this group, cannot be accommodated in any of these four genera and is described in the new genus Icuria . Three quite distinct species of Aphanocalyx are arranged in the new subgenus Antherodontus . The related genus Julbernardia , which was studies less comprehensive as the other taxa, proves to be a natural group when Paraberlinia is included.A small AFLP trial (DNA analysis) using only 7 ingroup samples resulted in a tree which is completely congruent with the tree based on morphological data. Moreover, the AFLP tree suggests that Bikinia is more closely related to Tetraberlinia than it is to Aphanocalyx , something which the morphological analysis was unable to prove. Ethanol extracts with different colour compounds found in various species were analyzed by chromatography (TLC). The method has proved to be useful, but our analysis was biased too much by initial problems to include the results in the cladistic analysis.The habitat of most species is dry-land evergreen forest. Only one species each of Bikinia and Tetraberlinia may be encountered in riverine forest as well, while a part of Aphanocalyx has become adapted primarily to riverine and gallery forests. Icuria comprises pure stands of coastal evergreen forest, but remains present in woodland-like vegetation after degradation by logging.Many trees of Macrolobieae have been recorded as growing gregariously. To assess such gregarious stands, the large trees in two 1-ha plots in Gabon have been mapped, and data from an existing line survey has been reanalysed. Several species of Bikinia and Tetraberlinia indeed grow in clusters. Most species of Bikinia occur in small clusters of only one or a few hectares, only B. le-testui and B. pellegrinii seem to form clusters with an area of a little less than 1 km <sup>2</SUP>. Four species of Tetraberlinia may become (co-)dominant over large areas, and such Tetraberlinia forests may extend over several square kilometres. Towards the margins of their distribution areas, the species tend to grow in smaller clusters or even singly. The gregarious nature of most Macrolobieae is considered the outcome of a rather complicated specialization process, including especially ectomycorrhizal relationships, but also factors of dispersal, germination and chemical defences. This set of ecological characters makes the species well adapted to life in a tropical evergreen forest.<CENTER><img src="/wda/abstracts/i2686_03.jpg" VSPACE="6" WIDTH="420" HEIGHT="288" ALT="Inline: Bikinia media Wieringa, type of the new genus" BORDER="0"/>Inflorescences of Tetraberlinia longiracemosa , showing yellow petals, characteristic for the genus</CENTER>The flowers of the species studied are pollinated by a variety of animals, most of them insects (like bees, beetles, flies and butterflies), but birds also seem to play a role in some species. The two species that were found to attract sunbirds both have 'strobiliform' racemes, which may be better suited for birds than 'umbelliform' or open and lax inflorescences. The 'umbelliform' inflorescences especially attract longhorn beetles. All different types of structures did attract large numbers of both medium-sized and small bees, probably the major pollinators. All species with yellow petals possess a deep hypanthium. However, an animal that appreciates these flowers is still unknown.Flowering in most of the species is restricted to a definite period, often correlated with the rainfall pattern of the area. Most Gabonese species flower just before or during the main rainy season, a few (especially Aphanocalyx heitzii and Tetraberlinia bifoliolata ) during the shorter rainy season. Only rarely was flowering recorded during dry periods.The group is considered to be quite interesting for forestry, especially for plantation purposes. The trees are adapted to poor soils, they grow quite fast and produce tall, straight cylindrical boles. Moreover, the trees grow gregariously and are probably not much subject to diseases.The distribution patterns of this group prove that one should be cautious in speaking about the location of glacial forest refuges. It depends on the requirements of a species whether a certain forested area during the ice ages could serve as a refuge or not. Within the Congo basin probably many riverine forests have existed during these periods, but they were only suited for species that could cope with inundation and other unstable conditions. Most species of Bikinia and Tetraberlinia were not adapted to these conditions, but several species of Aphanocalyx were and hence they still occur in the Congo basin.Following the newest IUCN criteria, two of the treated species are possibly extinct, four classify as critically endangered and two as endangered