Search

Pollination is a key step of plant reproduction, allowing individual plants to produce offspring as father, mother or both. However, few studies exist that consider together male and female pollination success. This implies studying both mating system, through paternity analyses, and seed set, by measuring the percentage of flowers giving a seed. Studying these two processes together is needed as they are not independent: gaining fitness advantage through one sex can incur fitness costs through the other due to various tradeoffs including direct sexual interference. Hence, we developed the first spatially explicit mixed-mating model integrating these two interactive processes, by coupling a mating model with a fruit set model, therefore jointly exploring pollen export and import. We used as model an insect-pollinated tree species, chestnut. We carried out a paternity analysis based on nearly exhaustive sampling of potential pollen donors in an intensively studied plot of 273 trees belonging to three interfertile chestnut species and including both male-fertile and male-sterile individuals. We collected a large dataset of 1924 mating events. We further performed fruit set measurements for 216 trees. Our process-based model predicts fruit set with great accuracy, but only if we account for self-pollen interference and associated ovule discounting, a form of sexual interference. This model represents an important step forward for fundamental pollination studies aiming at comprehensively exploring pollen emission, transport and reception in a single study, thus clarifying the consequences of pollination on male and female fitness.

International audience; Extant rear-edge populations located in former glacial refugia remain understudied despite their high conservation value. These populations should have experienced strong genetic drift due to their small size and long isolation. Moreover, the prolonged action of isolation by distance in refugial areas should result in stronger regional spatial genetic structure (SGS) than in recolonized areas, but empirical tests of this prediction are scarce. To fill this gap, we first used a set of 16 microsatellite markers to investigate the genetic structure of European beech in France in 65 populations from three refugial areas and one control recolonized (nonrefugial) area. Then, using the same approach, we reanalysed published isozyme data from 375 populations distributed across the entire species range. We found stronger genetic differentiation among populations in refugia than in recolonized areas. However, contrary to expectations, regional SGS was lower within refugia than within recolonized areas. Published studies presenting similar analyses suggest that our results could have generality across different biogeographical settings and types of organisms. Strong and prolonged genetic drift in refugial areas could have erased the signature of range expansions that is still visible in recolonized areas. Our results therefore suggest that Pleistocene population isolation has played a key role in increasing the genetic complexity of extant rear-edge populations.

Studying geographic variation in the rate of hybridization between closely related species could provide a useful window on the evolution of reproductive isolation. Reinforcement theory predicts greater prezygotic isolation in areas of prolonged contact between recently diverged species than in areas of recent contact, which implies that old contact zones would be dominated by parental phenotypes with few hybrids (bimodal hybrid zones), whereas recent contact zones would be characterized by hybrid swarms (unimodal hybrid zones). Here, we investigate how the hybrid zones of two closely related Chinese oaks, Quercus mongolica and Q. liaotungensis, are structured geographically using both nuclear and chloroplast markers. We found that populations of Q. liaotungensis located around the Changbai Mountains in Northeast China, an inferred glacial refugium, were introgressed by genes from Q. mongolica, suggesting historical contact between the two species in this region. However, these introgressed populations form sharp bimodal hybrid zones with Q. mongolica. In contrast, populations of Q. liaotungensis located in North China, which show no sign of ancient introgression with Q. mongolica, form unimodal hybrid zones with Q. mongolica. These results are consistent with the hypothesis that selection against hybrids has had sufficient time to reinforce the reproductive barriers between Q. liaotungensis and Q. mongolica in Northeast China but not in North China.

Phylogeography allows the inference of evolutionary processes that have shaped the current distribution of genealogical lineages across a landscape. In this perspective, comparative phylogeographical analyses are useful in detecting common historical patterns by either comparing different species within the same area within a continent or by comparing similar species in different areas. Here, we analyse one taxon (the white oak, genus Quercus , subgenus Quercus , section Quercus ) that is widespread worldwide, and we evaluate its phylogeographical pattern on two different continents: western North America and Western Europe. The goals of the present study are: (i) to compare the chloroplast genetic diversity found in one California oak species vs. that found in the extensively studied European oak species (in France and the Iberian Peninsula); (ii) to contrast the geographical structure of haplotypes between these two taxa and test for a phylogeographical structure for the California species. For this purpose, we used the same six maternally inherited chloroplast microsatellite markers and a similar sampling strategy. The haplotype diversity within site as well as the differentiation among sites was alike in both taxa, but the Californian species has higher allelic richness with a greater number of haplotypes (39 vs. 11 in the European white oak complex). Furthermore, in California these 39 haplotypes are distributed locally in patches while in the European oaks haplotypes are distributed into lineages partitioned longitudinally. These contrasted patterns could indicate that gene movement in California oak populations have been more stable in response to past climatic and geological events, in contrast to their European counterparts.

Rainforest tree species can be difficult to identify outside of their period of reproduction. Vascular tissues from Carapa spp. individuals were collected during a short field trip in French Guiana and analysed in the laboratory with nuclear and chloroplast markers. Using a Bayesian approach, > 90% of the samples could be assigned to one of two distinct clusters corresponding to previously described species, making it possible to estimate the genetic structure of each species and to identify cases of introgression. We argue that this blind procedure represents a first-choice rather than a fallback option whenever related taxa are investigated.

(sessile oak) and concludes that the lowgenetic differentiation among these species results fromshared ancestry rather than from high rates of gene flow.Although we agree that the two oaks might share allelesdue to common ancestry, we consider that the classicalhypothesis of weak interspecific barriers is more parsi-monious to explain the limited interspecific divergenceobserved at most loci. We further argue that these oaksrepresent one of the best examples of selection operatingon a subset of loci to effectively maintain species integritydespite high promiscuity. Our arguments are supportedby an abundant literature published in both specialized andmore general journals, a re-analysis and re-interpretationof Muir & Schlotterer’s own data, and recent data ona genomic scan for species differentiation in oaks (Scotti-Saintagne

Mutation rates at microsatellites tend to increase with the number of repeats of the motif, leading to higher levels of polymorphism at long microsatellites. To standardize levels of diversity when microsatellites differ in size, we investigate the relationship between tract length and variation and provide a formula to adjust allelic richness to a fixed mean number of repeats in the specific case of chloroplast microsatellites. A comparison between 39 loci from eight species of conifers (where chloroplast DNA is paternally inherited) and 64 loci from 12 species of angiosperms (where chloroplast DNA is generally predominantly maternally inherited) indicates that the greater allelic richness found in conifers remains significant after controlling for number of repeats. The approach stresses the advantage of reporting variation in number of repeats instead of relative fragment sizes.

Quercus rubra is one of the most important timber and ornamental tree species from eastern North America. It is a widespread species growing under variable ecological conditions. Chloroplast DNA variation was studied by PCR-RFLP (polymerase chain reaction-restriction fragment length polymorphism) in 290 individuals from 66 populations sampled throughout the natural range. A total of 12 haplotypes were detected, with one found in 75% of the trees. Population differentiation is relatively low ( G ST = 0.46), even when similarities between haplotypes are taken into account ( N ST = 0.50), pointing to a weak phylogeographical structure. Furthermore, no spatial structure of genetic diversity could be detected. The genetic differentiation increased northwards, reflecting the postglacial history of Q. rubra . The unusual aspect of this study was the low level of chloroplast DNA genetic differentiation in Q. rubra compared to that typically observed in other oak species. Palynological evidence indicates that during the last glacial maximum, Q. rubra had one major distribution range with populations located relatively far to the north, resulting in only modest movement northwards when climate improved, whereas European white oaks were largely restricted to the southern European peninsulas and experienced extensive movements during the postglacial period. The contrasted geographical features and levels of tree species richness of both continents might further explain why congeneric species sharing similar life history traits have genetic structures that are so different.

Plants offer excellent models to investigate how gene flow shapes the organization of genetic diversity. Their three genomes can have different modes of transmission and will hence experience varying levels of gene flow. We have compiled studies of genetic structure based on chloroplast DNA (cpDNA), mitochondrial DNA (mtDNA) and nuclear markers in seed plants. Based on a data set of 183 species belonging to 103 genera and 52 families, we show that the precision of estimates of genetic differentiation ( G ST ) used to infer gene flow is mostly constrained by the sampling of populations. Mode of inheritance appears to have a major effect on G ST . Maternally inherited genomes experience considerably more subdivision (median value of 0.67) than paternally or biparentally inherited genomes ( ∼ 0.10). G ST at cpDNA and mtDNA markers covary narrowly when both genomes are maternally inherited, whereas G ST at paternally and biparentally inherited markers also covary positively but more loosely and G ST at maternally inherited markers are largely independent of values based on nuclear markers. A model-based gross estimate suggests that, at the rangewide scale, historical levels of pollen flow are generally at least an order of magnitude larger than levels of seed flow (median of the pollen-to-seed migration ratio: 17) and that pollen and seed gene flow vary independently across species. Finally, we show that measures of subdivision that take into account the degree of similarity between haplotypes ( N ST or R ST ) make better use of the information inherent in haplotype data than standard measures based on allele frequencies only.

We studied the phylogeography of alder buckthorn (Frangula alnus), a bird-dispersed shrub or small tree distributed over most of Europe and West Asia and present in three of the four main refugia of West Palaearctic temperate woody plants: the Iberian Peninsula, the Balkans and Anatolia. A total of 78 populations from 21 countries were analysed for chloroplast DNA variation using polymerase chain reaction–restriction fragment length polymorphism (PCR–RFLP), and 21 different haplotypes were distinguished. We found a very strong overall population differentiation (GST = 0.81) and phylogeographical struc- ture, and a sharp contrast between the haplotype-rich refugia and the almost completely uniform area of postglacial colonization. The haplotype network comprises three lineages made up of haplotypes from the Iberian Peninsula, Anatolia with the Caucasus, and tem- perate Europe. The Iberian and the Anatolian branches represent parts of a major lineage that spans over the whole northern Mediterranean Basin and some neighbouring areas and probably dates back to the Tertiary. Many haplotypes of this lineage are distributed locally and most populations are fixed for a single haplotype; these populations have apparently been very stable since their establishment, experiencing negligible gene flow and few mutations. The temperate European lineage consists of one very widespread and abundant plus six locally distributed haplotypes. Four of them are located in Southeast Europe, the putative refugium of all extant temperate European populations. Contrary to populations from Iberia and Anatolia, F. alnus populations from the southeastern European refugium have most genetic variation within populations. Bird-mediated seed dispersal has apparently allowed not only a very rapid postglacial expansion of F. alnus but also subsequent regular seed exchanges between populations of the largely continuous species range in temperate Europe. In contrast, the disjunct F. alnus populations persisting in Mediterranean mountain ranges seem to have experienced little gene flow and have therefore accumulated a high degree of differentiation, even at short distances. Populations from the southern parts of the glacial refugia have contributed little to the postglacial recolonization of Europe, but their long- term historical continuity has allowed them to maintain a unique store of genetic variation

Range-wide variation of maritime pine was studied at maternally inherited and paternally inherited markers (mitochondrial DNA and chloroplast DNA). While chloroplast DNA exhibits the highest diversity, phylogeographic inferences from this marker are blurred by homoplasy and extensive pollen flow. In contrast, the only three mitochondrial haplotypes found provide a clear picture of nonoverlapping areas colonized from different refugia, with no single population having a mixed composition ( G ST = 1). Comparison of the genetic structure inferred from both organelle genomes allows the investigation of differential seed and pollen dispersal, pointing to pollen, but not seed, dispersal across the Strait of Gibraltar (from Morocco into Iberia). A comparison with already available genetic information, especially that of one of the maritime pine’s most threatening insect pests, the bast scale Matsucoccus feytaudi , further completes the picture.

We studied the phylogeography of ivy (Hedera sp.), a liana widespread in Europe, throughout its natural range. The populations sampled belong to four closely related species differing by ploidy levels and morphological characters. Chloroplast (cp) markers were used and 13 haplotypes were detected, usually shared across species, contrary to ribosomal internal transcribed spacer (ITS) variants. We demonstrated the existence of a strong overall cpDNA phylogeographical structure. Several methods of data analysis were conducted to describe how this structure and the genetic diversity change through space and time. Southern populations, especially those from Spain, are the most divergent. Pairwise estimates of differentiation point to isolation by distance, and the existence of a latitudinal gradient of divergence was demonstrated using a regression procedure. Similarly, latitudinal differences in haplotype richness and diversity exist, as shown by population permutations ('differentiation through space'). Finally, we measured differentiation by taking into account successive levels of divergence between haplotypes ('differentiation through time'). Genetic differentiation turns out to be much greater when differences between closely related haplotypes are not considered. Further, these results indicate that the phylogeographical structure is essentially due to the relative distribution of the most similar haplotypes. Diversity decreases from south to north, whereas haplotype frequencies change longitudinally. It appears that Hedera survived in Spanish and Balkan refugia during the last ice age. A third refugium must have been present in the Alps or in Italy. During the northward expansion, the decrease in overall diversity was attenuated by some mixing of lineages at intermediate latitudes, resulting in comparatively higher levels of differentiation in the south.

Chloroplast (cp) and mitochondrial (mt) DNA variation were studied in 97 populations of cork oak (Quercus suber) in Morocco; in 31 of these populations, holm oak (Quercus ilex), a clearly distinct species, also occurred and was compared with Q. suber. Three cpDNA and one mtDNA primer pairs were used in the survey, each in combination with one restriction enzyme. Six haplotypes belonging to two very divergent lineages were detected; one lineage predominates in each species, and is probably ancestral, as inferred from comparisons with other oak species. In the mixed-species populations, cytoplasmic genomes were frequently shared across species, as indicated by an introgression ratio of 0.63. This index is a new measure of the propensity of species to share locally genetic markers, varying from zero (complete differentiation) to one (no differentiation). By contrast, more closely related deciduous oak species (Q. robur, Q. petraea and Q. pubescens) have introgression ratios varying from 0.82 to 0.97. The introgression events appear to have been more frequent in the direction Q. ilex (female) x Q. suber (male), a finding which seems attributable to the flowering phenology of these two species. This asymmetry may have favoured immigration of Q. suber beyond its main range, in regions already colonized by Q. ilex. There, rare hybridization and further introgression through long distance pollen flow have established populations that are morphologically indistinguishable from Q. suber but that have cytoplasmic genomes originating from the local Q. ilex populations.

Polymorphisms in the chloroplast genome of the argan tree (Sapotaceae), an endemic species of south-western Morocco, have been detected by restriction site studies of PCR-amplified fragments. A total of 12 chloroplast DNA (cpDNA) and two mitochondrial DNA (mtDNA) fragments were amplified and digested with a single restriction enzyme (HinfI). Polymorphisms were identified in six of the cpDNA fragments, whereas no mtDNA polymorphisms were detected in a survey of 95 individuals from 19 populations encompassing most of the natural range of the species. The cpDNA polymorphisms allowed the identification of 11 haplotypes. Two lineages, one in the south-east and the other in the north-west, divide the range of the argan tree into two distinct areas. The level of genetic differentiation measured at the haplotype level (GSTc = 0.60) (i.e. with unordered haplotypes) was smaller than when phylogenetic relationships were taken into account (NSTc = 0.71-0.74) (ordered haplotypes), indicating that population history must be considered in the study of the geographical distribution of cpDNA lineages in this species. If contrasted with the level of nuclear genetic differentiation measured in a previous study with isozymes (GSTn = 0.25), the results indicate a relatively high level of gene flow by seeds, or conversely a relatively low level of gene flow by pollen, as compared with other tree species. Goats and camels could have played an important role in disseminating the fruits of this tree.

International audience; Natural hybridization is attracting much interest in modern speciation and conservation biology studies, but the underlying mechanisms remain poorly understood. In particular, it is unclear why environmental changes often increase hybridization rates. To study this question, we surveyed mating events in a mixed oak stand and developed a spatially explicit individual-based hybridization model. This model, where hybridization is frequency-dependent, pollen is nonlimiting and which allows immigrant pollen to compete with local pollen, takes into account species-specific pollen dispersal and sexual barriers to hybridization. The consequences of pollen limitation on hybridization were studied using another simple model. The results indicate that environmental changes could increase hybridization rates through two distinct mechanisms. First, by disrupting the spatial organization of communities, they should decrease the proportion of conspecific pollen available for mating, thus increasing hybridization rates. Second, by decreasing the density of conspecifics, they should increase pollen limitation and thus hybridization rates, as a consequence of chance pollination predominating over deterministic pollen competition. Altogether, our results point to a need for considering hybridization events at the appropriate level of organization and provide new insights into why hybridization rates generally increase in disturbed environments.

Studying geographic variation in the rate of hybridization between closely related species could provide a useful window on the evolution of reproductive isolation. Reinforcement theory predicts greater prezygotic isolation in areas of prolonged contact between recently diverged species than in areas of recent contact, which implies that old contact zones would be dominated by parental phenotypes with few hybrids (bimodal hybrid zones), whereas recent contact zones would be characterized by hybrid swarms (unimodal hybrid zones). Here, we investigate how the hybrid zones of two closely related Chinese oaks, Quercus mongolica and Q. liaotungensis, are structured geographically using both nuclear and chloroplast markers. We found that populations of Q. liaotungensis located around the Changbai Mountains in Northeast China, an inferred glacial refugium, were introgressed by genes from Q. mongolica, suggesting historical contact between the two species in this region. However, these introgressed populations form sharp bimodal hybrid zones with Q. mongolica. In contrast, populations of Q. liaotungensis located in North China, which show no sign of ancient introgression with Q. mongolica, form unimodal hybrid zones with Q. mongolica. These results are consistent with the hypothesis that selection against hybrids has had sufficient time to reinforce the reproductive barriers between Q. liaotungensis and Q. mongolica in Northeast China but not in North China.

International audience; Plant population genetic surveys are starting to take full advantage of technological advances in genotyping methods and of methodological advances in demographic inference. In this issue of Molecular Ecology, Keller et al. (2010) illustrate this trend with a particularly convincing study of rangewide genetic variation in a North American poplar, using both SNP and sequence data. They first investigate population genetic structure by estimating the most likely number of genetic clusters using a more formal approach than most other studies to date. They proceed by estimating gene flow among the inferred populations and by testing predictions on the distribution of low frequency alleles derived from recent work on range expansions.

Changes in species distributions after the last glacial maximum (c. 18 000 years bp) are beginning to be understood, but information diminishes quickly as one moves further back in time. In this issue of Molecular Ecology, Magri et al. (2007) present the fascinating case of a Mediterranean tree species whose populations preserve the genetic imprints of plate tectonic events that took place between 25 million years and 15 million years ago. The study provides a unique insight into the pace of evolution of trees, which, despite interspecific gene flow, can retain a cohesive species identity over timescales long enough to allow the diversification of entire plant and animal genera.

Rainforest tree species can be difficult to identify outside of their period of reproduction. Vascular tissues from Carapa spp. individuals were collected during a short field trip in French Guiana and analysed in the laboratory with nuclear and chloroplast markers. Using a Bayesian approach,90% of the samples could be assigned to one of two distinct clusters corresponding to previously described species, making it possible to estimate the genetic structure of each species and to identify cases of introgression. We argue that this blind procedure represents a first-choice rather than a fallback option whenever related taxa are investigated.

Mutation rates at microsatellites tend to increase with the number of repeats of the motif, leading to higher levels of polymorphism at long microsatellites. To standardize levels of diversity when microsatellites differ in size, we investigate the relationship between tract length and variation and provide a formula to adjust allelic richness to a fixed mean number of repeats in the specific case of chloroplast microsatellites. A comparison between 39 loci from eight species of conifers (where chloroplast DNA is paternally inherited) and 64 loci from 12 species of angiosperms (where chloroplast DNA is generally predominantly maternally inherited) indicates that the greater allelic richness found in conifers remains significant after controlling for number of repeats. The approach stresses the advantage of reporting variation in number of repeats instead of relative fragment sizes.

Plants offer excellent models to investigate how gene flow shapes the organization of genetic diversity. Their three genomes can have different modes of transmission and will hence experience varying levels of gene flow. We have compiled studies of genetic structure based on chloroplast DNA (cpDNA), mitochondrial DNA (mtDNA) and nuclear markers in seed plants. Based on a data set of 183 species belonging to 103 genera and 52 families, we show that the precision of estimates of genetic differentiation (G(ST)) used to infer gene flow is mostly constrained by the sampling of populations. Mode of inheritance appears to have a major effect on G(ST). Maternally inherited genomes experience considerably more subdivision (median value of 0.67) than paternally or biparentally inherited genomes (approximately 0.10). G(ST) at cpDNA and mtDNA markers covary narrowly when both genomes are maternally inherited, whereas G(ST) at paternally and biparentally inherited markers also covary positively but more loosely and G(ST) at maternally inherited markers are largely independent of values based on nuclear markers. A model-based gross estimate suggests that, at the rangewide scale, historical levels of pollen flow are generally at least an order of magnitude larger than levels of seed flow (median of the pollen-to-seed migration ratio: 17) and that pollen and seed gene flow vary independently across species. Finally, we show that measures of subdivision that take into account the degree of similarity between haplotypes (N(ST) or R(ST)) make better use of the information inherent in haplotype data than standard measures based on allele frequencies only.

New methods for better identification of timber geographical origin would constitute an important technical element in the forest industry, for phytosanitary certification procedures or in the chain of custody developed for the certification of timber from sustainably managed forests. In the case of the European white oaks, a detailed reference map of chloroplast (cp) DNA variation across the range exists, and we propose here to use the strong geographical structure, characterized by a differentiation of western vs. eastern populations, for the purpose of oak wood traceability. We first developed cpDNA markers permitting the characterization of haplotype on degraded DNA obtained from wood samples. The techniques were subsequently validated by confirming the full correspondence between genotypes obtained from living tissues (buds) and from wood collected from the same individual oak. Finally, a statistical procedure was used to test if the haplotype composition of a lot of wood samples is consistent with its presumed geographical origin. Clearly, the technique cannot permit the unambiguous identification of wood products of unknown origin but can be used to check the conformity of genetic composition of wood samples with the region of alleged origin. This could lead to major applications not only in the forest industry but also in archaeology or in palaeobotany.

Times Cited: 29; International audience; A polymorphic noncoding region of chloroplast DNA (cpDNA) was successfully amplified by the polymerase chain reaction (PCR) from various oak wood samples, including recent and more ancient (about 600-years-old) samples from different oak species. Adaptation of DNA isolation and amplification protocols was necessary to obtain this result. Polymorphisms useful to distinguish species or geographical origin of these samples could be scored through sequencing. These polymorphisms include one substitution and two microsatellite- type polymorphisms, due to a variable number of A/T repeats, identical results were obtained independently in two separate laboratories.

Recently, there has been a growing interest in the use of the chloroplast genome (cpDNA) in population genetics studies of plants. Indeed, despite its low substitution rate (Wolfe et al. 1987), its typically uniparental mode of transmission, and hence clonal mode of evolution is a unique feature of cytoplasmic genomes of great interest for evolution studies. The plant mitochondrial genome (mtDNA) shares these characteristics, with however, an even slower substitution rate, but is much less studied because it varies enormously in size and gene arrangement (Palmer 1992). Consensus primers which are homologous to the most conserved coding regions of cpDNA or mtDNA but amplify the more variable noncoding regions are likely therefore to be very useful. Here, we greatly enlarge a previous list of primers published in this journal (Demesure et al. 1995). A total of seven pairs of cpDNA primers and nine pairs of mtDNA primers are described. Moreover, the degree of conservation of the whole set (16 and 12 pairs of cpDNA and mtDNA primers, respectively) is tested by amplification on five species, three angiosperms (two dicotyledons and one monocotyledon), a pine (gymnosperm) and a fern (pteridophyte), i.e. on a broader taxonomic range than in the former study. The primers are also evaluated by comparison with sequences available in the molecular databases. The primers were designed in order to be as consensual as possible, although the chloroplast genome of tobacco (i.e. a dicotyledon) was used as reference, and full homology with dicotyledon sequences was also given priority in the case of the mtDNA primers. Their design was facilitated by the availability of an increasing number of complete sequences in the molecular databases. Conserved coding sequences flanking the regions to be amplified were identified with the computer software BLAST (Altschul et al. 1990). The cpDNA primers are all located in the relatively variable large single-copy region (Palmer 1985). For the mtDNA primers, introns of genes available in the molecu

International audience

Nested clade phylogeographic analysis (NCPA) has become a popular method for reconstructing the history of populations across species ranges. Ever since its invention in 1995, criticisms have been formulated, but the method, which has been regularly updated, continues to attract investigators. Molecular Ecology has published a large fraction of the literature on the topic — both pro and con. A recent study by Panchal and Beaumont (2007) finally allows a precise evaluation of the method by developing software that automates the somewhat complicated NCPA procedure. Using simulations of random-mating populations, Panchal and Beaumont find a high frequency of false-positives with their automated NCPA procedure (over 75%). These findings, which echo and amplify earlier warnings, appear serious enough to suggest to researchers to await further evaluation of the method. Although no other allencompassing method such as the NCAP currently exists to evaluate phylogeographic data sets, researchers have many alternative methods to test ever more refined hypotheses.

One of the dominant savannah colonists in Gabon is Aucoumea klaineana or Okoume (Burseraceae), an endemic species which belongs to a monotypic genus. Chloroplast DNA (cpDNA) variation was studied in this species by means of PCR amplification of 40 kb of cpDNA sequences, followed by restriction analysis of the resulting fragments. No insertion/deletion events were noted, and a single point mutation was found. The level of differentiation among the 19 populations studied was relatively low ( G ST = 0.54) compared to other plant species (mean of 0.76), in agreement with the pioneer status of the species. However, cpDNA diversity was geographically structured, with the less frequent haplotype occurring only in populations from southern Gabon. This distribution might suggest either that there were two ancient source populations of Okoume, one in the north and the other in the south, from which the colonizing process of the savannah began after the last ice age, or alternatively that there was one polymorphic source in the south. The low level of cpDNA diversity could indicate that Okoume populations in these refugia were quite small.