Clusia blattophila sp. nov. (Clusiaceae) from an inselberg in French Guiana

Clusia blattophila (Clusiaceae) is described as a new species. It is distinguished from related species in C. sect. Oedematopus by e.g. the relatively large and reflexed petals and from other Clusia spp. by having fewer (6–16) stamens and only four petals. Unique traits are the almost sessile inflorescence and the presence of a secretory pit at the floral apex in staminate flowers. The species grows epilithically and is known only from the Nouragues inselberg, a massive granite outcrop in eastern French Guiana. The name refers to its cockroach pollinators.     

Seedling recruitment in the invasive clonal shrub, Mahonia aquifolium Pursh (Nutt.)

The clonal shrub, Mahonia aquifolium, is an aggressive invader of some forests in central Germany. We analysed the importance of seedling recruitment for the local dynamics of these invasive populations. In contrast to many other clonal plants, repeated seedling recruitment takes place in M. aquifolium, contributing to the colonization of available space within populations. Thus, sexual reproduction is not only important for dispersal to new sites but also for the local invasion␣process. In situations with grass competition (Calamagrostis epigejos) the number of successful recruits is reduced, and clonal growth is the predominant mode of reproduction. Since seedling recruitment results in an increasing number of genets as well as ramets, the investigated populations are still below their carrying capacities. Seedling densities are higher beneath the canopy of adults than outside adult patches, and increase with increasing cover of adult ramets. The spatially clumped pattern of seedling emergence results in density-dependent mortality especially in the most crowded seedling clusters. However, adult neighbourhood imposes a much higher mortality. Therefore, we expect that the importance of seedling recruitment for the dynamics of local M. aquifolium populations will decline as the invasion process continues. Received: 15 January 1996 / Accepted: 28 September 1996     

Importance of demography and dispersal for the resilience and restoration of a critically endangered tropical conifer Araucaria nemorosa

Aims In situ species survival involves persistence at current sites and/or colonization of new locations. Determining the likelihood of these events requires an understanding of population dynamics and dispersal. We address these issues in populations of a critically endangered tropical conifer Araucaria nemorosa and provide conservation recommendations. Location Port Boise, island of Grand Terre, New Caledonia, Pacific Ocean. Methods We characterize the demographic structure of six populations of A. nemorosa based upon size class frequencies and relative basal area compared to competing angiosperm trees. Using genotype data from 280 adult individuals at seven microsatellite loci, we indirectly estimate dispersal distances and project a maximal dispersal envelope around the extant populations. Results Our survey detected marked differences in demographic parameters including the proportion of basal area occupied by A. nemorosa versus angiosperm trees (ranging from 25% to 77%), the size class frequency distributions of A. nemorosa and seedling densities (ranging from 500 to 37300 seedlings ha^?1) among some populations. Wright’s genetic neighbourhood ranged from 22 to 876 trees, and historic gene dispersal ranged from 10.8 to 82.4 m, indicating that most seed dispersal is <100 m. Main conclusions This study indicates that the risk of recruitment failure in remnant populations of Araucaria nemorosa is low in most of the stands but is high in the inland population Forêt Nord. In situ natural regeneration appears to be constrained in this population, most likely as a consequence of competitive exclusion. Our results also suggest that the majority of seed dispersal is too short to allow A. nemorosa to disperse to new more hospitable sites within an ecologically relevant time‐scale. Our findings have implications not only for this emblematic tree species but also for a wide range of fragmented and degraded plant species populations with limited dispersal that are vulnerable to competitive exclusion.     

Demographic bottlenecks in tropical plant regeneration: A comparative analysis of causal influences

Mortality factors that act sequentially through the demographic transitions from seed to sapling may have critical effects on recruitment success. Understanding how habitat heterogeneity influences the causal factors that limit propagule establishment in natural populations is central to assess these demographic bottlenecks and their consequences. Bamboos often influence forest structure and dynamics and are a major factor in generating landscape complexity and habitat heterogeneity in tropical forests. To understand how patch heterogeneity influences plant recruitment we studied critical establishment stages during early recruitment of Euterpe edulis, Sloanea guianensis and Virola bicuhyba in bamboo and non-bamboo stands in the Brazilian Atlantic forest. We combined observational studies of seed rain and seedling emergence with seed addition experiments to evaluate the transition probabilities among regeneration stages within bamboo and non-bamboo stands. The relative importance of each mortality factor was evaluated by determining how the loss of propagules affected stage-specific recruitment success. Our results revealed that the seed addition treatment significantly increased seedling survivorship for all three species. E. edulis seedling survival probability increased in the addition treatment in the two stand types. However, for S. guianensis and V. bicuhyba this effect depended strongly on artificially protecting the seeds, as both species experienced increased seed and seedling losses due to post-dispersal seed predators and herbivores. Propagules of all three species had a greater probability of reaching subsequent recruitment stages when protected. The recruitment of large-seeded V. bicuhyba and E. edulis appears to be much more limited by post-dispersal factors than by dispersal limitation, whereas the small-seeded S. guianensis showed an even stronger effect of post-dispersal factors causing recruitment collapse in some situations. We demonstrated that E. edulis, S. guianensis and V. bicuhyba are especially susceptible to predation during early compared with later establishment stages and this early stage mortality can be more crucial than stand differences as determinants of successful regeneration. Among-species differences in the relative importance of dispersal vs. establishment limitation are mediated by variability in species responses to patch heterogeneity. Thus, bamboo effects on the early recruitment of non-bamboo species are patchy and species-specific, with successional bamboo patches exerting a far-reaching influence on the heterogeneity of plant species composition and abundance.     

Dispersal of remnant endangered trees in a fragmented and disturbed forest by frugivorous birds

Most endangered plant species in a fragmented forest behave as a unique source population, with a high dependence on frugivorous birds for recruitment and persistence. In this study, we combined field data of dispersal behavior of birds and GIS information of patch attributes to estimate how frugivorous birds could affect the effective dispersal pattern of Chinese yew (Taxus chinensis) in a fragmented and disturbed forest. Nine bird species were observed to visit T. chinensis trees, with Urocissa erythrorhyncha, Zoothera dauma and Picus canus being the most common dispersers. After foraging, six disperser species exhibited different perching patterns. Three specialist species, P. canus, Turdus hortulorum, and Z. dauma stayed in the source patch, while three generalist species, U. erythrorhyncha, Hypsipetes mcclellandii, and H. castanonotus, could perch in bamboo patches and varied in movement ability due to body size. As a consequence of perching, dispersers significantly contributed to the seed bank, but indirectly affected seedling recruitment. Moreover, the recruitment of T. chinensis was also affected by patch attributes in a fragmented forest (distances to source patch, patch type, size). Our results highlighted the ability of unique source population regeneration of T. chinensis in a fragmented forest, with high dependence on both frugivorous birds and patch attributes, which should be considered in future planning for forest management and conservation.     

Recruitment limitations in Primula vulgaris in a fragmented landscape

Plant recruitment is limited by dispersal, if seeds cannot arrive at potential recruitment sites, and by establishment, due to a low availability of safe sites for recruitment. Seed-sowing experiments, scarcely applied along gradients of landscape alteration, are very useful to assess these limitations. Habitat loss and fragmentation may foster recruitment limitations by affecting all the processes from seed dispersal to seedling establishment. In this study, we perform a seed-sowing experiment to disentangle the importance of dispersal and establishment limitations in different stages of recruitment of the perennial herb Primula vulgaris in fragmented forests of the Cantabrian Range (Northwestern Spain). We evaluated the influence of ecological gradients resulting from habitat loss and fragmentation (modifications of habitat amount at the landscape and microhabitat scales, changes in the species’ population size, changes in seed predation and seedling herbivory) on seedling emergence, survival and early growth. We found strong evidence of dispersal limitation, as seedling emergence was very low in experimental replicates where no seeds were added. This limitation was independent of landscape alterations, as we found no relation with any of the ecological gradients studied. Establishment limitations at the germination phase were also unrelated to ecological gradients, probably because these limitations are more related to fine-scale environmental gradients. However, further monitoring revealed that seedling survival after summer and winter periods and seedling growth were conditioned by landscape alteration, as we found effects of habitat amount at the landscape and microhabitat scales, of presence of populations of P. vulgaris and of seedling herbivory. These effects were complex and sometimes opposite to what can be expected for adult plants, revealing the presence of different requirements between life stages.     

Avian seed dispersal and seedling distribution of the endangered tree species,Taxus chinensis,in patchy habitats

Background: There is limited understanding about bird dispersal behaviour and seedling distribution of endangered tree species in patchy environments, although these processes are important for plant species persistence.

Aims: We tested how patch features affected bird behaviour and seed dispersal, and thus seedling distribution of the endangered Chinese yew tree (Taxus chinensis).

Methods: In the present study, we combined field data of bird dispersal behaviour and GIS-based information to elucidate the influence of spatial features of habitat patches on bird dispersal behaviour, and the resulting effects on the seedling distribution of the endangered Chinese yew in two patchy habitats.

Results: Our results showed that the only seed source patch could attract eight bird species for dispersal at the two sites. Post-foraging movements of bird dispersers was strongly related to both topography and the relative locations of habitat patches. Yew seedlings aggregated only at the seed source and bamboo recruitment patches, which was affect by both the spatial distribution of recruitment patches and patch use by dispersers.

Conclusions: Our results emphasise that bamboo patches in both patchy environments provide the necessary conditions for germination of yew seeds, and the post-foraging behaviour of dispersers determines seed deposited in these patches. Our study highlights the importance of the dispersal behaviour of frugivorous birds in the successful regeneration and colonisation of yew populations in patchy habitats.     

Recruitment limitation in secondary forests dominated by an exotic tree

Question: What factors limit woody plant recruitment in a mosaic landscape where former agricultural lands are dominated by the invasive tree Ligustrum lucidum (Oleaceae)? Location: Subtropical northwestern Argentina. Methods: In secondary forest patches, we measured (1) tree, shrub and liana abundance in different size classes; (2) seed rain of Ligustrum and two native trees and (3) topographic, soil and light variables. We used spatial autoregressive models to test for effects of Ligustrum dominance and environment on native plant abundance in each size class. We used multiple regression on resemblance matrices to quantify the relative importance of spatial (e.g. dispersal) and environmental effects on native species composition. Results: Native tree abundance in the smallest size class was unrelated to Ligustrum canopy dominance, while native tree abundance in larger size classes and native liana abundance were negatively correlated with Ligustrum dominance. Native species composition was both environmentally and spatially structured, suggesting that some species are dispersal limited. Seed rain was spatially correlated with conspecific basal area for one of two native species, but not for Ligustrum. Conclusions: Native tree recruitment appears to be limited primarily by sapling mortality in patches dominated by the invasive Ligustrum. Ligustrum does not appear to be dispersal limited in our study area and is likely to continue spreading. Invaded patches may persist for hundreds of years.     

The relative importance of sexual and asexual reproduction in the spread of Spartina alterniflora using a spatially explicit individual-based model

This study investigates a spatially explicit, individual-based model for simulating the spread of invasive smooth cordgrass (Spartina alterniflora) in Yancheng coastal wetlands from 1995 to 2010. The model, which considers the landscape heterogeneity and changes detected by remote sensing, also reveals the relative importance of sexual and asexual reproduction in the spread by global sensitivity analysis. The model was verified as suitable for simulating the range expansion of S. alterniflora. The results show that: (1) although seedling recruitment is low, it significantly contributes to the range expansion of S. alterniflora. Removing sexual propagation greatly reduces the expansion rate. Rapid expansion requires both sexual and asexual reproduction; (2) in the global sensitivity analysis, the most significant affecters of S. alterniflora invasion were seed dispersal distance, adult survival rate and asexual recruitment survival rate. Sexual propagation contributes much more significantly to quick range expansion than asexual reproduction, but asexual reproduction is the main source of recruitment. Invasion control strategies should target a single reproduction mode. Here, limiting the germination and dispersal of seeds is suggested as a realistic strategy for controlling and managing invasion by this species.     

Distinguishing effects of juvenile mortality and dispersal on recruitment

Detailed data on juvenile survival are rare in the literature. Although many studies estimate recruitment, if you cannot distinguish between permanent dispersal and mortality, the management implications for a population may be unclear. We estimated juvenile survival in a reintroduced North Island robin (Petroica longipes) population in a protected sanctuary surrounded by an unprotected landscape where the species is extirpated. The population has had marginal population growth due to poor recruitment so we modeled 3 types of data (resighting of fledglings, radio-telemetry of independent juveniles, resighting of adults) in an integrated framework to determine the life stages where high mortality was occurring, and to distinguish mortality from dispersal. Approximately 16% of birds that fledged (n = 109) were present at the start of the next breeding season, consistent with recruitment rates from previous years. Low survival in the first 6 weeks after fledging was the primary cause of poor recruitment. Only 50% survived to independence (4 weeks after fledging), and 18% survived to the end of the radio-tracking period (14 weeks), after which juvenile survival matched adult survival. No dispersal from the sanctuary occurred during the radio-tracking period. Juveniles moved between adjacent forest fragments within the sanctuary, but did not leave the sanctuary. This information, which demonstrates the importance of distinguishing between natal mortality and dispersal, is important for ongoing management of the site and selection of future reintroduction sites for this species. © 2019 The Wildlife Society.     

Application of mathematical models in the spatial analysis of early tree seedling distribution patterns within a treefall gap at Gwangneung Experimental Forest, Korea

A spatial analysis was conducted on the characteristics of seedling recruitment and one-year mortality of two temperate deciduous tree species: Quercus serrata Thunb. ex Murray (QUSE) and Carpinus laxiflora (Siebold & Zucc.) Blume (CALA). A systematic grid of 64 seedling plots was established within a typhoon-induced treefall gap in a rare old-growth deciduous stand at Gwangneung Experimental Forest, South Korea. Inverse model, Ripley’s K function, and semivariogram tools were introduced to examine spatial relationships among seedling recruitment, one-year seedling mortality, and trees. Overall QUSE germinant density more than tripled that of CALA. Results suggest that dispersal patterns differ between the two species: for QUSE, 2Dt model yielded the best fit whereas the lognormal model was best fitted to CALA. Those differences may reflect different regeneration strategies between the two species. One-year mortality rates were similar for QUSE (76.15%) and CALA (72.25%). Mortality for both species exhibited substantial variance that was indicated as autocorrelation by the semivariogram, suggesting that mortality was not random for either species. Tree spatial patterns were also shown to be autocorrelated. Together, the observed spatial patterns in germination and first-year mortality indicate that the single treefall gap is among the factors that maintain the community structure associated with this forest type.     

Isolation of polymorphic microsatellite loci in the genus Clusia (Clusiaceae)

Microsatellite flanking region sequences may provide phylogenetically useful information. We isolated 13 polymorphic microsatellite loci from two species, Clusia minor (five loci) and Clusia nemorosa (eight loci), to aid in the determination of phylogenetic relationships within the genus Clusia. Eleven loci amplified across all 17 Clusia species tested, while two loci amplified in 10 out of 17 species. The extensive cross‐species amplification suggests that these loci may be useful for an examination of phylogenetic relationships in this genus.     

Estimation of the relative importance of sexual and vegetative reproduction in the clonal woodland herb Anemone nemorosa

Anemone nemorosa is a perennial rhizomatous plant of European woodlands. The “probability of clonal identity” method estimated the relative proportion of sexual to vegetative reproduction in this species to be 4.4% from allozyme genotype distributions. This result is congruent with investigations on the germination, short-term demography, population genetics, and breeding system of this species, and supports the hypothesis that even low levels of seedling recruitment can maintain considerable intrapopulational genetic diversity. Received: 2 March 1998 / Accepted: 24 July 1998     

Bottlenecks for plant recruitment in woodland remnants: An ornithochorous shrub in a Mediterranean ‘relictual’ landscape

Habitat fragmentation may lead to declines in plant populations and ultimately to extinction through a disruption of demographic processes, which may result in bottlenecks or even a collapse in regeneration. Nevertheless, very few studies have assessed the net effects of habitat fragmentation on plant recruitment integrating its multiple demographic processes. Using data from a four-year field study, we analyze how habitat fragmentation affects particular demographic processes and the overall magnitude of plant recruitment. We use as a case study the bird-dispersed shrub Myrtus communis in Mediterranean woodland patches within an extremely fragmented landscape (～1% woodland cover). By means of observations and experiments, we quantified fecundity, fruit removal by frugivores, seed rain, post-dispersal seed predation by rodents and seedling emergence and survival. Within each patch, we quantified post-dispersal processes in different target microhabitats. We considered the life cycle to be a combination of consecutive life stages connected by transitional processes with specific probabilities. We calculated the overall probability of recruitment for each patch as the product of all of these probabilities. The demographic processes negatively affected by fragmentation were bird-generated seed rain and seedling emergence and survival, which were attributable, respectively, to lower fruit abundance and poorer habitat quality in the smaller patches. The negative effect of fragmentation on M. communis recruitment became stronger when all the demographic processes were integrated. Of all processes, seedling emergence and, above all, seedling survival were clearly bottlenecks for recruitment that were associated with habitat fragmentation. Results from our observations and experiments were consistent with natural patterns of regeneration given that we found higher seedling densities in larger patches and old population structures (with no saplings/juveniles) in some small patches. Our study shows that habitat fragmentation has serious negative effects on recruitment in M. communis due to demographic bottlenecks in seedling establishment. The available evidence (this and a companion study) suggests that the impoverishment of habitat quality associated with habitat fragmentation (edge effects, disturbances associated with management and microhabitat availability) can explain these results. Given that restoration at a landscape scale is likely to be extremely difficult, initial management actions should aim to improve habitat quality in the smallest woodland remnants.     

Bet‐hedging and germination in the Australian arid zone shrub Acacia ligulata

The diaspore of the Australian arid zone shrub Acacia ligulata is dispersed by birds and ants. To investigate the benefits of providing a dispersal structure attractive to both groups, we compared the germination response and viability of seeds eaten by birds, handled by ants or collected from trees to simulated precursors of germination: scarification, fire and rainfall were simulated. Seed germination and viability were related to the degree of preheating disturbance to the seed coat. Heating increased the germinability of seeds not scarified or eaten by birds. In the absence of heating, ingestion by birds increased germinability. Heating increased the mortality of seeds. Our results suggest that ingestion of seeds by birds may break seed dormancy and hence enable some seeds to germinate soon after dispersal. Alternatively, seeds not eaten by birds are likely to remain dormant until sufficiently scarified by soil or stimulated by fire. Consequently, in areas such as the Simpson Desert, A. ligulata may be able to use a range of seedling establishment ‘windows’ provided by monsoon rains, post‐fire environments and unseasonal winter rains, and also spread the risk of unsuccessful seedling establishment by retaining dormant seeds in the seedbank.     

Predicted changes in vegetation structure affect the susceptibility to invasion of bryophyte-dominated subarctic heath

Background and Aims

A meta-analysis of global change experiments in arctic tundra sites suggests that plant productivity and the cover of shrubs, grasses and dead plant material (i.e. litter) will increase and the cover of bryophytes will decrease in response to higher air temperatures. However, little is known about which effects these changes in vegetation structure will have on seedling recruitment of species and invasibility of arctic ecosystems.

Methods

A field experiment was done in a bryophyte-dominated, species-rich subarctic heath by manipulating the cover of bryophytes and litter in a factorial design. Three phases of seedling recruitment (seedling emergence, summer seedling survival, first-year recruitment) of the grass Anthoxanthum alpinum and the shrub Betula nana were analysed after they were sown into the experimental plots.

Key Results

Bryophyte and litter removal significantly increased seedling emergence of both species but the effects of manipulations of vegetation structure varied strongly for the later phases of recruitment. Summer survival and first-year recruitment were significantly higher in Anthoxanthum. Although bryophyte removal generally increased summer survival and recruitment, seedlings of Betula showed high mortality in early August on plots where bryophytes had been removed.

Conclusions

Large species-specific variation and significant effects of experimental manipulations on seedling recruitment suggest that changes in vegetation structure as a consequence of global warming will affect the abundance of grasses and shrubs, the species composition and the susceptibility to invasion of subarctic heath vegetation.     

Variability of seedling recruitment under dominant, moss, and litter removal over four years

In a four-year removal experiment we investigated the effect of removing a dominant species (Nardus stricta), the litter layer and moss layer on species composition of established vegetation and on numbers and species composition of seedlings. The experiment was conducted in an oligotrophic wet meadow (Molinion with some features ofViolion caninae according to phytosociological classification). After four years of the experiment, the redundancy analysis (RDA) did not reveal any changes in the composition of the established plant community. Seedling numbers and their species composition, however, varied significantly both among treatments and over years. Whereas the differences between treatments were mainly in total seedling numbers, the years differed in both seedling numbers and relative representation of individual species. In particular,Myosotis nemorosa seedlings highly increased their abundance in a favourable year 1995 constituting one third of all the seedlings found, whereas in other years their abundance was between 4% and 10%. The effects of year and treatment were not additive (significant year × treatment interaction). This means that plots under different treatments are differently affected by climatic conditions of individual years. The greatest increase of seedling recruitment was observed in the plots with the moss layer removed. Microsites with high seedling recruitment persisted in the plots for no more than two years. Why the established vegetation composition does not differ among treatments even after four years, despite the relatively pronounced and immediate response of seedlings? The number of species with regular seedling recruitment is limited. They are mostly perennials, and consequently, it might take even longer than four years before recruitment limitation affects the population size of established plants.     

Effects of seed dispersal on Juniperus communis recruitment on a Mediterranean mountain

Abstract. The recruitment of the relict shrub Juniperus communis on a mountain in SE Spain was studied during the period 1994–1998. The main objective was to determine both the quantitative and qualitative effects of bird dispersal on seedling establishment. Seed removal by birds, seed rain, post‐dispersal seed predation, germination, and seedling emergence and survival were analysed in different microhabitats. Birds removed 53 ‐ 89% of the seeds produced by plants. Seed rain was spatially irregular as most seeds accumulated near stones used by birds as perches and below mother plants while a few seeds were dropped in wet meadows and open ground areas. Post‐dispersal seed predation by rodents affected < 10% of dispersed seeds but varied significantly among microhabitats. Only 3.6 ‐ 5.5% of dispersed seeds appeared viable, as many seeds had aborted or showed wasp damage. Seeds germinated in the second and third springs after sowing, reaching a germination percentage of 36%. Seedling emergence was concentrated in wet meadows. Seedling mortality was high (75–80%), but significantly lower in wet meadows, the only microhabitat where seedlings could escape from summer drought, the main mortality cause. Seed abortion, germination and seedling mortality proved to be the main regeneration constraints of J. communis on Mediterranean mountains. Birds exerted a strong demographic effect, although their qualitative effect was limited by abiotic factors which caused the pattern of seed rain to differ from the final pattern of recruitment between microhabitats.     

Effects of insect herbivory on induced chemical defences and compensation during early plant development in Penstemon virgatus

Background and Aims

The lack of studies assessing the simultaneous expression of tolerance and resistance traits during seedling development and overall seedling defences as compared with adult plants, in general, constitutes a significant research need that can greatly improve our understanding of overall investment in defences during plant ontogeny.

Methods

Using two seedling and two juvenile stages of the perennial herb Penstemon virgatus (Plantaginaceae) evaluations were made of (a) patterns of investment in constitutive chemical defences [i.e. iridoid glycosides (IGs)], and (b) simultaneous variation in the short-term ability of seedling and juvenile stages to induce resistance traits, measured as induced chemical defences, or tolerance traits, measured as compensatory re-growth following moderate levels of damage by a specialist insect herbivore.

Key Results

Plants were highly defended during most of their transition from seedling to early juvenile stages, reaching a constant approx. 20 % dry weight total IGs. Furthermore, following 30 % above-ground tissue damage, seedlings and juvenile stages were equally able to induce resistance, by raising their IG concentration by approx. 8 %, whereas compensatory re-growth was only achieved at young juvenile but not seedling stages.

Conclusions

Two major trends emerged from this study: (1) in contrast to expected and previously observed trends, in this perennial plant species, seedlings seem to be one of the most well-defended stages as compared with adult ones; (2) high levels of constitutive defences did not limit the ability of young developmental stages to induce resistance following damage, although this response may come with a cost (i.e. decreased compensation) in young seedling stages. Hence, as has been previously demonstrated in few other systems, these results points towards an indirect evidence for a trade-off between tolerance and resistance traits at some, but not all, developmental stages; making them often difficult to detect.     

The inselberg flora of Atlantic Central Africa. I. Determinants of species assemblages

Aims To identify the relative contributions of environmental determinism, dispersal limitation and historical factors in the spatial structure of the floristic data of inselbergs at the local and regional scales, and to test if the extent of species spatial aggregation is related to dispersal abilities. Location Rain forest inselbergs of Equatorial Guinea, northern Gabon and southern Cameroon (western central Africa). Methods We use phytosociological relevés and herbarium collections obtained from 27 inselbergs using a stratified sampling scheme considering six plant formations. Data analysis focused on Rubiaceae, Orchidaceae, Melastomataceae, Poaceae, Commelinaceae, Acanthaceae, Begoniaceae and Pteridophytes. Data were investigated using ordination methods (detrended correspondence analysis, DCA; canonical correspondence analysis, CCA), Sørensen's coefficient of similarity and spatial autocorrelation statistics. Comparisons were made at the local and regional scales using ordinations of life‐form spectra and ordinations of species data. Results At the local scale, the forest‐inselberg ecotone is the main gradient structuring the floristic data. At the regional scale, this is still the main gradient in the ordination of life‐form spectra, but other factors become predominant in analyses of species assemblages. CCA identified three environmental variables explaining a significant part of the variation in floristic data. Spatial autocorrelation analyses showed that both the flora and the environmental factors are spatially autocorrelated: the similarity of species compositions within plant formations decreasing approximately linearly with the logarithm of the spatial distance. The extent of species distribution was correlated with their a priori dispersal abilities as assessed by their diaspore types. Main conclusions At a local scale, species composition is best explained by a continuous cline of edaphic conditions along the forest‐inselberg ecotone, generating a wide array of ecological niches. At a regional scale, these ecological niches are occupied by different species depending on the available local species pool. These subregional species pools probably result from varying environmental conditions, dispersal limitation and the history of past vegetation changes due to climatic fluctuations.