The degree to which variation in plant community composition (beta-diversity) is predictable from environmental variation, relative to other spatial processes, is of considerable current interest. We addressed this question in Costa Rican rain forest pteridophytes (1,045 plots, 127 species). We also tested the effect of data quality on the results, which has largely been overlooked in earlier studies. To do so, we compared two alternative spatial models [polynomial vs. principal coordinates of neighbour matrices (PCNM)] and ten alternative environmental models (all available environmental variables vs. four subsets, and including their polynomials vs. not). Of the environmental data types, soil chemistry contributed most to explaining pteridophyte community variation, followed in decreasing order of contribution by topography, soil type and forest structure. Environmentally explained variation increased moderately when polynomials of the environmental variables were included. Spatially explained variation increased substantially when the multi-scale PCNM spatial model was used instead of the traditional, broad-scale polynomial spatial model. The best model combination (PCNM spatial model and full environmental model including polynomials) explained 32% of pteridophyte community variation, after correcting for the number of sampling sites and explanatory variables. Overall evidence for environmental control of beta-diversity was strong, and the main floristic gradients detected were correlated with environmental variation at all scales encompassed by the study (c. 100-2,000 m). Depending on model choice, however, total explained variation differed more than fourfold, and the apparent relative importance of space and environment could be reversed. Therefore, we advocate a broader recognition of the impacts that data quality has on analysis results. A general understanding of the relative contributions of spatial and environmental processes to species distributions and beta-diversity requires that methodological artefacts are separated from real ecological differences.

Summary1 Field studies to evaluate the roles of environmental variation and random dispersal in explaining variation in the floristic composition of rain forest plants at landscape to regional scales have yet to reach a consensus. Moreover, only one study has focused on scales below 10 km 2 , where the effects of dispersal limitation are expected to be easiest to observe. 2 In the present study, we estimate the importance of differences in some key environmental variables (describing canopy openness, soils and topography) relative to the geographical distances between sample plots as determinants of differences in pteridophyte (ferns and fern allies) species composition between plots within a c . 5.7 km 2 lowland rain forest site in Costa Rica. 3 To assess the relative importance of environmental vs. geographical distances in relation to the length of environmental gradient covered, we compared the results obtained over the full range of soil types, including swamps, with those for upland soils alone. 4 Environmental variability was found to be a far stronger predictor of changes in floristic differences than the geographical distance between sample plots. In particular, differences in soil nutrient content, drainage and canopy openness correlated with floristic differences. 5 The decline in mean floristic similarity with increasing geographical distance was stronger than proposed by the random dispersal model over short distances (up to c . 100 m), which is probably attributable to both dispersal limitation and environmental changes. The scatter around the mean was large at all distances. 6 Our initial expectation was that the effects of dispersal limitation (represented by geographical distance) on observed patterns of floristic similarity would be stronger, and those of environmental differences weaker, than at broader spatial scales. Instead, these results suggest that the niche assembly view is a more accurate representation of pteridophyte communities at local to mesoscales than the dispersal assembly view.

Aim A major problem for conservation in Amazonia is that species distribution maps are inaccurate. Consequently, conservation planning needs to be based on other information sources such as vegetation and soil maps, which are also inaccurate. We propose and test the use of biotic data on a common and relatively easily inventoried group of plants to infer environmental conditions that can be used to improve maps of floristic patterns for plants in general. Location Brazilian Amazonia. Methods We sampled 326 plots of 250 m × 2 m separated by distances of 1–1800 km. Terrestrial fern individuals were identified and counted. Edaphic data were obtained from soil samples and analysed for cation concentration and texture. Climatic data were obtained from Worldclim. We used a multivariate regression tree to evaluate the hierarchical importance of soils and climate for fern communities and identified significant indicator species for the resultant classification. We then tested how well the edaphic properties of the plots could be predicted on the basis of their floristic composition using two calibration methods, weighted averaging and k‐nearest neighbour estimation. Results Soil cation concentration emerged as the most important variable in the regression tree, whereas soil textural and climatic variation played secondary roles. Almost all the plot classes had several fern species with high indicator values for that class. Soil cation concentration was also the variable most accurately predicted on the basis of fern community composition (R2 = 0.65–0.75 for log‐transformed data). Predictive accuracy varied little among the calibration methods, and was not improved by the use of abundance data instead of presence–absence data. Conclusions Fern species composition can be used as an indicator of soil cation concentration, which can be expected to be relevant also for other components of rain forests. Presence–absence data are adequate for this purpose, which makes the collecting of additional data potentially very rapid. Comparison with earlier studies suggests that edaphic preferences of fern species have good transferability across geographical regions within lowland Amazonia. Therefore, species and environmental data sets already available in the Amazon region represent a good starting point for generating better environmental and floristic maps for conservation planning.

Summary1. Plant species turnover in central Panamanian forests has been principally attributed to the effects of dispersal limitation and a strong Caribbean to Pacific gradient in rainfall seasonality. Despite marked geological heterogeneity, the role of soil variation has not been rigorously examined. 2. We modelled the compositional turnover of trees and ferns in the Panama Canal watershed as a function of soil chemistry, climate and geographical separation, using generalized dissimilarity models (GDMs). 3. Predictability in both plant groups was strong, with 74% of turnover explained in trees and 49% in ferns. Major trends in the two plant groups were strikingly similar. The independent effects of soils, and of climate for trees, were sizeable, but those of geographical distance were minor. In both plant groups, distance and climatic effects on species turnover covaried strongly. 4. Including floristic dissimilarity of the other taxon as a predictor increased explained deviance to 81% in trees and 59% in ferns. Controlling for differences in plant density among plots reduced deviance explained by climate and distance, while soil effects remained strong. Limiting the analyses to soils of volcanic origin increased deviance explained by climate, soils and distance, but their effects covaried strongly. Independent soil effects on tree turnover were reduced, but their effects on fern turnover remained pronounced. 5. Dry season length was the most important climatic predictor for both taxa, and P and pH were the most important soil predictors. Particularly, rapid species turnover was associated with the driest end of the seasonality gradient, linked to declining individual densities and species richness, and with the low end of the phosphorus gradient. 6. Synthesis. While changes in rainfall and seasonality undoubtedly limit plant distributions in this region, soil effects are at least as important, and interactions between the two are sizeable. This is likely to hold elsewhere in the Caribbean region, where mosaics of marine and volcanic soils combined with pronounced rainfall gradients are common. Strong congruence between our focal taxa suggests that our results can be extrapolated to other plant groups, particularly as trees and ferns are distantly related and represent different life-forms.

Aim To assess the degree of fine-scale environmental determinism in fern community composition in a diverse, mid-montane tropical forest, and to test whether species local microhabitat distributions are linked to their elevational ranges regionally.Location Tunquini biological station, eastern Bolivian Andes. MethodsWe recorded fern species composition and environmental descriptors in two montane forest transects on opposite valley slopes at c. 1650 m elevation. Redundancy analysis and variation partitioning were run to examine how much floristic variation was explained by spatial and environmental (soil, forest structural, topoclimatic) variables. We then linked species local environmental distributions, as evidenced in these analyses, to their elevational ranges in Bolivia using herbarium data. ResultsThere was a striking floristic difference between the transects, attributable to differences in exposure, temperature and humidity. Environmental and floristic variation were also strongly linked within transects, with soil variation being most important. The overlap between spatially and environmentally explained variation was high, but there was congruence in the main environmental predictors selected in both transects, suggesting considerable spatial structure arose from species responses to a patchy environment. Species local microclimatic distributions were significantly linked to their regional elevational ranges. Species associated with the drier north-easterly transect had average range mid-points and maxima significantly higher (by 211 m and 402 m, respectively) than those associated with the more humid south-westerly transect. Within transects, species associated with more acidic soils and warmer microsites had higher average range minima, midpoints or maxima (by > 200 m in all cases) than those associated with less acidic and cooler microsites.Main conclusions Strong environmental determinism of community composition was apparent in relation to both local topoclimatic and soil gradients. Moreover, our results suggest that fine-scale variation in environmental conditions enhances the mixing of lowland and highland floras in mid-montane forests, thereby contributing to mid-elevation diversity.

Aim To examine the relative importance of gradients in soil fertility and rainfall for the phylogenetic structure of Neotropical forest fern communities, and to quantify how much the results are affected by phylogenetic resolution.Location Tropical lowland forests in Brazil (central Amazonia) and Panama (the Panama canal watershed). MethodsWe inventoried local fern communities at a total of 87 sites to model how their species richness and phylogenetic relatedness varied along gradients in soil fertility and rainfall. We produced a time-calibrated species-level molecular phylogeny of ferns, and quantified the phylogenetic relatedness of species within local communities using Faith's phylogenetic diversity (PD) and mean phylogenetic diversity (MPD). Calculations were compared for phylogenies resolved to the species and genus levels.Results (1) We found significant and consistent effects of soil nutrient status on local phylogenetic community structure in both regions. In contrast, phylogenetic structure showed only a weak or no relationship with rainfall. (2) In both regions, MPD declined with increasing soil fertility, which means that fern communities on poor soils consisted, on average, of less closely related lineages than fern communities on rich soils. (3) Different fern genera were over-represented in different sections of the soil nutrient gradient. (4) Qualitatively similar results were obtained whether phylogenies were resolved to the species or the genus level. Main conclusionsOur results highlight the importance of edaphic variation in structuring plant communities over evolutionary time-scales. Within the tropical forests studied, the effects of soil variation on local phylogenetic community structure seem to outweigh those of climate. Several fern genera show strong edaphic niche conservatism to either poor or rich soils, whereas many other genera have radiated to span a rather broad edaphic range. PD-based measures are so dominated by deep relationships that genus-level phylogenies are sufficient to investigate these patterns in tropical fern communities.

Summary 1.Tropical rain forest tree species are commonly perceived to have more generalized habitat distributions than understorey species. However, the correspondence between floristic patterns in large trees and smaller understorey plants in relation to environmental differences has rarely been investigated. Comparative analyses are complicated by the fact that tree data are often much more noisy, with higher overall species richness and poorer spatial and temporal representation of species composition within sampling plots. 2. Using data on trees and pteridophytes collected in 1125 plots located within 6 km 2 of lowland rain forest in Costa Rica, we asked: (i) Is there evidence that trees are more generalized in their edaphic and topographic distributions than pteridophytes, and (ii) How much might differences in sampling efficiency affect the results? 3. We ran community-level analyses using Mantel tests, and for the five most frequent tree and pteridophyte species we ran logistic regression analyses of species occurrence. To investigate the influence of differing sampling efficiency, the analyses were repeated with four rarefied pteridophyte data sets approximating key structural features of the tree data set. 4. Differences in community composition in trees and in all pteridophyte data sets were most strongly correlated with environmental differences in the same five soil variables. The most frequent species in both plant groups had non-random edaphic and topographic distributions. 5. Both Mantel correlations and the predictive performance of single species models (AUC values) were higher for the full pteridophyte data set than for trees, but these differences were much reduced after pteridophyte rarefaction. 6. The performance of the species occurrence models calibrated with rarefied pteridophyte data appeared poor when tested on rarefied data. However, their performance in predicting occurrences in the full data set was usually moderate (AUC > 0.70). This suggests that the tree species models may also predict species' potential habitat distributions better than is apparent from their AUC values. 7. Synthesis . Our results indicate that rain forest trees and pteridophytes may respond to mesoscale environmental variation in qualitatively similar ways, but that trees will appear much more generalized in their response to the environment unless differences in sampling efficiency are taken into account.

Questions:To what extent are the distributions of tropical rain forest tree ferns (Cyatheaceae) related to environmental variation, and is habitat specialization likely to play a role in their local coexistence? Location: Lowland rain forest at La Selva Biological Station, Costa Rica. Methods: Generalized linear (GLM) and generalized additive (GAM) logistic regression were used to model the incidence of four tree fern species in relation to environmental and neighbourhood variables in 1154 inventory plots regularly distributed across 6 km 2 of old-growth forest. Small and large size classes of the two most abundant species were modelled separately to see whether habitat associations change with ontogeny. Results: GLM and GAM model results were similar. All species had significant distributional biases with respect to microhabitat. Environmental variables describing soil variation were included in the models most often, followed by topographic and forest structural variables. The distributions of small individuals were more strongly related to environmental variation than those of larger individuals. Significant neighbourhood effects (spatial autocorrelation in intraspecific distributions and nonrandom overlaps in the distributions of certain species pairs) were also identified. Overlaps between congeners did not differ from random, but there was a highly significant overlap in the distributions of the two most common species. Conclusions: Our results support the view that habitat specialization is an important determinant of where on the rain forest landscape tree ferns grow, especially for juvenile plants. However, other factors, such as dispersal limitation, may also contribute to their local coexistence.