Faculté des Sciences

Aims Relationships between local habitat heterogeneity and tree communities in miombo woodlands have been very little studied. While some studies have addressed this topic at broad scales and based on few environmental parameters, this study aims at (i) detecting fine-scale habitats (≤10 ha) on the basis of a detailed characterisation of soil explicitly considering past anthropogenic disturbances, and an exhaustive census of the tree community, and at (ii) searching for indicator tree species corresponding to the resulting habitats. Methods The study was carried out in the miombo woodland of Mikembo Forest Reserve, Upper Katanga, The Democratic Republic of the congo. A complete census of the tree community was conducted in a 10-ha forest dynamics plot comprising 160 adjacent quadrats of 25 × 25 m, with a total of 4604 trees (diameter at breast height > 10 cm). Thirty-six physicochemical soil parameters were measured. Studying the frequency distribution of soil charcoal content allowed identifying local signature of past human agriculture in the soil. Two strategies were used to define habitats: (i) a combination of principal component analysis (PcA) on soil variables and Ward clustering and (ii) multivariate regression trees (MRT) to search for key soil parameters allowing the best prediction of species composition. Tree-habitat associations were tested by means of a robust statistical framework combining the IndVal index and torus randomisations. Important Findings The forest contained 82 tree species and a significant proportion of wet miombo species (e.g. Marquesia macroura). We detected a strong east–west edaphic gradient driven by soil texture; most chemical soil parameters followed this pattern. Five habitats were identified based on soil factors and floristic composition. Nine indicator species of these habitats were found. The key soil factors discriminating habitats were total calcium, available forms of phosphorus and clay content. Even though past agricultural practices were successfully detected in soils, they did not display any significant influence neither on habitat differentiation nor on the associated tree communities. Based on an unprecedented large number of soil parameters, fine-scale soil heterogeneity and niche partitioning were shown to contribute to the variability of the floristic composition in this forest. Our results indicated that considering the most variable environmental parameters, as in PcA, is a poor manner for defining habitats. In contrast, combining MRT with the IndVal index and torus randomisation has proved to be a much more robust and sensitive approach to highlight tree-habitat associations at this scale. The common dichotomous viewpoint of considering deterministic and neutral effects as acting at broad and fine scales, respectively, is not confirmed when measuring suitable environmental variables, even in a case where the physical environment does not exhibit strong heterogeneity.

One sentence summary: Functional traits of host tree species and chemical soil parameters shape the spatial distribution of the ectomycorrhizal fungal community of a miombo forest. ABSTRACT Ectomycorrhizal fungi (EMF) are highly diversified and dominant in a number of forest ecosystems. Nevertheless, their scales of spatial distribution and the underlying ecological processes remain poorly understood. Although most EMF are considered to be generalists regarding host identity, a preference toward functional strategies of host trees has never been tested. Here, the EMF community was characterised by DNA sequencing in a 10-ha tropical dry season forest—referred to as miombo—an understudied ecosystem from a mycorrhizal perspective. We used 36 soil parameters and 21 host functional traits (FTs) as candidate explanatory variables in spatial constrained ordinations for explaining the EMF community assemblage. Results highlighted that the community variability was explained by host FTs related to the 'leaf economics spectrum' (adjusted R 2 = 11%; SLA, leaf area, foliar Mg content), and by soil parameters (adjusted R 2 = 17%), notably total forms of micronutrients or correlated available elements (Al, N, K, P). Both FTs and soil generated patterns in the community at scales ranging from 75 to 375 m. Our results indicate that soil is more important than previously thought for EMF in miombo woodlands, and show that FTs of host species can be better predictors of symbiont distribution than taxonomical identity.

1. Species distribution within plant communities results from both the influence of deterministic processes , related to environmental conditions, and neutral processes related to dispersal limitation and stochastic events, the relative importance of each factor depending on the observation scale. 2. Assessing the relative contribution of environment necessitates controlling for spatial dependences among data points. Recent methods, combining multiple regression and Moran's eigenvectors maps (MEM), have been proved successful in disentangling the influence of pure spatial processes related to dispersal limitation, pure environmental variables (not spatially structured) and spatially structured environmental properties. However, the latter influence is usually not testable when using advanced spatial models like MEM. 3. To overcome this issue, we propose an original approach, based on torus-translations and Moran spectral randomizations, to test the fraction of species abundance variation that is jointly explained by space and seven soil variables, using three environmental and tree species abundance data sets (consisting of 120, 52 and 34 plots of 0Á2 ha each, located along 101-, 66-and 35-km-long transect-like inventories, respectively) collected in tropical moist forests in southern Cameroon. 4. The overall abundance of species represented by ≥30 individuals, and 27% of these species taken individually, were significantly explained by fine-scale (<5 km) and/or broad-scale (5–100 km) spatially structured variations in soil nutrient concentrations (essentially the concentration of available Mn, Mg and Ca) along the 120-plots area. The number of significant tests considerably decreased when investigating the two smaller data sets, which mostly resulted from low statistical power rather than weaker floristic and/or edaphic variation captured among plots. 5. Synthesis. Our results provide evidence that tree species turnovers are partly controlled by spatially structured concentrations in soil nutrients at scales ranging from few hundreds of metres to c. 100 km, a poorly documented subject in Central African forests. We also highlight the usefulness of our testing procedure to correctly interpret the space-soil fraction of variation partitioning analyses (which always accounted here for the most important part of the soil contribution), as this fraction was sometimes relatively high (R 2 values up to c. 0Á3) but nearly or not significant.

Eigenvector mapping techniques are widely used by ecologists and evolutionary biologists to describe and control for spatial and/or phylogenetic patterns in their data. The selection of an appropriate subset of eigenvectors is a critical step (misspecifica-tion can lead to highly biased results and interpretations), and there is no consensus yet on how to proceed. We conducted a ten-year review of the practices of eigenvector selection and highlighted three main procedures: selecting the subset of descriptors minimising the Akaike information criterion (AIC), using a forward selection with double stopping criterion after testing the global model significance (FWD), and selecting the subset minimising the autocorrelation in the model residuals (MIR). We compared the type I error rates, statistical power, and R² estimation accuracy of these methods using simulated data. Finally, a real dataset was analysed using variation partitioning analysis to illustrate to what extent the different selection approaches affected the ecological interpretation of the results. We show that, while the FWD and MIR approaches presented a correct type I error rate and were accurate, the AIC approach displayed extreme type I error rates (100%), and strongly overestimated the R². Moreover, the AIC approach resulted in wrong ecological interpretations, as it overestimated the pure spatial fraction (and the joint spatial-environmental fraction to a lesser extent) of the variation partitioning. Both the FWD and MIR methods performed well at broad and medium scales but had a very low power to detect fine-scale patterns. The FWD approach selected more eigenvectors than the MIR approach but also returned more accurate R² estimates. Hence, we discourage any future use of the AIC approach, and advocate choosing between the MIR and FWD approaches depending on the objective of the study: controlling for spatial or phylogenetic autocorrelation (MIR) or describing the patterns as accurately as possible (FWD).

1. Species distribution within plant communities results from both the influence of deterministic processes , related to environmental conditions, and neutral processes related to dispersal limitation and stochastic events, the relative importance of each factor depending on the observation scale. 2. Assessing the relative contribution of environment necessitates controlling for spatial dependences among data points. Recent methods, combining multiple regression and Moran&amp;#39;s eigenvectors maps (MEM), have been proved successful in disentangling the influence of pure spatial processes related to dispersal limitation, pure environmental variables (not spatially structured) and spatially structured environmental properties. However, the latter influence is usually not testable when using advanced spatial models like MEM. 3. To overcome this issue, we propose an original approach, based on torus-translations and Moran spectral randomizations, to test the fraction of species abundance variation that is jointly explained by space and seven soil variables, using three environmental and tree species abundance data sets (consisting of 120, 52 and 34 plots of 0Á2 ha each, located along 101-, 66-and 35-km-long transect-like inventories, respectively) collected in tropical moist forests in southern Cameroon. 4. The overall abundance of species represented by ≥30 individuals, and 27% of these species taken individually, were significantly explained by fine-scale (&amp;lt;5 km) and/or broad-scale (5–100 km) spatially structured variations in soil nutrient concentrations (essentially the concentration of available Mn, Mg and Ca) along the 120-plots area. The number of significant tests considerably decreased when investigating the two smaller data sets, which mostly resulted from low statistical power rather than weaker floristic and/or edaphic variation captured among plots. 5. Synthesis. Our results provide evidence that tree species turnovers are partly controlled by spatially structured concentrations in soil nutrients at scales ranging from few hundreds of metres to c. 100 km, a poorly documented subject in Central African forests. We also highlight the usefulness of our testing procedure to correctly interpret the space-soil fraction of variation partitioning analyses (which always accounted here for the most important part of the soil contribution), as this fraction was sometimes relatively high (R 2 values up to c. 0.3) but nearly or not significant.

Bacteria can damage sperm and thus reduce the reproductive success of both males and females; selection should therefore favour the evolution of antimicrobial protection. Eusocial hymenopterans might be particularly affected by such bacterial infections because of their mating ecology. In both sexes, mating is restricted to a short window early in the adult stage; there are no further chances to mate later in life. Males die shortly after mating, but queens use the acquired sperm to fertilise their eggs for years, sometimes decades. The reproductive success of both sexes is, thus, ultimately sperm-limited, which maintains strong selection for high sperm viability before and after storage. We tested the antibacterial activity of the contents of the male and female sperm-storage organs - the accessory testes and the spermatheca, respectively. As our study species, we used the bacterium Escherichia coli and the garden ant Lasius niger, whose queens can live for several decades. Our results provide the first empirical evidence that male and female sperm-storage organs display different antibacterial activity. While the contents of the accessory testes actually enhanced bacterial growth, the contents of the spermatheca strongly inhibited it. Furthermore, mating appears to activate the general immune system in queens. However, antimicrobial activity in both the spermatheca and the control tissue (head-thorax homogenate) declined rapidly post-mating, consistent with a trade-off between immunity and reproduction. Overall, this study suggests that ejaculates undergo an immune 'flush' at the time of mating, allowing storage of sperm cells free of bacteria.

Community assembly rules have been extensively studied, but its association with regional environmental variation, while land use history remains largely unexplored. Land use history might be especially important in Mediterranean forests, considering their historical deforestation and recent afforestation. Using forest inventories and historical (1956) and recent (2000) land cover maps, we explored the following hypotheses: 1) woody species assembly is driven by environmental factors, but also by historical landscape attributes; 2) recent forests exhibit lower woody species richness than pre-existing due to the existence of colonization credits; 3) these credits are modulated by species’ life-forms and dispersal mechanisms. We examined the association of forest historical type (pre-existing versus recent) with total species richness and that of diverse life-forms and dispersal groups, also considering the effects of current environment and past landscape factors. When accounting for these effects, no significant differences in woody species richness were found between forest historical types except for vertebrate-dispersed species. Species richness of this group was affected by the interaction of forest historical type with distance to coast and rainfall: vertebrate-dispersed species richness increased with rainfall and distance to the coast in recent forests, while it was higher in dryer sites in pre-existing forests. In addition, forest historical types showed differences in woody species composition associated to diverse environmental and past landscape factors. In view of these results we can conclude that: 1) community assembly in terms of species richness is fast enough to exhaust most colonization credit in recent Mediterranean forests except for vertebrate-dispersed species; 2) for these species, colonization credit is affected by the interplay of forest history and a set of proxies of niche and landscape constraints of species dispersal and establishment; 3) woody species assemblage is mostly shaped by the species’ ecological niches in these forests.

Eigenvector-mapping methods such as Moran's eigenvector maps (MEM) are derived from a spatial weighting matrix (SWM) that describes the relations among a set of sampled sites. The specification of the SWM is a crucial step, but the SWM is generally chosen arbitrarily, regardless of the sampling design characteristics. Here, we compare the statistical performances of different types of SWMs (distance-based or graph-based) in contrasted realistic simulation scenarios. Then, we present an optimization method and evaluate its performances compared to the arbitrary choice of the most-widely used distance-based SWM. Results showed that the distance-based SWMs generally had lower power and accuracy than other specifications, and strongly underestimated spatial signals. The optimization method, using a correction procedure for multiple tests, had a correct type I error rate, and had higher power and accuracy than an arbitrary choice of the SWM. Nevertheless, the power decreased when too many SWMs were compared, resulting in a trade-off between the gain of accuracy and the loss of power. We advocate that future studies should optimize the choice of the SWM using a small set of appropriate candidates. R functions to implement the optimization are available in the adespatial package and are detailed in a tutorial.

Variation partitioning analyses combined with spatial predictors (Moran’s eigenvector maps, MEM) are commonly used in ecology to test the fractions of species abundance variation purely explained by environment and space. However, while these pure fractions can be tested using a classical residuals permutation procedure, no specific method has been developed to test the shared space-environment fraction (SSEF). Yet, the SSEF is expected to encompass a major driver of community assembly, that is, an induced spatial dependence effect (ISD; i.e. the reflection of a spatially structured habitat filter on a species distribution). A reliable test of this fraction is therefore crucial to properly test the presence of an ISD on ecological data. To bridge the gap, we propose to test the SSEF through spatially-constrained null models: torus- translations, and Moran spectral randomisations. We investigated the type I error rate and statistical power of our method based on two real environmental datasets and simulations of tree distributions. Ten types of tree distribution displaying contrasted aggregation properties were simulated, and their abundances were sampled in 153 regularly-distributed 20 × 20 m quadrats. The SSEF was tested for 1000 simulated tree distributions either unrelated to the environment, or filtered by environmental variables displaying contrasting spatial structures. The method proposed provided a correct type I error rate (< 0.05). The statistical power was high (> 0.9) when abundances were filtered by an environmental variable structured at broad scale. However, the spatial resolution allowed by the sampling design limited the power of the method when using a fine-scale filtering variable. This highlighted that an ISD can be properly detected providing that the spatial pattern of the filtering process is correctly captured by the sampling design of the study. An R function to apply the SSEF testing method is provided and detailed in a tutorial.

Question: Do termitophilous and non- termitophilous trees of dry tropical woodlands show local adaptation? Location: Region of Lubumbashi, Upper Katanga, DR Congo. Methods: Three pairs of congeneric tree species showing strict edaphic specializa- tion with respect to termite mounds, Combretum molle (termitophilous, T)/C. collinum (non- termitophilous, NT); Strychnos potatorum (T)/S. spinosa (NT), Ziziphus mucronata (T)/Z. abyssinica (NT), were used in a reciprocal transplant experiment in situ. Seedlings were reciprocally transplanted on termite mounds and in the surrounding matrix in a miombo woodland. Growth (height and number of leaves) and survival were monitored for 30 months. Soil physical and chemical properties, and available water, were assessed on and off mounds. Results: Growth was little affected by habitat; only one species showed better growth in its home habitat (Strychnos spinosa in the matrix). Survival was strongly affected by habi- tat, in opposite directions consistent with species’ habitat specialization. Termitophilous species experienced a very high mortality rate in the matrix, especially during the dry season. Available water content was higher in termite mound soil than in the matrix soil. Conclusions: Termitophilous and non- termitophilous tree species show local adapta- tion at the seedling stage, expressed mostly as different patterns of mortality in the dry season. The results point to water supply as a critical factor in the edaphic spe- cialization of termitophilous species. In contrast, the higher mortality of non- termitophilous species on termite mounds is not explained by water stress.