M. Leponce

Quantifying the spatio-temporal distribution of arthropods in tropical rainforests represents a first step towards scrutinizing the global distribution of biodiversity on Earth. To date most studies have focused on narrow taxonomic groups or lack a design that allows partitioning of the components of diversity. Here, we consider an exceptionally large dataset (113,952 individuals representing 5,858 species), obtained from the San Lorenzo forest in Panama, where the phylogenetic breadth of arthropod taxa was surveyed using 14 protocols targeting the soil, litter, understory, lower and upper canopy habitats, replicated across seasons in 2003 and 2004. This dataset is used to explore the relative influence of horizontal, vertical and seasonal drivers of arthropod distribution in this forest. We considered arthropod abundance, observed and estimated species richness, additive decomposition of species richness, multiplicative partitioning of species diversity, variation in species composition, species turnover and guild structure as components of diversity. At the scale of our study (2km of distance, 40m in height and 400 days), the effects related to the vertical and seasonal dimensions were most important. Most adult arthropods were collected from the soil/litter or the upper canopy and species richness was highest in the canopy. We compared the distribution of arthropods and trees within our study system. Effects related to the seasonal dimension were stronger for arthropods than for trees. We conclude that: (1) models of beta diversity developed for tropical trees are unlikely to be applicable to tropical arthropods; (2) it is imperative that estimates of global biodiversity derived from mass collecting of arthropods in tropical rainforests embrace the strong vertical and seasonal partitioning observed here; and (3) given the high species turnover observed between seasons, global climate change may have severe consequences for rainforest arthropods.

Physico-chemical properties of soils have an indirect effect on ground-dwelling ant assemblages at regional scale. By contrast, little is known on their effect at the scale of the meter, at which direct interactions between subterranean ants and their environments occur.
We aimed to study the effect of soil properties on the small-scale distribution of highly diverse subterranean ant assemblages. We expected ant species to be differentially associated with soil variables.
This study was conducted in Bombuscaro and Copalinga (Ecuador), two premontane tropical forests 2 km apart, sharing the same climate and altitude. In each site, we delineated a 100 m line transect and every meter collected a sample of mineral soil (15 cm × 15 cm × 10 cm) and the associated ant fauna.
Compared with traditional baiting this method should provide a relatively unbiased measure of subterranean ant diversity. Organic matter content, pH, electrical conductivity, degree of humification, soil texture and land slope were measured at each sampling point.
Soil attributes of the two sites were contrasting and strongly spatially autocorrelated. The two ant assemblages shared only 12% of their 76 species. Ant diversity was higher in Bombuscaro (Fisher-alpha = 21.7 vs. 16 in Copalinga). Species turnover was remarkably high at both regional (intersite) and local (intrasite) scales. Ant species richness was not correlated with any of the measured variables. Soil properties were not associated with the occurrence of any ant species in Copalinga. In Bombuscaro, Hypoponera sp.01RI was correlated with lower pH values, and Acropyga fuhrmanni to higher clay content. A. fuhrmanni did not show correlation with this variable in Copalinga, where it is the dominant species (present in 48% of the samples). However, this species is associated with higher clay content at Copalinga (36.7 ± 6% SD) compared with Bombuscaro (6.5 ± 2%).
Although weakly explained by the soil factors investigated, a high ant species turnover was observed at both small (1 m) and regional (2 km) spatial scales. The latter observation stresses the importance of designing large nature reserves including mosaics of habitats to conserve subterranean ant diversity in the Andes.

ABSTRACT The fire ant Solenopsis saevissima is a major pest frequent in human-disturbed areas of its native range where it forms 'supercolonies'. We determined that its natural habitat in French Guiana is likely the sporadically flooded riparian forest and aimed to evaluate this ant's impact on the abundance and diversity of other ants by comparing different habitats at two sites. We noted a significant decrease in ant species richness between the rainforest and human-disturbed habitats (but not between the former and the naturally disturbed riparian forest), and between extreme habitats and all others. The number of ant nests per surface unit (i.e., quadrats of equal surface area), a proxy of ant abundance, globally followed this pattern. Solenopsis saevissima was absent from pristine rainforest (as expected) and from extreme habitats, showing the limits of its adaptability, whereas some other native ants can develop in these habitats. Ant species richness was significantly lower in the presence of S. saevissima in the riparian forest, forest edges and meadows, illustrating that this ant species has a negative impact on the ant communities in addition to the impact of natural- and man-made disturbances. Only some ant species can develop in its presence, and certain of these can even thrive. Because it has been recorded in Africa, Guadeloupe and the Galápagos Islands, we concluded that, due to the increasing volume of global trade and forest destruction, S. saevissima could become a pantropical invasive species.

DI-fusion, le Dépôt institutionnel numérique de l'ULB, est l'outil de référencementde la production scientifique de l'ULB.L'interface de recherche DI-fusion permet de consulter les publications des chercheurs de l'ULB et les thèses qui y ont été défendues.