Tropical forests contain an important proportion of the carbon stored in terrestrial vegetation, but estimated aboveground biomass (AGB) in tropical forests varies two-fold, with little consensus on the relative importance of climate, soil and forest structure in explaining spatial patterns. Here, we present analyses from a plot network designed to examine differences among contrasting forest habitats (terra firme, seasonally flooded, and white-sand forests) that span the gradient of climate and soil conditions of the Amazon basin. We installed 0.5-ha plots in 74 sites representing the three lowland forest habitats in both Loreto, Peru and French Guiana, and we integrated data describing climate, soil physical and chemical characteristics and stand variables, including local measures of wood specific gravity (WSG). We use a hierarchical model to separate the contributions of stand variables from climate and soil variables in explaining spatial variation in AGB. AGB differed among both habitats and regions, varying from 78Mgha_1 in white-sand forest in Peru to 605Mgha_1 in terra firme clay forest of French Guiana. Stand variables including tree size and basal area, and to a lesser extent WSG, were strong predictors of spatial variation in AGB. In contrast, soil and climate variables explained little overall variation in AGB, though they did co-vary to a limited extent with stand parameters that explained AGB. Our results suggest that positive feedbacks in forest structure and turnover control AGB in Amazonian forests, with richer soils (Peruvian terra firme and all seasonally flooded habitats) supporting smaller trees with lower wood density and moderate soils (French Guianan terra firme) supporting many larger trees with high wood density. The weak direct relationships we observed between soil and climate variables and AGB suggest that the most appropriate approaches to landscape scale modeling of AGB in the Amazon would be based on remote sensing methods to map stand structure. (Résumé d'auteur

We describe patterns of tree community change along a 700‐km transect through terra firme forests of western Amazonia, running from the base of the Andes in Ecuador to the Peru–Brazil border. Our primary question is whether floristic variation at large scales arises from many gradual changes or a few abrupt ones. Data from 54 1‐ha tree plots along the transect support the latter model, showing two sharp discontinuities in community structure at the genus level. One is located near the Ecuador–Peru border, where the suite of species that dominates large areas of Ecuadorean forest declines abruptly in importance to the east. This discontinuity is underlain by a subterranean paleoarch and congruent with a change in soil texture. A second discontinuity is associated with the shift from clay to white sand soils near Iquitos. We hypothesize that the first discontinuity is part of an edaphic boundary that runs along the Andean piedmont and causes a transition from tree communities preferring richer, younger soils near the base of the Andes to those preferring poorer, older soils farther east. Because the floristic changes observed at this discontinuity are conserved for large distances to the east and west of it, the discontinuity is potentially key for understanding floristic variation in western Amazonia. The significant floristic turnover at the Ecuador–Peru border suggests that the only large protected area in the region—Ecuador's Yasuní National Park—is not adequate protection for the very diverse tree communities that cover vast areas of northern Peru.

MO 63121-4499, USA. Present address for CB: Univ. Nacional de Corboda, Argentina.Within the tropics, a marked gradient in rainfall between dry and wet forests correlates with a well documented turnover of plant species. While water availability along these gradients is an important determinant of species distributions, other abiotic and biotic factors correlate with rainfall and may also contribute to limit species distribution. One of these is soil fertility, which is often lower in the wetter forests. To test its possible role in species distribution along a rainfall gradient, we performed a screen-house experiment where we measured the growth performance of seedlings of 23 species with contrasting distributions across the Isthmus of Panama. We grew seedlings in soils collected from the drier Pacific side and the wetter Atlantic side. Differences in soil fertility across the Isthmus were large enough to significantly influence the growth performance of the seedlings. However, we found no evidence of home-soil advantage among species with contrasting distributions. Dry-distribution species grew on average slower than wet-distribution species suggesting a cost to drought adaptations. The response to soil differences correlated with the growth rate of the species, such that fastgrowing species responded more to changes in soil quality. We hypothesize that inherently slow growth rates of some drydistribution tropical species may be a more important factor limiting their colonization of wetter sites along the rainfall gradient.

The form and functioning of peatlands depend strongly on their hydrological status, but there are few data available on the hydraulic properties of tropical peatlands. In particular, the saturated hydraulic conductivity (K) has not previously been measured in neotropical peatlands. Piezometer slug tests were used to measure K at two depths (50 and 90 cm) in three contrasting forested peatlands in the Peruvian Amazon: Quistococha, San Jorge and Buena Vista. Measured K at 50 cm depth varies between 0.00032 and 0.11 cm s À1 , and at 90 cm, it varies between 0.00027 and 0.057 cm s À1 . Measurements of K taken from different areas of Quistococha showed that spatial heterogeneity accounts for~20% of the within-site variance and that depth is a good predictor of K. However, K did not vary significantly with depth at Buena Vista and San Jorge. Statistical analysis showed that~18% of the variance in the K data can be explained by between-site differences. Simulations using a simple hydrological model suggest that the relatively high K values could lead to lowering of the water table by >10 cm within~48 m of the peatland edge for domed peatlands, if subjected to a drought lasting 30 days. However, under current climatic conditions, even with high K, peatlands would be unable to shed the large amount of water entering the system via rainfall through subsurface flow alone. We conclude that most of the water leaves these peatlands via overland flow and/or evapotranspiration.