Résumé
In this study, we investigated the spatial distribution of an earthworm community together with the heterogeneity of selected soil properties in a gallery forest (GF) of the Colombian "Llanos". We performed fine-scale spatial variability by intensively sampling 100 points distributed in the nodes of a regular grid with 5 in inter-sample distance. Non-parametric statistics were used and included SADIE analysis and partial Mantel test, in addition to geostatistics (semi-variograms) and correlogram computation. Our results indicated that the spatial distribution of earthworms was characterized by areas of presence (patches) and absence (gaps), although the general pattern was random at the scale of this study (< 5 m), while soil physico-chemical characteristics showed a clumped spatial distribution. Contrary to previous results reported for the nearby savanna, a significant spatial-association was found for two competing endogeic species Andiodrilus sp. and Glossodrilus sp. in the GF. Semi-variograms of soil environmental factors were adjusted to model families most commonly used (spherical and linear), and correlograms for earthworms showed significant positive and negative spatial autocorrelation for lag distances < 15 m and > 30 m, respectively. Partial Mantel test revealed specific significant relationships between soil variables and some species. The earthworm community of the GF displayed a random structure in a spatially clumped soil environment, and our results suggested that spatial distribution observed for some species could be the result of preferential selection of soil environmental factors. In other words, soil heterogeneity contributed to the formation of population patches for some earthworm species. The variability of suitable sites (resource availability patchiness) exerted an influence in the spatial distribution of earthworms at the scale used in this study, and we identified the spatial scale at which both environmental heterogeneity could influence and express earthworm impact on soil properties. (C) 2011 Elsevier Ltd. All rights reserved.