Earthworm populations exhibit an irregular and aggregated distribution that may be related to vegetation, soil characteristics and biotic interactions. Their contribution to processes such as decomposition, soil aggregation and plant production is presumably limited to patches where earthworms are active. The spatial distribution of earthworm populations and how that distribution changes through time in temperate regions is not well known. This study mapped the spatial distribution of earthworm populations in 25 subplots of 50 m x 50 m sampling grids established in: (1) a cultivated corn field; (2) a mowed hayfield and (3) a deciduous forest in Quebec, Canada. Earthworms were collected three to four times a year using a combination of handsorting and formaldehyde extraction. Ten lumbricid species were found, the most common being Aporrectodea rosea Savigny. In 2001, we collected, on average, 46-177 earthworms m(-2) in the corn field, from 138 to 224 earthworms m(-2)- in the hayfield and between 124 and 253 earthworms m-2 in the forest. Due to the spatial variation in earthworm distribution, at least 40 samples ha -1 are required to make an unbiased estimate of earthworm numbers and biomass. Five subplots (patches) from each site with the highest and lowest earthworm populations were re-sampled in 2002. Patches with high earthworm populations were generally distinguishable from those with low earthworm populations in the forest, but not in the agricultural sites; this may be due to greater heterogeneity in resource distribution in the forest than agricultural sites. Further work on earthworm spatial and temporal distribution is required to determine the scale at which earthworm induced changes in soil processes and plant production may be detected. (C) 2004 Elsevier B.V. All rights reserved.