Earthworm populations have been little studied in forest systems in South America, and the choice of appropriate sampling methods is an important factor to consider. Hence, the present study evaluated earthworm populations using two methods: handsorting of large soil monoliths (40 x 40 cm x 20 cm depth), and application of formalin (0.5%) to the soil surface. Worms were sampled in fifteen areas at the Embrapa Forestry Research Station, including native Atlantic forest fragments, and native (Araucaria angustifolia) or exotic (Pinus elliottii) tree species plantations. Eight species of earthworms were collected overall, including four native species (Urobenus brasiliensis, Andiorrhinus duseni, Fimoscolex nivae and Glossoscolex embrapaensis) and one exotic (M. schmardae) in low abundance, and three other exotics in higher abundance (Pontoscolex corethrurus, Amynthas gracilis and Amynthas corticis). Total earthworm density ranged from 11.5 to 102 individuals m(-2) and biomass from 5.6 to 31.6 g m(-2), with higher values being found in P. elliottii plantations. Handsorting was more effective than formalin for sampling earthworms, resulting in higher species richness, overall abundance and biomass (particularly of the endogeic P. corethrurus) in both Pine and Araucaria plantations. Formalin extraction was better for extracting epi-endogeic species (particularly M. schmardae). Eveness and A. corticis densities were higher in native forests, while A. gracilis abundance was higher in Araucaria plantations and P. corethrurus predominated in both plantations, likely due to soil-related differences between the vegetation types. Earthworm population assessments in Brazilian forest systems should prioritize handsorting, while biodiversity inventories may benefit from combining handsorting with formalin extraction.
