We investigated the nitrogen source for main taxa of soil fauna in two beech forests of contrasted humus type using 15 N-labelled beech litter and 15 N analysis of soil fauna. 15 Nlabelled beech litter was deposited on the topsoil in December 2000 in four stands of different ages at Leinefelde (Germany) with mull humus and in one mature stand at Sorø (Denmark) with moder humus. The fate of the tracer isotope was measured in litter and soil, as well as in the soil fauna, and for each taxa, we calculated the proportion of N in the animal derived from the labelled substrate. Of the original N contained in the litter, 20% to 41% was lost after 9 months at Leinefelde, and only 10% at Sorø. This loss was counterbalanced by the incorporation of 24% to 31% external N at Leinefelde, and 31% at Sorø, partly originating from fungal colonisation of the added litter. The proportion of N assimilated from the labelled litter by the different soil animals varied in relation to their mobility and feeding preferences.Large and mobile soil animals, especially predators, derived on average less 15 N because they were also able to feed outside the labelled litter boxes. Detritivores assimilated at most 15% of their nitrogen content at Leinefelde and 11% at Sorø from the decomposing labelled litter.The most labelled taxa at Leinefelde were small fungivorous and coprophagous species, mainly isotomid Collembola such as Isotomiella and Folsomia. At Sorø, best labelled taxa were saprophagous species such as Enchytraeidae, Glomeridae and Phthiracaroidea. These low rates of 15 N assimilation indicate that fresh litter is not directly the main N source for soil animals. The results obtained suggest that soil fauna fed preferentially upon microorganisms colonising the litter at Leinefelde (mull) and from litter itself at Sorø (moder).