Résumé
1. Scientific understanding of acidification in aquatic ecosystems relies on effective assessment, which at present is mostly limited to chemical and sometimes structural biological variables. Effects on ecosystem functioning are, in contrast, largely neglected. Litter breakdown is a potentially useful, highly integrative and crucial process that could enhance such assessment programmes.2. Breakdown rates of beech Fagus sylvatica leaves were determined in 25 woodland headwater streams along an acidification gradient in the Vosges Mountains, France. Additional data relating to micro-organisms (microbial respiration, fungal biomass and degree of conditioning measured as leaf palatability) and macroinvertebrates (shredder diversity, abundance and biomass) associated with decomposing leaves were collected to elucidate the mechanisms underlying leaf breakdown.3. Breakdown rates varied more than 20-fold between the most acidified and circum- neutral sites (k = 0·0002–0·0055 day−1). Stream water alkalinity and total Al concen- tration together accounted for 88% of the variation in litter breakdown rates among streams. Microbial factors associated with decaying leaves, particularly microbial respiration, declined with increasing stream acidity and were significantly related to Ca2+ and total Al concentrations.4. Total abundance, biomass and richness of leaf-shredding invertebrates associated with decomposing leaves were not related to stream acidity. However, the abundance and biomass of the amphipod Gammarus fossarum, an acid-sensitive and particularly efficient leaf-shredder, showed a strong positive relationship with leaf breakdown rate. Gammarus abundance and microbial respiration together accounted for 85% of the variation in litter breakdown rates among streams.5. Synthesis and applications. These results indicate that leaf-litter breakdown responds strongly to stream acidification, with both microbial decomposers and invertebrate detritivores markedly affected. Measuring leaf breakdown rate may be developed into a simple, powerful and low-cost tool for assessing a critical component of ecosystem functioning. We advocate further investigation of this approach for the routine bio- monitoring of freshwaters affected by, or recovering from, other anthropogenic stresses.