Decomposition and associated nutrient cycling is a key process for ecosystem functioning, in particular in tropical rainforests with highly weathered and nutrient-impoverished soils. Previous studies in an Amazonian rainforest of French Guiana showed high variation in leaf litter quality among co-occurring tree species. For example, litter phosphorus (P) concentrations varied by a factor of seven among tree species, resulting in widely different litter C:N:P stoichiometries. Stoichiometric constraints of microbial decomposers might be an important control factor of decomposition. We tested this hypothesis with a microcosm experiment under controlled conditions using leaf litter from six tropical tree species of French Guiana that are distinctly separated along a C:N:P gradient. To test whether non-limiting external resources can compensate for non-optimal wide litter elemental ratios, we additionally supplemented a series of microcosms with external resources of C (as cellulose), N (as NH4NO3), and P (as KH2PO4) in all possible combinations. Litter was incubated in a total of 240 microcosms (6 litter types x 8 resource treatments x 5 replicates) for 236 days to determine litter mass loss, substrate induced microbial respiration (SIR) and activities of nine extracellular enzymes.

Results/Conclusions

Litter mass loss was negatively correlated with litter inherent C:N, but not with C:P and N:P. Moreover, interspecific variation in decomposition was related to differences in C quality, with leaf litter rich in condensed tannins decomposing more slowly. The addition of mineral N and P generally increased decomposition and SIR, but this effect depended on the species identity of the litter. The effects of a combined addition of N and P were superior to the individual effects of N or P. Both, the P and the NP treatment effects increased significantly with increasing litter N:P ratio. We found no significant effect of external C addition. Apart of a significant reduction in phosphatase activity with the addition of P, there were no clear treatment effects on enzyme activities. Our results do partly support the hypothesis of litter inherent stoichiometry control on decomposition. Collectively, our results suggest that tree species-specific litter C quality traits exhibit a strong control on microbial decomposers in this microcosm experiment, with a noticable effect of non-limiting external N and P availability.