-Difficulties to simply transfer trophic cascade theory from temperate to tropical lakes are now well recognized. Many mechanisms trying to explain top-down divergences between these systems have been proposed, such as lack of key species of herbivorous zooplankton, absence of seasonality in fish reproduction, cyanobacteria development, or differences in fish foraging behaviour. Very few studies have considered bottom-up mechanisms, in particular differences in nutrient recycling and nutrient limitation between the two types of ecosystems. According to the ecological stoichiometry theory, fish-induced alterations of food-web structure could modify the efficiency of consumer-driven nutrient recycling by changing the relative biomass contribution of species in food webs. Consequently, by mostly considering top-down processes, one could underestimate consequences on nutrient availability for phytoplankton growth. In this paper, we compared the results of two mesocosm experiments carried out in temperate and tropical areas, each manipulating foodweb structure via the presence or absence of fish. We found trophic cascades in both experiments, but differences between fishless and fish treatments were greater in temperate than in tropical systems. In the tropical experiment, the observed effects could not be supported by classical zooplankton community alteration or by cyanobacteria prevalence. Our results suggest a key contribution of fish nitrogen-excretion to phytoplankton growth in mostly nitrogen-limited tropical systems. Differences in stoichiometric response to food-web structure alteration between temperate and tropical lakes could thus represent a major difference between the two systems. Our study stresses the need for further studies that would allow robust generalization on the functioning of freshwater temperate and tropical ecosystems.