A manipulative mesocosm experiment was done in a saline coastal lake in Australia to compare the top-down and bottom-up influences of jellyfish on primary production and planktonic assemblages. We hypothesized that non-zooxanthellate jellyfish (Catostylus mosaicus) would exert both 'top-down' (grazing) and 'bottom-up' (nutrient excretion) effects, whereas zooxanthellate jellyfish (Phyllorhiza punctata) would only exert top-down influences, as their dissolved excretory products are internally recycled to their zooxanthellae rather than released into the water column. Experimental treatments consisted of triplicate mesocosms that contained two C. mosaicus, two P. punctata, a combination of C. mosaicus and P. punctata, no jellyfish, and an open-water control (sampling outside mesocosms). Both species of jellyfish preyed heavily on mesozooplankton, initiating a topdown trophic cascade that resulted in increased production of the heterotrophic dinoflagellate Protoperidinium sp. However, no increase in primary production or phytoplankton biomass was observed in the treatments containing P. punctata, indicating that top-down processes did not extend to primary producers. Bottom-up excretion of nutrients, however, caused phytoplankton biomass to more than double in the C. mosaicus treatment compared to all other treatments. Increased primary production was due largely to a 10-fold increase in the diatom Chaetoceros sp. and was predominantly driven by C. mosaicus excreting phosphate (PO43-), which was the limiting nutrient in the lake. Blooms of both zooxanthellate and non-zooxanthellate jellyfish will deplete mesozooplankton and alter the composition of microzooplankton assemblages via top-down processes. Excretion of nutrients by blooms of non-zooxanthellate jellyfish, however, can also greatly increase phytoplankton production and could favor algal blooms.