The short-term effects of pulses of nutrients and jellyfish Catostylus mosaicus on planktonic assemblages were investigated in field-based experiments using 3 m 3 mesocosms. Experiments ran for 5 d and were repeated in autumn and spring at Lake Illawarra, a coastal lagoon in New South Wales, Australia. Experiments consisted of 2 orthogonal treatments, addition/non-addition of nutrients and presence/absence of jellyfish, and were designed to determine how bottom-up (i.e. addition of nutrients) and top-down (i.e. predation by jellyfish) processes influence planktonic assemblages, both independently and in combination. During both experiments, nutrients stimulated primary production and caused changes in phytoplankton assemblages. Nutrients also stimulated production of large tintinnids, suggesting that bottom-up processes may influence 2 trophic levels. Mesozooplankton were consistently depleted in mesocosms containing jellyfish. Jellyfish also caused changes in microzooplankton assemblages, indicating that top-down processes also cascade to at least 2 trophic levels. In mesocosms to which both nutrients and jellyfish were added during spring, concentrations of the red-tide forming, heterotrophic dinoflagellate Noctiluca scintillans were 20 times greater than in mesocosms to which nutrients were added alone. We hypothesize that addition of nutrients stimulated production of centric diatoms, the main prey of N. scintillans, but that a bloom of N. scintillans only formed when jellyfish were also present because jellyfish grazed on populations of herbivorous mesozooplankton (particularly the calanoid copepod Gladioferens), which generally out-competed N. scintillans for diatom prey. These data provide the first empirical evidence linking jellyfish to the formation of red tides.