A set of eight large (20 m 3 ) mesocosms were moored in Johnson's Dock (62839.5769S, 60822.4089W, Livingston Island, Antarctica) to experimentally generate a gradient of phytoplankton biomass and production in order to test the extent of coupling between bacteria (heterotrophic Bacteria and Archaea) and phytoplankton, as well as the role of bacterial losses to protist grazers. This was achieved by imposing four light levels (100%, 50%, 25%, and 10%) in the presence or absence of nutrient additions (0.1 mol NH4Cl, 0.1 mol F6Na2Si, and 0.01 mol KH2PO4 per day per mesocosm). The experimental treatments resulted in a broad range of chlorophyll a (Chl a) (0.31-93.5 mg Chl a L 21 ) and average primary production rates, while bacteria responded in a much narrower range of biomass (3-447 m gCL 21 ) and production (0.21- 15.71 m gCL 21 d 21 ). Results confirm that bacteria-chlorophyll and bacterial production-primary production relationships in the Southern Ocean differ from the typical relationships applicable to aquatic ecosystems elsewhere. Bacteria respond to phytoplankton blooms, but they respond so weakly that bacterial production represents a small percentage of primary production (1-10%). Although other mechanisms might also contribute to the weak bacterial response to phytoplankton blooms, we demonstrate that the reason for it is likely the tight control of bacterial populations by their predators. Protist grazers are able to sustain faster growth rates in the cold waters of the Southern Ocean than are bacteria, thereby preventing bacteria from responding to phytoplankton blooms more forcibly.