Eight hundred photosynthesis-irradiance relationships were determined in the shelf waters adjacent to Palmer Station, Antarctica during the spring/summer periods of 1991–94. Biomass specific maximum photosynthetic rate, PBmax, and the light limited photosynthetic efficiency, αB, were poorly correlated to the physical forcing and nutrient regimes at the sampling stations. The two photosynthetic parameters, however, did strongly covary indicating the minimum irradiance required to saturate photosynthesis, Ik, was relatively constant in this highly variable environment. The variability in Ik could partially be attributed to both depth in the water column and time of the year, with the highest values occurring for surface samples during the summer period of peak incident irradiance. Given this and the significant dependence of Pmax on phytoplankton biomass, a simple empirical model for primary productivity was developed. An independent test of the model was performed on data collected in a mesoscale offshore grid and predicted primary production was found to be within 13% of measured values. Although there are limitations to this approach (i.e. exclusion of diel periodicity in photosynthetic response), these results provide relatively robust estimates of daily primary production for the Southern Ocean.