We examine ship-based observations of atmospheric O2 and CO2 in the southern Pacific Ocean made during two voyages: in February 2003 and in April 2004. We found, for the Austral late summer to autumn, evidence of a maximum in Atmospheric Potential Oxygen (APO) (an atmospheric tracer that is conservative with respect to terrestrial biological activity, APO ˜ O2 + CO2) centered around 50°S owing to biologically driven O2 outgassing (associated with strong productivity in the Sub-Tropical Convergence Zone) and evidence in the February voyage of a decreasing APO trend from 57°S south toward the Antarctic coast. The observed APO variability appears to be primarily determined by the interplay between atmospheric transport and the spatial distribution of O2 air-sea flux resulting from regional differences in biological production. Comparisons of these observations with APO derived from coupling the TM3 atmospheric transport model with flux estimates from (1) the PISCES-T ocean biogeochemistry model, and (2) O2 and CO2 climatologies, show that both PISCES-T and the climatology reproduce the observed gradient and curvature of APO but underestimate the magnitude of the observed APO maximum in the mid southern latitudes. Although the temporal limitation of our data does not permit us to calculate the annual mean APO gradient, the results support previous model predictions of a decreasing APO trend from the mid to high southern latitudes.