Dispersion of the tracer sulphur hexafluoride (SF6) during the Southern Ocean Iron Enrichment Experiment (SOIREE) provided an estimate of vertical exchange at the base of the surface mixed layer (60 m) at 61¿S 140¿E. Budget analysis confirmed that the SF6 patch was well constrained by surface mapping, with the decline in total SF6 showing good agreement with that predicted from wind speed parameterizations. Two approaches were used to calculate the mean effective vertical diffusivity Kz from the diapycnal transfer of SF6, with complementary error function and second-moment fits to the SF6 depth profiles indicating that Kz was less than 0.3 ¿ 10-4 m2s-1. This result was examined using a three-dimensional diffusion model that incorporated lateral dispersion and air-sea exchange losses, which confirmed that vertical shear and subpycnocline dispersion did not influence the Kz estimate. Current shear at the base of the mixed layer was generated by wind-driven inertial oscillation, with a decrease in wind speed and increasing stratification in the latter half of the experiment reducing diapycnal transfer of SF6. A compilation was used to examine the potential of both N (Brunt-V¿is¿la frequency) and Ri (gradient Richardson number) on the basis of parameterizations of Kz in the seasonal pycnocline. Application of Kz to nutrient gradients in the seasonal pycnocline suggests that vertical diffusion represents a significant pathway for silicic acid supply in late summer. Furthermore, use of the mean effective Kz (0.11 ¿ 0.2 ¿ 10-4 m2s-1) indicates that vertical diffusion supplies a large proportion of the iron required for new production in this region.