Eddy correlation instruments mounted above a plantation of Pinus radiato near Mt. Gambier, South Australia, have been operated during two periods of intensive effort, in May and October, 1972. Measurements of the Reynolds stress and of wind speed gradients show that the zero plane for momentum is located at about d = 0.8h (where h is the height of the trees), and that the roughness length of the surface is about 30% of the difference (h—d).Sensible heat fluxes and temperature gradients give a displacement length not significantly different from that applicable in the momentum case, but the roughness length for sensible heat transfer is smaller than that for momentum, by about a factor of 3.Advective effects are found to be important, particularly when the fetch across the canopy is less than about 0.8 km (corresponding in our case to an effective fetch/height ratio of between 100 and 200). Long-fetch cases allow an evaluation of the heat storage (S) in the canopy and in the air below the height of eddy flux measurement. The rate of heat storage is found to be about 60 ± 20 W m−2 per °C h−1 of canopy temperature change (for a densely packed forest with trees about 13 m high), which is compatible with measurements of the biomass and assumed specific heats. The residual heat energy at about 6 m above the effective zero plane, unaccounted for by the various measured fluxes, is found to be related to the difference in net radiation over grassland and forest.During daytime, the forest is found to lose heat by turbulence in much the same manner as pasture, with fluxes of similar magnitude (although possibly differing to the extent of differences in ground flux, albedo and emissivity, for example) and giving similar Bowen ratios. At night, however, the evaporation from the forest tends to continue as heat is supplied by the cooling canopy. This is in direct contrast to the usual situation over pasture, where the heat storage is not of sufficient magnitude to result in this behavior.