Vertical turbulent transports of latent and sensible heat through the tropical marine boundary layer were measured with ship and airborne instruments in the GATE experimental area. The measurements from the two systems are compared for cases of simultaneous sampling in the same locations during undisturbed weather and during the wakes of convective disturbances. The paired average fluxes from the airborne eddy correlation measurements and the shipboard bulk aerodynamic measurements, for undisturbed weather, are related with correlation coefficients in the range of 0.6 to 0.9; the correlations depend primarily on stratifications of the aircraft data according to sampling altitude (15-153 m msl) and flight pattern relative to the mean wind. The agreement between the two types of measurements is best for ship data paired with aircraft samples from the lowest altitudes (15 and 30 m) and from alongwind rather than crosswind Rights, as is appropriate since the stationary ships monitor the alongwind characteristics of the turbulence. The use of long (l-3 hr) versus short (10 min) ship samples did not significantly affect the comparability with the aircraft samples during undisturbed periods.The good baseline comparability demonstrated by the undisturbed-weather comparisons was applied to interpret the results from the wakes behind showers. Although these paired measurements were few, not only the sensible heat, but also the latent heat flux was shown to decrease substantially with altitude from 10 m to at least 150 m in the wakes, where the mixing was relatively intense. Variations in the fluxes were much greater at the higher level than nearer the sea surface.