SUMMARYEnhancement of bulk surface fluxes by subgrid variability is reviewed, with special attention to conditions ranging from fair-weather free convection to precipitating deep convection over uniform surfaces. Aircraft measurements in the Tropical OceanlGlobal Atmosphere Coupled Ocean-Atmosphere Response Experiment subcloud layer indicate that the predominant mechanism for flux enhancement is wind variability associated with both boundary-layer-scale coherent eddies (dry convective gustiness) and density currents beneath precipitating convection (moist convective gustiness). A new evaluation of the dry convective gustiness parameter, /?, supports recent indications that its effective value is less than previously thought (0.60-0.80, rather than 1 .OO-1.25), due both to a discrepancy in the technique often used for its estimation from field data, and to contamination of datasets by meso-y-scale (-10 km) motions. A strong relationship is found in the aircraft time series between the moist convective gustiness velocity and the theoretical speed of density-current gust fronts driven by the observed meso-y -scale temperature variability. This prompts development of a simple theoretical model for the latter, based upon a known convective rainfall rate and ambient thermodynamic structure for the lower atmosphere. Parameters in the model are estimated using data from simultaneous aircraft and shipboard radar measurements, and model behaviour is compared with a simpler scheme developed by MCtto-France. It is concluded that an explicit representation of the rain evaporation process in the generation of density perturbations is crucial for accurate determination of the strength of moist convective gustiness in the model, and through it the degree of surface flux enhancement.