The turbulent transfer of heat from Arctic leads in winter is one of the largest terms in the Arctic heat budget. Results from the AIDJEX Lead Experiment (ALEX) suggest that the sensible component of this turbulent heat flux can be predicted from bulk quantities. Both the exponential relation N = 0.14R:' 72 and the linear relation N = 1.6 x lO-'R, + 1400 fit our data well. In these, N is the Nusselt number formed with the integrated surface heat flux, and R, is the Reynolds number based on fetch across the lead. Because of the similarity between heat and moisture transfer, these equations also predict the latent heat flux. Over leads in winter, the sensible heat flux is two to four times larger than the latent heat flux.The internal boundary layer (IBL) that develops when cold air encounters the relatively warm lead is most evident in the modified downwind temperature profiles. The height of this boundary layer, S, depends on the fetch, x, on the surface roughness of the lead, zO, and on both downwind and upwind stability. A tentative, empirical model for boundary layer growth iswhere L is the Obukhov length based on the values of the momentum and sensible heat fluxes at the surface of the lead, and p is a constant reflecting upwind stability. Velocity profiles over leads are also affected by the surface nonhomogeneity. Besides being warmer than the upwind ice, the surface of the lead is usually somewhat rougher. The velocity profiles therefore tend to decelerate near the surface, accelerate in the mid-region of the IBL because of the intense mixing driven by the upward heat flux, and rejoin the upwind profiles above the boundary layer. The profiles thus have distinctly different shapes for stable and unstable upwind conditions.