Here, the Reynolds stresses hu9w9i and hy9w9i, where u9, y9, and w9 are the cross-shore, alongshore, and vertical turbulence velocities, respectively, and the angle brackets represent time averaging, are used to diagnose turbulence dynamics beneath natural breaking surf-zone waves. The data were collected at Truc Vert Beach, France, during a 12-day period in 1-3-m water depth with strong cross-shore and alongshore currents under high-energy wave conditions (offshore significant wave heights ranged between 2 and 8 m). The hu9w9i term is predominantly negative, increases with the ratio of wave height H s to water depth h (;degree of wave breaking), and decreases in magnitude toward the bed. This supports the view that the cross-shore shear stress is due to breaking-induced vortices that transport high-speed cross-shore flow downward and disintegrate close to the bed. The occasional positive sign of hu9w9i within the lower 15%-20% of the water column indicates that sometimes surface-generated turbulence is overwhelmed by bed-generated turbulence, but the conditions when this happens are not clear from the data. The term hy9w9i is persistently of opposite sign to the alongshore mean current and decreases with height above the seabed, implying that hy9w9i is due to bottom boundary layer processes rather than surface-generated turbulence. The bottom drag coefficient amounted to 1.6 3 10 23 , similar to earlier observations. As in other high-Reynolds-number geophysical flows, time series of u9w9 and y9w9 comprise intermittently large, short-duration (here, ;1 s) stress events that in the data contribute considerably to the net stress in only 3%-15% of the time. The data further show that the turbulent kinetic energy is depth uniform and increases with H s /h. The depth-averaged Froudescaled turbulent kinetic energy beneath surf-zone bores is 0.025, a factor of 2 to 3 less than observed beneath regular laboratory waves.