Near-neutral measurements of the turbulent wind field within and above a sequence of 15 parallel windbreaks on a flat pastoral site are presented. The windbreak fences each had a porosity of 60% and were equally-spaced at 6 times their height (h = 2 m). The following conclusions seem justified for wind directions within 10 ° of the normal to the array: 1. Above the windbreaks (2 h), mean windspeeds first decreased and then increased asymptotically to a value in equilibrium with the new surface roughness. At 0.5 h, windspeeds exhibited a slow increase down the entire array. 2. Reflecting differences in approach flows, the drag on the initial fence was almost twice that on barriers farther downstream. This reduction in momentum extraction per windbreak was associated with an elevation in the zero-plane displacement to a level equal to 0.8 h. 3. At positions well-removed from the initial fences, mean windspeeds were reduced throughout the entire region below shelter height. In this region, the flow became increasingly dominated by downward moving air with velocities much greater than the local average. The zone of reduced turbulence was small, extending only 2 h downstream of a barrier at a height of 0.25 h. This corresponded with the region excluded from smoke trails released at the top of windbreaks. 4. An approximate TKE budget mid-way between windbreaks 7 and 8 suggests that shear and wake production peak near z = h and that production is balanced by dissipation and vertical transport components. Advective and inertial interaction terms are negligible at this midway position but are likely to be major sources of TKE closer to the windbreak. Local equilibrium is attained above z = 1.5 h implying the existence of a constant-stress layer. The measurements show the practical difficulty of simultaneously reducing both mean windspeeds and turbulence levels with repeated windbreaks at conventional spacings for horticultural applications.