Wind sensors mounted on towers and smokestacks do not always indicate the true free-air flow. To determine the probable errors in measurements of wind speed and direction around such structures, quarter-scale models have been tested in a large wind tunnel. Data on changes in wind speed and direction were obtained by using smoke, very small wind vanes, and a scale model propeller anemometer. Most emphasis has been placed on a relatively open lattice-type tower, but a solid tower and a stack were also studied. The analysis shows that in the wake of lattice-type towers disturbance is moderate to severe, and that in the wake of solid towers and stacks there is extreme turbulence, with reversal of flow. Recommendations for locating wind sensors in the wind field relative to the supporting structure are given for each of the three structures studied. Guidelines are suggested regarding probable errors in measurements of wind speed and direction around different supporting structures, as outlined below. For an open triangular tower with equal sides D, the wake is about 1-1/2D in width for a distance downwind of at least 6D. Sensors mounted 2 D out from the corner of such a tower will usually measure speeds within ± 10° of that of the undisturbed flow for an arc of about 330°. The disturbance by very dense towers and stacks is much greater. Wind sensors mounted 3 diameters out from the face of a stack will measure wind speeds within ± 10%, and directions within ± 10° of the undisturbed flow for an arc of about 180°.