An experiment has been performed to measure directly the effect of atmospheric turbulence on spherical wave propagation of light on a vertical path. The results permit determination of the magnitude of the effect at a variety of locations and under different conditions. Agreement is found with the functional form predicted by the inertial theory of turbulence. General agreement is found with scattering data obtained by a low-frequency radar and with predictions of some turbulence profile models. The results also demonstrate some reasons for poor agreement with time integrating sensors.