An equation for the daytime lapse rate of temperature as a function of z/L is derived assuming the lapse rate varies as z−⅔ in very light winds. This variation is presumed to result from the geometry of buoyant plumes. The equation, which agrees very well with some data taken by Dyer, predicts a maximum value for the lapse rate at a fixed height and with given heat flux, at intermediate values of z/L. This variation suggests the concept of eddy viscosity as usually applied does not adequately describe the heat flux in unstable conditions.