AbstractA definition of convective vertical heat flux is proposed which leads directly to the expressionp1w1h1wherep1,w1, andh1are, respectively, simultaneous values of the fluctuating density, vertical velocity relative to the surface, and total heat per unit mass at a point in a constant pressure surface. Neglect (apparently justifiable) of the pressure fluctuations yields the expressioncpp±w1/Rfor the convective flux of sensible heat. The distinction between free and forced convection is made. Recent treatments by Montgomery and by Priestley and Swinbank are examined with particular attention to the importance of free convection. It is submitted that neither Montgomery's neglect of the buoyancy forces nor Priestley and Swinbank's arguments suggesting their universal importance can at present be justified. In particular the latter authors explanation of the normal tropospheric lapse rate is questioned