Abstract
It is suggested that an additional source of semidiurnal forcing due to daily variations in tropical rainfall could correct the discrepancy between the calculated phase (boned on forcing due to insulation absorption by ozone and water vapor) and that observed for the surface pressure oscillation. It is also shown that the 180° phase shift in horizontal wind oscillations at 28 km which current calculations predict, but which is not observed, would he eliminated by the proposed additional forcing. The magnitude and phase of the required rainfall oscillation is calculated and found to be consistent with existing observations. Finally, it is shown that the convergence field due to the tide could not directly account for the rainfall oscillation.