Meridional heat transport in the tropical Pacific is estimated using a linear numerical model with realistic boundaries and forced by 18 years of observed wind, covering the period from January 1962 to December 1979. The long-term mean heat transport estimated in this study is similar to the estimates based on heat balance and radiation considerations and on complex numerical models that account for thermodynamics as well. This points to the dominant role played by the adiabatic process, the only heat transport mechanism present in this study, in the heat balance for the equatorial Pacific. The combined Ekman and geostrophic heat transport can account for the net meridional heat transport, except near the equator, where continuity requirements dictate. The Ekman and geostrophic transport oppose each other, and their small difference in magnitude gives rise to the net meridional heat transport, resulting in transport away from the equator for the southern hemisphere and north of 6¿N, while for the band between 6¿N and the equator an equatorward transport is present. Seasonal and interannual variations are found to be as large as, or even larger than, the long-term mean. Seasonal variations in meridional heat transport are in accordance with seasonal variations in zonal winds via Ekman transport, while the geostrophic transport remains more or less constant on the time scale. The results are a net poleward heat transport in the winter hemisphere and a equatorward transport in the summer hemisphere. At the interannual time scale, variabilities in both Ekman and geostrophic components contribute to the interannual variability in heat transport. Major features of the interannual variations in meridional heat transport appear to be associated with the El Nino events. It is interesting to note that the interannual variations associated with El Nino events are not restricted to the near-equatorial region. Phase locking between the interannual variations and the annual cycle is evident in the data set. Major findings in this study, based on an adiabatic model, are expected to carry over to more realistic nonlinear numerical models.