Observational evidence analyzed by Eriksson <1965> and Weiss and Roether <1980> suggests that over the globe the ratio of tritium deposition into the ocean by vapor diffusion to that by rainfall should be near or slightly greater than two, while Koster et al. <1989> found in a general circulation model study that the diffusion to rainout ratio was closer to one. This study investigates the convective transport of tritium from the atmosphere to the ocean using a two-dimensional warm-rain cloud model. It is found that the deposition ratio is strongly dependent on the frequency and duration of rain events, with typical values for a monotonic tritium profile being about 1.0, but with values as low as 0.7 when long-duration events occur frequently and as high as 1.9 when convective events are short-lived and infrequent. On the basis of this study it appears that explicit treatment of convection in the general circulation model would not resolve the discrepancy in the deposition ratio between the model and the observations. It is also shown that the process of isotopic adjustment of tritium between the rain and the vapor phases is a key factor in determining the deposition ratio, since this process allows tritium to escape from the raindrops and ultimately diffuse to the surface. When tritium is ''frozen'' in the droplets, deposition ratios are reduced significantly.