Recent work has determined the depth of the Mohorovicic discontinuity at sea and has made it likely that peridotite xenoliths in basaltic volcanic rocks are samples of material from below the discontinuity. It is now possible to produce a hypothetical section showing the transition from a continent to an ocean. This section is consistent with both the seismic and gravity results. The possible reactions of the crust to changes in the total volume of sea water are discussed. It seems possible that the oceans were shallower and the crust thinner in the Archean than they are now. If this were so, some features of the oldest rocks of Canada and Southern Rhodesia could be explained. Three processes are described that might lead to the formation of oceanic ridges; one of these involves tension, one compression and the other quiet tectonic conditions. It is likely that not all ridges are formed in the same way. It is possible that serpentization of olivine by water rising from the interior of the earth plays an important part in producing changes of level in the ocean floor and anomalies in heat flow. Finally, a method of reducing gravity observations at sea is discussed.