An investigation was made of the meteorological factors which a ffect the thermal structure of the ocean, and an attempt was made to develop techniques for predicting this structure. A prediction technique is proposed which is based on relationships giving the magnitude of the effects of radiation, evaporation and molecular conduction, convection, and turbulent mixing on the thermal structure. A sample prediction showed promise, but the data indicated that an accurate and operationally useful prediction method required a consideration of the internal waves which produce vertical motion. This factor was investigated for a 2-layer system and for a model in which the density was a continuous function of depth. The main considerations in the first model were seasonal and monthly variations, semidiurnal and diurnal tidal variations, short period variations, and random variations, Study of semidiurnal and diurnal tidal variations led to an empirical formula giving the amplitudes of oscillation in one area as a function of the thickness and density discontinuity of the 2-layer system. Where the data were adequate for a forecast, the formula predicted amplitudes with a fair degree of accuracy. Methods were also developed for predicting the internal oscillations of the continuous density model. A sample prediction showed good agreement at the greater depths; however, discrepancies occurred near the surface.