Convection in a layer of fluid with a free surface due to a combination of thermocapillary stresses and buoyancy is a classic problem of fluid mechanics. It has attracted increasing attentions recently due to its relevance for two-phase cooling. Many of the modern thermal management technologies exploit the large latent heats associated with phase change at the interface of volatile liquids, allowing compact devices to handle very high heat fluxes. To enhance phase change, such cooling devices usually employ a sealed cavity from which almost all noncondensable gases, such as air, have been evacuated. Heating one end of the cavity, and cooling the other, establishes a horizontal temperature gradient that drives the flow of the coolant. Although such flows have been studied extensively at atmospheric conditions, our fundamental understanding of the heat and mass transport for volatile fluids at reduced pressures remains limited. A comprehensive and quantitative numerical model of two-phase buoyancy-thermocapillary convection of confined volatile fluids subject to a horizontal temperature gradient has been developed, implemented, and validated against experiments as a part of this thesis research. Unlike previous simplified models used in the field, this new model incorporates a complete description of the momentum, mass, and heat transport in both the liquid and the gas phase, as well as phase change across the entire liquid-gas interface. Numerical simulations were used to improve our fundamental understanding of the importance of various physical effects (buoyancy, thermocapillary stresses, wetting properties of the liquid, etc.) on confined two-phase flows. In particular, the effect of noncondensables (air) was investigated by varying their average concentration from that corresponding to ambient conditions to zero, in which case the gas phase becomes a pure vapor. It was found that the composition of the gas phase has a crucial impact on heat and mass transport as well as on the flow stability. A simplified theoretical description of the flow and its stability was developed and used to explain many features of the numerical solutions and experimental observations that were not well understood previously. In particular, an analytical solution for the base return flow in the liquid layer was extended to the gas phase, justifying the previous ad-hoc assumption of the linear interfacial temperature profile. Linear stability analysis of this two-layer solution was also performed. It was found that as the concentration of noncondensables decreases, the instability responsible for the emergence of a convective pattern is delayed, which is mainly due to the enhancement of phase change. Finally, a simplified transport model was developed for heat pipes with wicks or microchannels that gives a closed-form analytical prediction for the heat transfer coefficient and the optimal size of the pores of the wick (or the width of the microchannels).