Results of a wide-ranging investigation of the scaling of gas injection processes are reported. The research described examines how the physical mechanisms at work during a gas injection project interact to determine process performance. In particular, we examine: 1. The interactions of equilibrium phase behavior and two-phase flow that determine local displacement efficiency and minimum miscibility pressure, 2. The combined effects of viscous fingering, gravity segregation and heterogeneity that control sweep efficiency in two- and three-dimensional porous media, 3. The use of streamtube/streamline methods to create very efficient simulation technique for multiphase compositional displacements, 4. The scaling of viscous, capillary and gravity forces for heterogeneous reservoirs, and 5. The effects of the thin films and spreading behavior on three-phase flow. The following key results are documented: 1. Rigorous procedures for determination of minimum miscibility pressure (MMP) or minimum miscibility enrichment (MME) for miscibility have been developed for multicomponent systems. 2. The complex dependence of MMP's for nitrogen/methane floods on oil and injection gas composition observed experimentally is explained for the first time. 3. When gravity segregation and viscous fingering are both important, two-dimensional calculations of displacement performance do not reflect accurately what happens in three-dimensional flows. Gravity segregation is more important in three-dimensional flow than it is in two-dimensional flow. 4. The presence of layer-like heterogeneities strongly influences the interplay of gravity segregation and viscous fingering, as viscous fingers adapt to preferential flow paths and low permeability layers restrict vertical flow. 5. Streamtube/streamline simulation techniques are demonstrated for a variety of injection processes in two and three dimensions. The resulting calculations are orders of magnitude faster than conventional finite difference simulation, and they are free of adverse effects of numerical dispersion. 6. Quantitative scaling estimates for the transitions from capillary-dominated to gravity dominated to viscous-dominated flows are reported. 7. Experimental results are given that demonstrate that high pressure CO2 can be used to generate low IFT gravity drainage in fractured reservoirs if a suitably connected fracture network exists. 8. The effect of the underlying physical phenomena of wetting and spreading behavior on three-phase flow is described.