Abstract This study of the effect of gravity segregation was carried out in a vertical core of unconsolidated sand, 4 ft long and 2 in. in diameter. Fluids of varying viscosity and density were injected into the top of the core and moved at constant rates of 25, 50 or 100 ft/day by use of a positive displacement pump. The composition of the efflux was analyzed by measurement of the refractive index. The main variables in this study were in viscosity ratio, the density difference between in-place and displacing phases, and the rate of flow. The experiments fell into four logical groups:favorable viscosity ratio and favorable density difference;favorable viscosity ratio and unfavorable density difference;unfavorable viscosity ratio and favorable density difference; andunfavorable viscosity ratio and unfavorable density difference. The behavior of these four systems at various rates of flow was determined by measuring the length of the mixing or transition zone which developed between the displaced and displacing phases. The data indicated that gravity segregation could act to shorten this mixing zone when the displacing material was the less dense phase, and lengthen the zone for unfavorable density differences. The magnitude of the effect was most marked at slow rates when sufficient residence time existed to allow significant flow to occur. The change in the length of the mixing zone with density difference, rate and viscosity ratio was plotted, and from these graphs it became evident that the length of the mixing zone was really dependent upon the ratio of the viscous forces to the gravity forces. If the ratio of these quantities is used as a parameter, we find that the plots of mixing zone length vs the dimensionless quantity V/Vc, yields a greatly simplified presentation of the data. Such plots confirm the importance of the ratio of the viscous forces to the gravity forces in analyzing the flow behavior in vertical systems. Introduction The idea of using miscible phase displacements to increase recovery of crude oil has been under active study for more than 15 years. Most of this work has been carried out in the research laboratories of the industry and in universities. In such studies a great deal of attention has been directed toward determining the amount of miscible material required to recover all of the oil from a porous medium. The nature and extent of the mixing zone which develops between the two miscible materials has also been observed for many situations, including those where the fluids used have different densities. In many of the cases where density differences exist, however, the effect of gravity segregation on the mixing zone has not been explicitly studied. It seems quite likely that serious errors might be present in such experimental work. That is, the literature presents certain data which suggest changes in efficiency of the process as a result of some explicit change in a variable, when it may well be that the change in the efficiency results from gravity effects. The objective of this study, therefore is to determine the effect of density differences of miscible fluids on the observed efficiency of the displacement process. Since such density differences cause selective movement of fluids within the porous medium, the study becomes vitally concerned with gravity segregation. Moreover, since the amount of gravity segregation which can occur for a given set of fluids depends upon residence time, the rate of flow in these experiments also becomes an important variable. Finally, since the actual distribution of the fluids within the system (fingers, etc.) also affects the results, the viscosity ratio is another pertinent variable. Hence, this study may be characterized as a determination of the nature of the mixing zone between miscible phases as a function of density differences, viscosity ratio, and rate of flow. Fortunately these variables are inter-related. SPEJ P. 1ˆ