Theoretical and experimental investigations of a constant pressure gravitydrainage system are reported. Experimental data are presented to show thatrecovery to gas breakthrough by gravity drainage is inversely proportional torate. The gravity drainage reference rate, which is numerically equal to theso-called "maximum theoretical rate of gravity drainage" is shown tohave no particular significance from a recovery standpoint. Before this ratecan be used as a basis of comparison for recoveries, it is necessary that therelative permeability and capillary pressure characteristics and displacingfluid viscosities be identical for the systems compared. A method is presented by which accurate prediction of the performance of agravity drainage system can be made. Close agreement between experimental andcalculated drainage performance shows that steady state relative permeabilityand static capillary pressure data can be used to describe fluid displacementbehavior. The very wide range of liquid recoveries before gas breakthroughwhich result from production rate variation alone demonstrates the importanceof this factor in planning depletion of a gravity drainage reservoir.Calculated results are presented which show that little additional recovery canbe expected from a high pressure gravity drainage system between the times ofgas breakthrough and attainment of such high gas/liquid ratios as to makefurther pressure maintenance impractical. Introduction: It has been recognized for some time that gravity forces play an importantpart in the recovery of oil from some types of reservoirs. Field experience hasshown that under certain conditions, gravity drainage can result in very highoil recoveries. Qualitative reasoning has led most engineers to the generalconclusions that:Where gravity drainage is important, the reservoir pressure should bemaintained by gas injection at the crest of the structure to prevent shrinkageof the oil in place and to keep a low viscosity so the oil can drain at thefastest possible rate.Recovery by gravity drainage is rate sensitive. A survey of the literature indicates that while considerable work has beendone on the effects of gravity in oil production problems, no satisfactorymethod of calculating the performance of gravity drainage reservoirs has beenreported. In the absence of any proven method of calculating reservoirperformance, the level at which pressure should be maintained and the rate ofproduction for most efficient operation has been open to debate. T.P. 3199