Although the North Cross (Devonian) Unit CO2 flood has not performed as dramatically as predicted, response to CO2 injection has been encouraging. Sharp, sustained production increases in responding wells and minimal CO2 breakthrough indicate that CO2 is displacing oil efficiently. No major operating problems have been encountered so far. CO2 injectivities have been significantly below initial predictions. Introduction: The Crossett Field (Crane and Upton counties, west Texas) was discovered in 1944 with the Texaco Hobbs A-1 well. Development of the 28-well field essentially was completed by the mid-1950's. The major producing horizon is the Devonian "Main Pay," a classical Archie Type 11-A rock composed of microscopic, tripolitic chert grains bound by limestone cement. The formation is characterized by high porosity, averaging 22%, and low uniform permeability, averaging 5 md. Connate water saturation averages 35%. The Crossett Devonian is a stratigraphic trap, bounded on the north, south, and west flanks by a deterioration of porosity, and on the east flank by the producing water porosity, and on the east flank by the producing water level of - 3,040 ft (Fig. 1). The reservoir oil was initially saturated (PVT properties are presented in Table 1), and a gas-oil contact existed at -2,860 ft. The gross pay distribution is described in Fig. 2. Net-to-gross pay ratios for individual wells range from 70 to 100%, and pay continuity among wells is very good. The primary producing mechanism of solution gas drive ultimately will recover 12.9% of the original oil in place. The field was developed on regular 40-acre spacing and produced under primary recovery until 1964. Crossett Field was unitized in 1964, when the North Cross (Devonian) Unit (operated by Shell Oil Co. with a working interest of 58.3%) was formed. Simultaneously, a program of partial pressure maintenance by updip program of partial pressure maintenance by updip residue-gas injection began. That program added 45% of the field's ultimate recovery to its recoverable reserves. Waterflooding is not feasible because of the reservoir's extremely low permeability to water; various other means of enhanced recovery were studied. Of these, miscible displacement by CO2 injection was the most attractive. When a CO2 source became available in 1970, a comprehensive CO2-flood design began. Initial Flood Design Implementation: A leasing agreement was signed between Shell and Canyon Reef Carriers (CRC), operator of the SACROC CO2 pipeline, whereby up to 20 MMcf/D of that system's capacity would be made available for CO2 transport to North Cross Unit. The CO2 was to be supplied from Shell's production in the JM and East Brown Bassett gas fields. The 20 MMcf/D (93% CO2, 7% methane) was to be delivered over a 10-year period and would constitute an ultimate solvent slug of 73 Bcf (about 40% of an oil-column hydrocarbon pore volume). Shell's miscible reservoir simulator was used to try to define an optimum flood plan. The field's past production performance was history matched, and project production performance was history matched, and project predictions for several injection patterns and operating predictions for several injection patterns and operating policies were generated and compared. An inverted policies were generated and compared. An inverted nine-spot pattern was found optimal. Wells 46 and 62 were converted to CO2 injection initially, with Wells 44 and 64 to be converted after 4 years of flooding. Continuous solvent injection with recycling was planned. JPT P. 1706