To compare oil recovery and alkali consumption in alkaline flooding using sodium bicarbonate with other alkaline agents, coreflooding experiments were performed in turn with viscosified sodium bicarbonate and viscosified sodium carbonate solutions. Oil recovery was monitored, and the effluent brine from these corefloods was analyzed for silicon, aluminum, pH, and total inorganic carbon. The results indicate that viscosified sodium bicarbonate recovered more of the asphaltic Cerro-Negro crude than of the less asphaltic Wilmington crude oil. The recovery efficiency using the viscosified sodium carbonate was similar for the two crudes. For both crudes, the percent oil recovery using viscosified sodium carbonate was slightly higher than that using the viscosified sodium bicarbonate. Mineral dissolution and decrease in pH were found to be greater in corefloods using viscosified sodium carbonate. Total inorganic carbon recovery can be obtained in corefloods with either agent, provided that a sufficient water drive follows the chemical slug. Long-term experiments were performed by recirculating alkaline solutions through oil-free, unfired Berea sandstone to monitor the rock/alkali interactions. The experimental results indicate an eight-fold decrease in quartz dissolution by sodium bicarbonate compared with sodium carbonate. Moderate magnesium solubility was observed at the pH of the bicarbonate solution. Low solubility of magnesium and aluminum at the pH of the carbonate indicates the possible formation of precipitates. In these experiments 13% of the carbonate was converted to bicarbonate. Total alkalinity was not significantly decreased with either agent. 18 refs., 5 tabs.