In rat jejunal brush-border membranes (BBM), ATP hydrolysis activity was specifically stimulated by CaCl2 and by MgCl2, allowing to identify Ca(2+)-ATPase and Mg(2+)-ATPase activities with a broad pH optimum near 8.0. Nonspecific ATPase activity (in the absence of cations) had a pH optimum above 9.5 as alkaline phosphatase. The effects of Ca2+ and Mg2+ concentrations on ATPase activity evidenced two apparent KA for each cation. At high concentrations, a similar affinity for both cations was recorded (KA: 0.35 mM). At low concentrations, the affinity for Mg2+ was greater than for Ca2+ (KA: 0.02 mM and 0.07 mM respectively). In an attempt to differentially solubilize alkaline phosphatase and ATPase activities, eleven different detergents were assayed. They more or less successfully released Ca(2+)-ATPase and Mg(2+)-ATPase activities from BBM but the more membranes were solubilized by a detergent, the more activities were lost, suggesting a close dependence on integration in BBM. As to alkaline phosphatase and nonspecific ATPase, they almost co-solubilized with Ca(2+)-ATPase and Mg(2+)-ATPase but their total activity was little affected. After treatment of BBM with phosphatidylinositol-specific phospholipase C (E.C. 3.1.4.10), 58% of alkaline phosphatase activity and 45% of nonspecific ATPase activity were released in the supernatant while Ca(2+)-ATPase and Mg(2+)-ATPase activities remained totally incorporated in BBM pellets. These last results definitively demonstrate that Ca(2+)-ATPase and Mg(2+)-ATPase activities are not manifestations of alkaline phosphatase, as earlier suggested, but rather result from the existence of one or several intrinsic membrane enzymes.