Inositol 1,4,5-trisphosphate [Ins(1,4,5)P3] and inositol 1,4-bisphosphate [Ins(1,4)P2] phosphatase activities were measured in both 180,000 g (60 min) particulate and supernatant fractions of rat brain homogenates. Although Ins(1,4,5)P3 was mostly hydrolysed by a particulate phosphatase [Erneux, Delvaux, Moreau & Dumont (1986) Biochem. Biophys. Res. Commun. 134, 351-358], Ins(1,4)P2 phosphatase was predominantly soluble. The latter enzyme was Mg2+-dependent and sensitive to thiol-blocking agents (e.g. p-hydroxymercuribenzoate). In contrast with Ins(1,4,5)P3 phosphatase activity measured in the soluble fraction, Ins(1,4)P2 phosphatase was insensitive to 0.001-1 mM-2,3-bisphosphoglycerate. Lithium salts, widely used in psychiatric treatment, inhibited both Ins(1,4)P2 and Ins(1)P1 phosphatase activities of the crude soluble fraction. In particular, 50% inhibition of phosphatase activity, with 2 microM-Ins(1,4)P2 as substrate, was achieved at 3-5 mM-LiCl. At these concentrations, LiCl did not change Ins(1,4,5)P3 phosphatase activity measured in the same fraction with 1-4 microM-Ins(1,4,5)P3 as substrate. Chromatography of the soluble fraction of a rat brain homogenate on DEAE-cellulose resolved three phosphatase activities. These forms, peaks I, II and III, dephosphorylated Ins(1,4,5)P3, Ins(1)P1 and Ins(1,4)P2 respectively. If LiCl (10 mM) was included in the assay mixture, it inhibited both peak-II Ins(1)P1 phosphatase and peak-III Ins(1,4)P2 phosphatase, suggesting the existence of at least two Li+-sensitive phosphatases.