A complete separation of myo-inositol 1,4,5-[4,5-(32)P]trisphosphate prepared from human erythrocytes, and myo-[2-3H]inositol 1,3,4-trisphosphate prepared from carbachol-stimulated rat parotid glands [Irvine, Letcher, Lander & Downes (1984) Biochem. J. 223, 237-243], was achieved by anion-exchange high-performance liquid chromatography. This separation technique was then used to study the metabolism of these two isomers of inositol trisphosphate in carbachol-stimulated rat parotid glands. Fragments of glands were pre-labelled with myo-[2-3H]inositol, washed, and then stimulated with carbachol. At 5s after stimulation a clear increase in inositol 1,4,5-trisphosphate was detected, with no significant increase in inositol 1,3,4-trisphosphate. After this initial lag however, inositol 1,3,4-phosphate rose rapidly; by 15s it predominated over inositol 1,4,5-trisphosphate, and continued to rise so that after 15 min it was at 10-20 times the radiolabelling level of the 1,4,5-isomer. In contrast, after the initial rapid rise (maximal within 15s), inositol 1,4,5-trisphosphate levels declined to near control levels after 1 min and then rose again very gradually over the next 15 min. When a muscarinic blocker (atropine) was added after 15 min of carbachol stimulation, inositol 1,4,5-trisphosphate levels dropped to control levels within 2-3 min, whereas inositol 1,3,4-trisphosphate levels took at least 15 min to fall, consistent with the kinetics observed earlier for total parotid inositol trisphosphates [Downes & Wusteman (1983) Biochem. J. 216, 633-640]. Phosphatidylinositol bisphosphate (PtdInsP2) from stimulated and control cells were degraded chemically to inositol trisphosphate to seek evidence for 3H-labelled PtdIns(3,4)P2. No evidence could be obtained that a significant proportion of PtdInsP2 was this isomer; in control tissues it must be less than 5% of the total PtdInsP2 radiolabelled by myo-[2-3H]inositol. These data indicate that, provided that inositol 1,4,5-trisphosphate is studied independently of inositol 1,3,4-trisphosphate, the former shows metabolic characteristics consistent with its proposed role as a second messenger for calcium mobilization. The metabolic profile of inositol 1,3,4-trisphosphate is entirely different, and its function and source remain unclear.