1. Paired iris smooth muscles from rabbits were incubated for 30min at 37°C in an iso-osmotic salt medium containing glucose, inositol, cytidine and [32P]phosphate. 2. One of the pair was then incubated at 37'C for 10min in unlabelled medium containing lOmM-2-deoxyglucose and the other was incubated in the presence of acetylcholine plus eserine (0.05mM each). 2-Deoxyglucose, which was included in the incubation medium to minimize the biosynthesis of triphosphoinositide from ATP and diphosphoinositide, decreased the amount of labelled ATP by 71 % and inhibited further 32p incorporation from ATP into triphosphoinositide by almost 30 %. 3. Acetylcholine (0.05mM) increased significantly the loss of 32P from triphosphoinositide (the 'triphosphoinositide effect') in 32P-labelled iris muscle. This effect was measured both chemically and radiochemically. It was also observed when 32P1 was replaced by myo-[3H]inositol in the incubation medium.4. The triphosphoinositide effect was blocked by atropine but not by D-tubocurarine. Further, muscarinic but not nicotinic agonists were found to provoke this effect. 5. Acetylcholine decreased by 28% the 32p incorporation into triphosphoinositide, presumably by stimulating its breakdown. This decrement in triphosphoinositide was blocked by atropine, but not by D-tubocurarine. 6. The triphosphoinositide effect was accompanied by a significant increase in 32p labelling, but not tissue concentration, of phosphatidylinositol and phosphatidic acid. The possible relationship between the loss of 32p label from triphosphoinositide in response to acetylcholine and the concomitant increase in that of phosphatidylinositol and phosphatidic acid is discussed. 7. The presence of triphosphoinositide phosphomonoesterase, the enzyme that might be stimulated in the iris smooth muscle by the neurotransmitter, was demonstrated, and, under our methods of homogenization and assay, more than 80% of its activity was localized in the particulate fraction.