It is now well-established that phosphorylation of the 20,000-dalton light chain of smooth muscle myosin (LC20) is a prerequisite for muscle contraction. However, the relationship between myosin dephosphorylation and muscle relaxation remains controversial. In the present study, we utilized a highly purified catalytic subunit of a type-2, skeletal muscle phosphoprotein phosphatase (protein phosphatase 2A) and a glycerinated smooth muscle preparation to determine if myosin dephosphorylation, in the presence of saturating calcium and calmodulin, would cause relaxation of contracted uterine smooth muscle. Addition of the phosphatase catalytic subunit (0.28 microM) to the muscle bath produced complete relaxation of the muscle. The phosphatase-induced relaxation could be reversed by adding to the muscle bath either purified, thiophosphorylated, chicken gizzard 20,000-dalton myosin light chains or purified, chicken gizzard myosin light chain kinase. Incubation of skinned muscles with adenosine 5'-O-(thiotriphosphate) prior to the addition of phosphatase resulted in the incorporation of 0.93 mol of PO4/mol of LC20 and prevented phosphatase-induced relaxation. Under all of the above conditions, changes in steady-state isometric force were associated with parallel changes in myosin light chain phosphorylation over a range of phosphorylation extending from 0.01 to 0.97 mol of PO4/mol of LC20. We found no evidence that dephosphorylation of contracted uterine smooth muscles, in the presence of calcium and calmodulin, could produce a latch-state where isometric force was maintained in the absence of myosin light chain phosphorylation. These results show that phosphorylation or dephosphorylation of the 20,000-dalton myosin light chain is adequate for the regulation of contraction or relaxation, respectively, in glycerinated uterine smooth muscle.