The authors fing lag phases exceeding 20 min in measuring creatine kinase activity, by using the kinetic creatine phosphate → creatine assay, in sera of some patients with carcinoma metastatic to the liver. Such long lag phases are accompanied by a decreased apparent enzyme activity. These problems are eliminated by adding sulfhydryl agents to the serum before assay, but not by adding more of such agents to the assay reagent. β Mercaptoethanol is superior to Cleland's reagents, glutathione, and cysteine. The long lag phases could not be explained by inadequate activity of the coupling enzymes, interference with the coupling steps, high proportions of cardiac isoenzyme activity, simple oxidation of the enzyme, low concentrations of albumin, or increased concentrations of glutathione reductase, lactate dehydrogenase, or uric acid. It is concluded that the prolonged lag phases reflect inadequate reactivation of the enzyme by sulfhydryl agents under the usual assay conditions. Reactivation before assay can prevent potentially serious negative errors in the assay of creatine kinase.