We used confluent cultures of dog gallbladder epithelial cells, stimulated by conditioned medium from a culture of human neonatal foreskin fibroblasts, to establish the presence of inducible nitric oxide synthase (NOS, EC 1.14.13.39). Assay was by conversion of radiolabeled arginine to citrulline. By 4 days after addition of the conditioned medium, a relatively high level of activity was observed. However, further study showed that the enzyme did not require addition of the usual cofactors for maximal activity (NADPH, FAD, FMN and tetrahydrobiopterin) and was stable in the absence of anti-proteolytic agents. Our suspicion that this enzyme might not be NOS but arginine deiminase (EC 3.5.3.6) was confirmed by enzyme purification and by the liberation of ammonia during enzyme reaction. This enzyme, which is absent from primates and virtually confined to single-cell organisms, suggested the presence of Mycoplasma, a common contaminant of cell cultures, and it was subsequently confirmed that the fibroblast culture was a source of Mycoplasma. With the widespread interest in nitric oxide and NOS, and common use of the convenient [3H]arginine assay, there is a considerable danger of the two enzymes being confused. At the very least, it is necessary to check for activity in the absence of added cofactors.