Following up the earlier findings by Fuchs on surgeon-induced demand, this paper makes numerous data and econometric improvements in conducting a test of neoclassical and inducement theories. A simultaneous equation model is used to estimate physician demand and equilibrium fees for surgery from a sample of 350 PSUs over the 1969-76 period. The results provide definite support for the notion of competitive market failure--particularly in large metropolitan areas. Other things equal, fees and utilization are higher in surgeon-rich areas although our estimated shift elasticities were only about one-third those found by Fuchs. A statistically significant, albeit small price elasticity of demand for surgery was also obtained, in contrast to Fuchs. Increasing monopoly and disequilibrium models are also tested without altering the basic findings. Where surgeons were in short supply, their availability significantly affected surgery rates, although a small supply effect was found in plentiful areas as well.