In nursing this essay through several drafts, I have benefited greatly from suggestions by Edward Denison, Robert Evenson, Zvi Griliches, Richard Levin, John Kendrick, Edwin Mansfield, and Richard Murnane. Moses Abramovitz has been a source ofencouragement and good, substantive editorial advice, for which I am most grateful. The heterodox views are my own, although I share many of them with Sidney Winter.

I. Introduction: Neoclassical theory and the evolutionary alternative, 90, — II. A specific evolutionary model, 93. — III. Calibration of the model, 98. — IV. Simulation results, 102. — V. Conclusion, 116.