IN HIS stimulating article, Convexity Assumption in the Theory of Competitive Markets,1 Mr. M. J. Farrell proposes reduce the minimum assumptions necessary for the validity of the 'received theory' demonstrating the traditional assumptions of convexity are no means essential to the optimality of competitive markets. In the process, he chides Mr. J. de V. Graaff and me for having engaged in facile attacks on the theory of competitive allocation, in debunking which should be taken with a grain of salt. In particular, he claims that Graaff and I were led into elementary error by a rather farfetched analogy between competitive markets and one-man economies. The purpose of this note is (1) to substantiate a plea of not guilty to Farrell's charge; and (2) to raise a question about his suggestion for broadening the basis of the received theory. I shall restrict myself to discussing only those parts of the Farrell paper which deal with non-convexity in production functions, partly because it is production non-convexities which are alleged to have led me astray, and partly because, in an article in the October, 1960, issue of this Journal,2 Jerome Rothenberg has dealt rather fully with Farrell's treatment of nonconvex preferences.3