S EAGRAVES AND PASOUR (hereafter S-P) offer a strong argument for accepting a cost-oriented definition of the "uneconomic regions" of the production function in the theory of the firm [3]. We hope in this note to make a few amendments to the S-P argument and, in so doing, to cast further doubt on the usefulness of the traditional "stages of production" notion as a pedagogical device. To illustrate the problems with conventional definitions of "economic region," S-P have employed a homogeneous production function. They also, implicitly at least, assumed the "fixed factor" to be perfectly divisible.1 We shall argue that: 1. In the case of a two-input production function, when one input is fixed and indivisible, the firm facing constant input and product prices will choose not to produce if output price falls below average variable costs. The reason for this is not that the marginal physical productivity of the "fixed input" is zero at the output corresponding to minimum average variable costs (this is true, however, for linear and homogeneous production functions), but rather that if the firm produces when product price is less than average variable costs, it will earn a negative quasi-rent. For some functions, the marginal physical productivity of the "fixed factor" may be negative at the minimum average variable cost output. The firm may wish to employ fewer units of that factor but is unable to do so because of its indivisible nature. Thus, production where the marginal physical productivity of an input is negative need not imply "uneconomic" production. 2. In the case of a two-input production