During the Winter 1871-72, in preparation of a series of conferences to be soon after delivered in Geneva, Walras jotted down some extended Notes where a striking number of concepts and constructs that will characterize the subsequent developments of Walras’s general equilibrium theory can be found. What is altogether missing in these Notes is the idea of optimizing behavior. Such idea, in fact, came to Walras’s mind only later, in the Fall 1872, when Walras, in the context of a pure-exchange, two-commodity economy, was able for the first time to establish an analytical link between utility maximization and the consumers-traders' demand functions. All the theoretical advances made by Walras over the highly productive period 1872-1877 can be viewed as attempts to revise his pre-existing theory in the light of the Fall 1872 discovery. Yet, the peculiar way in which Walras came to learn how to use the new approach, together with the persistence of his old ideas, not always compatible with a generalized assumption of optimizing behavior and its implications, can be shown to have conditioned and constrained Walras’s endeavors. In this paper we critically assess the exchange and production equilibrium models put forward by Walras in the period 1872-1877, explaining how the inconsistencies and shortcomings marring such models, of which Walras will try to get rid for the rest of his life, can be traced back to the mixed origin of Walras’s general equilibrium theory.