Stock models, in which production is interpreted as if it were the population growth of a stock, have been the preferred tool for fishery economics since Clark and Munro (1975) introduced capital theory in these models. Ravn-Jonsen (2009c) applied capital theory to a model in which the production in the ecosystem is a consequence of predator-prey interaction and the somatic growth of the predator as a result of this interaction. By deducing the results of Clark and Munro anew, the assumptions of the stock model are clarified. Four different biomass measures are introduced in the ecosystem model as stocks. The optimum point found with the stock model approach is compared with the optimum point found in the ecosystem model with the capital value calculations of the occurring rent flow. A comparison shows that the stock model fails to generate the correct optimal point. The assumptions behind the use of stock models for species population models are discussed. The population stock model corresponds to a holistic community view, which has in fact failed to explain various phenomena. The production of the marine ecosystem cannot be reduced to a model as if the production were a consequence of the growth of a stock. The concept of a stock is rather an illusion, as is the concept of an optimal stock level. It is essential to liberate fishery economics from a simplified view of population and communities.