Abstract
This paper investigates the justification of the concept of a balance of nature in population ecology as a case of model based reasoning. The ecologist A.J. Nicholson understood balance as an outcome of intraspecific competition in populations. His models implied density dependent growth of populations oscillating around an equilibrium state. Today the assumption of density dependence is tested statistically by using models that represent certain data dynamics. This however, does not test for density dependence in the sense suggested by Nicholson. From a suggested mechanism in nature equilibrium has become a property of a data set. I call this a change of reference of the term equilibrium . The new equilibrium is a necessary but not sufficient condition for the old equilibrium. However, some ecologists suggest using the new definition instead of the old one. I argue that the new definition of equilibrium points to a change in justification of the assumption of a balance in nature. This change has to be seen in relation to modeling strategies pursued by the scientific community. I argue that there is an interdependence between models, data, values, and evidence. Model based reasoning, though abstract and idealizing, has to be considered in a historical, philosophical, and social context