This paper analyses the traditionally recognized dependence between observation statements and theories. The analysis proceeds by working out the interrelationship between classification systems and theoretical frameworks. Cuvier's and Darwin's theories are used as examples to illustrate this issue. The second part of the paper develops a model designed to give an account of the historical development of this interrelationship. It is argued that the interdependence is not circular and that it is an integral part of scientific research. It is suggested (...) that an interdependence of the type reflected in the model can be helpful in evaluating research programmes in the history of all the sciences. (shrink)