Both in Europe and the United States there is a growing interest in design research. One example is the design and validation of topic-oriented teaching-learning sequences. This research may be said to have two objectives. One is to design and test "useful products", such as teachers guides and study material for students, which may be put into practice in various ways. The second is to contribute to the development of educational science; for example, understanding conditions for learning of given topics under regular classroom conditions. This article concerns the latter objective and deals with the development of content-oriented theories stating conditions that promote learning with long-term understanding of given topics. We present one such theory, concerning evolution by natural selection, and describe the arguments and evidence that underlie the theory, which we regard as a well-founded hypothesis. Some methodological problems associated with testing this type of theory are discussed, as well as the role of content-oriented theories in strengthening science education research as an autonomous specialization within educational science. (Contains 3 notes.)