Taking into current literature of didactics of Geography, there is a lack of clear and coherent criteria concerning teaching-related experiments. Up to now there is a wide range of different criteria to refer to so that in concrete teaching situations the teachers’ own perception of how experiments should be used plays an important role. Against this background this paper firstly is meant to develop a workable definition of using experiments in Geography. Secondly this paper aims at analyzing and evaluating the results of six guided and standardized interviews with teachers on the subject of “Experiments in Geography lessons” based on the model of Didactical Reconstruction. These findings show that teachers are able to name main criteria for experiments, however the meaning of a control approach often seems to be unclear. Furthermore, the teachers‘ goal to achieve more precise clarity by using experiments partly correlates with the undifferentiated equating of experiments and models.