When one encounters a problem that one has no adequate way of representing, a new mental model for the problem may have to be found. When a possible model is found, an important next step is to evaluate the validity of the model. In this document, examples of subjects finding and evaluating mental models used as problem representations are discussed in several case studies which depict expert scientists solving problems. This paper focuses on a non-deductive strategy called "bridging" that is used to evaluate the validity of a mental model and that has been observed in solutions to both science and mathematics problems. In constructing a bridge, the subject creates an intermediate case that is seen as "in between" the proposed model and the problem situation because it shares important features of both. A bridge can help the subject confirm or deny the validity of the analogy relation between the model and the problem. It is suggested that the bridging strategy observed in experts can be used to help students construct and refine new mental models. Numerous figures are provided. (Author/TW)