The history of anomalies shows that empirical theories are by no means defenseless against conflicts with observational data. By contrast, they have refined immunization mechanisms at their disposal, which can be utilized to protect their fundamental principles. This lecture is devoted to some considerations concerning these immunization mechanisms. I will argue that a special variant of the semantic view of empirical theories, namely the structuralist approach, is especially well suited for the treatment of conflicts of this type. As an illustrative example, I will use one of the most famous anomalies discovered in the nineteenth century: the anomaly of Mercury's perihelion. First, let me present a brief sketch of Mercury's anomaly.