Fossil butterflies are extremely rare. Yet, they are the only direct evidence of the first appearance of particular characters and as such, they are crucial for calibrating a molecular clock, from which divergence ages are estimated. In turn, these estimates, in combination with paleogeographic information, are most important in paleobiogeographic considerations. The key issue here is the correct allocation of fossils on the phylogenetic tree from which the molecular clock is calibrated.The allocation of a fossil on a tree should be based on an apomorphic character found in a tree based on extant species, similar to the allocation of a new extant species. In practice, the latter is not done, at least not explicitly, on the basis of apomorphy, but rather on overall similarity or on a phylogenetic analysis, which is not possible for most butterfly fossils since they usually are very fragmentary. Characters most often preserved are in the venation of the wings. Therefore, special attention is given to possible apomorphies in venational characters in extant butterflies. For estimation of divergence times, not only the correct allocation of the fossil on the tree is important, but also the tree itself influences the outcome as well as the correct determination of the age of the fossil. These three aspects are discussed. All known butterfly fossils, consisting of 49 taxa, are critically reviewed and their relationship to extant taxa is discussed as an aid for correctly calibrating a molecular clock for papilionoid Lepidoptera. In this context some aspects of age estimation and biogeographic conclusions are briefly mentioned in review. Specific information has been summarized in four appendices.