Many fossil animals bear traits such as crests or horns that probably functioned as sexually selected signals or weapons. Interpretations of these structures as functioning in mate choice or intrasexual contests are often controversial, with interpretations based on biomechanics or physiology being favoured by many. Although testing hypotheses based on sexual selection can be difficult, especially given that there is no single, reliable means of recognising sexual selection, we argue that it is not impossible; indeed, there are now several cases where sexual selection is strongly supported. In other cases, a careful study of features such as sexual dimorphism, ontogeny, and allometry, coupled with testing of alternative hypotheses, will be necessary to distinguish between possible explanations for exaggerated features. © 2012 Elsevier Ltd.