Gene duplication is of central interest to evolutionary developmental biology, having been implicated in evolutionary increases in complexity. These ideas stem principally from the Lewis model for the evolution of the BX-C and Ohno's proposal for genome duplications during chordate evolution. Here I revisit these models and show how recent data have confirmed their essential features, but forced some important revisions. These include revised dates for homeotic gene duplications and for widespread gene duplication in vertebrate evolution. I also outline the major unresolved questions in the study of gene duplication, and its relevance to evolution and development.