Five orders of late Paleozoic-Mesozoic seed ferns have, at one time or another, figured in discussions on the origin of angiosperms, even before the application of phylogenetic systematics. These are the Glossopteridales, Peltaspermales, Corystospermales, Caytoniales, and Petriellales. Although vegetative features have been used to suggest homologies, most discussion has focused on ovulate structures, which are generally interpreted as megasporophylls bearing seeds, with the seeds partially to almost completely enclosed by the megasporophyll (or cupule). Here we discuss current information about the reproductive parts of these plants. Since most specimens are impression-compression remains, homologizing the ovulate organs, deriving angiospermous homologues, and defining synapomorphies remain somewhat speculative. Although new specimens have increased the known diversity in these groups, a reconstruction of an entire plant is available only for the corystosperms, and thus hypotheses about phylogenetic position are of limited value. We conclude that, in the case of these seed plants, phylogenetic analysis techniques have surpassed the hard data needed to formulate meaningful phylogenetic hypotheses. Speculation on angiosperm origins and transitional stages in these fossils provides for interesting discussion, but currently it is still speculation, as the role of these groups in the origin of angiospermy continues to be cloaked in Darwin's mystery.