Seed plant phylogeny and the origin of angiosperms: An experimental cladistic approach

We present a numerical cladistic (parsimony) analysis of seed plants plus progymnosperms, using characters from all parts of the plant body, outgroup comparison, and a method of character coding that avoids biases for or against alternative morphological theories. The robustness of the results was tested by construction of alternative trees and analysis of subsets of the data. These experiments show that although some clades are strongly supported, they can often be related to each other in very different but nearly equally parsimonious ways, apparently because of extensive homoplasy. Our results support Rothwell’s idea that coniferopsids are derived fromCallistophyton- like platyspermic seed ferns with saccate pollen, but the hypothesis that they evolved fromArchaeopteris- like progymnosperms and the seed arose twice is nearly as parsimonious. Meyen’s division of seed plants into radiospermic and primarily and secondarily platyspermic lines is highly unparsimonious, but his suggestion that ginkgos are related to peltasperms deserves attention. Angiosperms belong among the platyspermic groups, as the sister group of Bennettitales,Pentoxylon, and Gnetales, and this “anthophyte” clade is best related toCaytonia and glossopterids, although relationships with other combinations of Mesozoic seed fern taxa are nearly as parsimonious. These results imply that the angiosperm carpel can be interpreted as a modified pinnate sporophyll bearing anatropous cupules (=bitegmic ovules), while gnetalian strobili are best interpreted as aggregations of highly reduced bennettitalian flowers, as anticipated by Arber and Parkin and Crane. Our most parsimonious trees imply that the angiosperm line (though not necessarily all its modern features) extended back to the Triassic, but a later derivation of angiosperms from some species ofCaytonia or Bennettitales, which would be nearly as parsimonious, should also be considered. These results raise the possibility that many features considered key adaptations in the origin and rise of angiosperms (insectpollinated flowers, rapid reproduction, drought tolerance) were actually inherited from their gymnospermous precursors. The explosive diversification of angiosperms may instead have been a consequence of carpel closure, resulting in increased speciation rates due to potential for stigmatic isolating mechanisms and/or new means of dispersal. DNA sequencing of extant plants and better information on anatomy, chemistry, sporophyll morphology, and embryology of Bennettitales and Caytoniales and the morphological diversity of Mesozoic anthophytes could provide critical tests of relationships.