Abstract

Principles of phylogenetic analysis (cladistics) are introduced with an examination of relationships among extant genera of Gnetales. The Gnetales can be supported as a monophyletic group, with Gnetum and Welwitschia more closely related to each other than either is to Ephedra. Characters of the progymnosperm Archaeopteris and 19 extinct and extant seed plant taxa are then reviewed as a basis for a cladistic analysis of their interrelationships. The seed plant taxa included are: medullosans, cycads, Lyginopteris, Cordaixylon, Mesoxylon, Lebachia, extant conifers, Ginkgo, Callistophyton, peltasperms, glossopterids, Caytonia, corystosperms, Bennettitales, Pentoxylon, Gnetum, Welwitschia, Ephedra, and angiosperms. Preliminary analyses of relationships within conifers, glossopterids, and Bennettitales are also presented. Results suggest that seed plants are a monophyletic group, and Lyginopteris is resolved as the sister taxon to all other seed plants considered. The cordaites, Cor- daixylon and Mesoxylon, along with Lebachia, extant conifers, and Ginkgo constitute a monophyletic group. Pentoxylon is the sister taxon to Bennettitales, and the Gnetales are the sister group to angio- sperms. Together the Gnetales plus the angiosperms form the sister group to the Bennettitales plus Pentoxylon. If the outer integument of bennettitalean and angiosperm ovules is interpreted as ho- mologous with the "cupule" of Caytonia and corystosperms, then the corystosperms are resolved as the sister group to the Bennettitales plus Pentoxylon plus Gnetales plus angiosperm clade. Under this interpretation all the seed plant taxa considered except Lyginopteris, cycads, and medul- losans are part of a single clade in which flattened seeds and saccate pollen are primitive. The principal difficulties with the cladistic analysis concern necessary inferences on unknown characters in certain plants, and the current absence of a large base of comparative data. Results of the analysis suggest that the seed ferns as currently circumscribed are not a meaningful group for phylogenetic purposes and permit an evaluation of the possible phylogenetic position of Eremopteris, Nystroemia, Sper- mopteris, Phasmatocycas, Vojnovskiales, Leptostrobus, and several other enigmatic groups of plant fossils. Comparison of the phylogenetic analysis with previous theories of angiosperm origin shows that it reconciles the ideas of Arber and Parkin, that Bennettitales and Gnetales are closely related to flowering plants, with more recent hypotheses that invoke Caytonia and corystosperms as highly relevant to the angiosperm problem. The results suggest that increased understanding of Triassic Bennettitales, Gnetales, and corystosperms will be of maximum interest in further elucidating the phylogenetic relationships of flowering plants.