FOSSILS AND SEED PLANT PHYLOGENY REANALYZED

In a cladistic analysis of Recent seed plants, Loconte and Stevenson (1990) obtained results that conflict with our 1986 analysis of both extant and fossil groups and argued that fossil data had led us to incorrect conclusions. To explore this result and the general influence of fossils on phylogeny reconstruction, we assembled new "Recent" and "Complete" (extant plus fossil) data sets incorporating new data, advances in treatment of characters, and those changes of Loconte and Stevenson that we consider valid. Our Recent analysis yields only one most parsimonious tree, that of Loconte and Stevenson, in which conifers are linked with Gnetales and angiosperms (anthophytes), rather than with Ginkgo, as in our earlier Recent and Complete analyses. However, the shortest trees derived from our Complete analysis show five arrangements of extant groups, including that of Loconte and Stevenson and our previous arrangements, suggesting that the result obtained from extant taxa alone may be misleading. This increased ambiguity occurs because features that appear to unite extant conifers and anthophytes are seen as convergences when fossil taxa are interpolated between them. All trees found in the Complete analysis lead to inferences on character evolution that conflict with those that would be drawn from Recent taxa alone (e.g., origin of anthophytes from plants with a "seed fem" morphology). These results imply that conclusions on many aspects of seed plant phylogeny are premature; new evidence, which is most likely to come from the fossil record, is needed to resolve the uncertainties.