— Several prominent cladists have questioned the importance of fossils in phylogenctic inference, and it is becoming increasingly popular to simply fit extinct forms, if they are considered at all, to a cladogram of Recent taxa. Gardiner's (1982) and Løvtrup's (1985) study of amniote phylogeny exemplifies this differential treatment, and we focused on that group of organisms to test the ... [Show full abstract] proposition that fossils cannot overturn a theory of relationships based only on the Recent biota. Our parsimony analysis of amniote phylogeny, special knowledge contributed by fossils being scrupulously avoided, led to the following best fitting classification, which is similar to the novel hypothesis Gardiner published: (lepidosaurs (turtles (mammals (birds, crocodiles)))). However, adding fossils resulted in a markedly different most parsimonious cladogram of the extant taxa: (mammals (turtles (lepidosaurs (birds, crocodiles)))). That classification is like the traditional hypothesis, and it provides a better fit to the stratigraphic record. To isolate the extinct taxa responsible for the latter classification, the data were successively partitioned with each phylogenetic analysis, and we concluded that: (1) the ingroup, not the outgroup, fossils were important; (2) synapsid, not reptile, fossils were pivotal; (3) certain synapsid fossils, not the earliest or latest, were responsible. The critical nature of the synapsid fossils seemed to lie in the particular combination of primitive and derived character slates they exhibited. Classifying those fossils, along with mammals, as the sister group to the lineage consisting of birds and crocodiles resulted in a relatively poor fit to data; one involving a 2—4 fold increase in evolutionary reversals! Thus, the importance of the critical fossils, collectively or individually, seems to reside in their relative primitive-ness, and the simplest explanation for their more conservative nature is that they have had less time to evolve. While fossils may be important in phylogenetic inference only under certain conditions, there is no compelling reason to prejudge their contribution. We urge systematists to evaluate fairly all of the available evidence.