Consider the changing morphospace within the stem group of a phylum when it is traced back through successively earlier nodes from the ancestor of the crown group (the crown ancestor) to its last common ancestor with a sister phylum. The crown synapomorphies are lost immediately. Some stem-group branches may be quite diverse, and their derived features may be so morphologically striking that in Linnean classification these branches would rank as extinct classes or orders of the phylum, ranks they can not be assigned in cladistic classifications. Proceeding toward the last common ancestor with a sister phylum, the constellation of features continue to erode. Eventually the morphologies at the nodes on the main line toward the common ancestor become dominated by features that are plesiomorphic. Finally, unique features of the phylum can no longer be recognized, and by definition the phylum is no longer present.
The above quote occurs on page thirty-one of Valentine’s book On the Origin of Phyla. Earlier Valentine had said:
Stem-group taxa expand the morphological features and, usually, the morphological disparity of a phylum beyond that of a crown group, while at the same time they lack crown-group synapomorphies.
This has some interesting implications for primate and human evolution. To see how this plays out in primate evolution you should read this post by Hawks. I had just finished reading Valentine’s book when Berger announced Australopithecus sebida and Curnoe announced Homo gautengensis. Some anthropologists (Clarke, for example) have long thought there were more than several species of australopithecine in South Africa and the same can be said for early Homo. Although Valentine was referring specifically to phyla above, the basic ideas in the above quotes can be applied to any crown group and stem group(s) – collectively known as a total group. More on this later.