Cambrian—because on their definition phyla appear in younger strata. They effectively
reduce the length of the phylogenetic ‘fuse’. The problems with this approach are
exemplified by the discovery of a fossil with n−1 synapomorphies—it seems perverse to
‘define out’ such a fossil from other animals it resembles in significant features merely
because such an animal did not win the lottery of survival (Wills and Fortey 2000). The
argument is well displayed in the discussion concerning a recently discovered ostracode-
like Early Cambrian fossil (see Siveter et al. 2001, and replies). Regardless of phylum
definitions, there are striking morphological differences between Early Cambrian
arthropods in the Chengjiang fauna of China (Bergstrom and Hou 1998) sufficient to place
them on different stem-lineages, which must themselves have split earlier: the problem of
deep divergences will not simply go away. Molecular estimates of divergences (e.g.
between chelicerates and crustaceans) will be based upon the split between the total-
groups, regardless of definitions of convenience.
Figure 3.2 Alternative topologies which have been used as consensus trees in divergence time
estimation papers. A, shown in Wang et al. (1998), Nei et al. (2001) and Hausdorf (2000). B, shown
in Wray et al. (1996) and Lynch (1999). Key: D=Deuterostomata, A=Arthropoda, N=Nematoda,
F=Fungi, P=Plantae.
PHYLOGENETIC FUSES AND EVOLUTIONARY ‘EXPLOSIONS’ 47