imply in turn having a model that links abundance to diversity during an evolutionary
radiation. In part, this link may depend on the balance between sympatric and allopatric
speciation mode.
Recent modelling, for example, has suggested that during a diversification, the carrying
capacity of a particular environment is first attained rather quickly, followed by instability
in the population distribution and break-up into distinct subpopulations and eventual
sympatric speciation (Drossel and McKane 2000). Competitive displacement of characters
through selection may lead to the introduction of phenotypic and therefore resource
exploitation ‘no-go’ zones, with the surprising theoretical result being that the total
number of individuals may actually drop during a radiation (see figs 1 and 2 of Drossel and
McKane 2000). The effect of increasing ecological complexity on carrying capacity is of
course a famous problem, first mentioned by Darwin, but not yet satisfactorily answered.
The initial stages of the radiation will indeed be marked by a rapid increase of population
as the newly available resource spectrum is exploited, but this may be on a geologically
instantaneous timescale. In any case, the relationship between abundance (proportional to
rate of fossilization) and diversity is clearly not at all a straightforward one, and it seems to
us that one cannot simply use the latter as a proxy for the former. For these reasons, we
remain sceptical of current attempts to model the preservability of the early stages of
radiations (Foote et al. 1999; Tavaré et al. 2002).
Even if the above objections could be taken into account, there remain the particular
problems presented by the Cambrian fossil record (see Smith and Peterson 2002 for a
concise summary). And here, some attempts to reconcile the fossil record with molecular
estimates do not seem to do the job required. Smith and Peterson (2002), following Knoll
and Carroll (1999), emphasized that, in general, bilaterian divergence times are being
measured by molecular clocks, not metazoan origin itself. They both adopt the model that
a tail of stem-group bilaterian diversification in the Precambrian precedes crown-group
bilaterian radiation in the Cambrian explosion itself. Yet this formulation conceals an
ambiguity between the members of the stem-group of Bilateria and the members of the
stem-groups of bilaterian phyla. The importance of this distinction is that we can have
some idea about what stem-group members of bilaterian phyla were like on phylogenetic
grounds, whereas we have much less idea currently about the stem-group of the Bilateria.
The stem-group of the Bilateria is in fact somewhat irrelevant to the molecular problem,
because molecular clock estimates are typically of the time of divergence of crown-, not
stem-group bilaterians. Even if some stem-group bilaterians were tiny planktonic animals
(but for critical discussion of this idea see Budd and Jensen 2000), the upper stem-group
would still have been characterized by the progressive appearance of the monophyletic
features of crown-group Bilateria and these features must have been assembled in large
animals (Budd and Jensen 2000).
Molecular methods have, in general, dated splits between existing phyla such as
vertebrates and echinoderms, and as discussed in Budd and Jensen (2000), these sorts of
taxa are bound, on phylogenetic grounds, to share important features such as a coelom.
Even if caution should be employed in inferring morphological homology from shared
developmental systems (e.g. Budd 2001; Erwin and Davidson 2002), there are still
classical grounds for thinking that many important metazoan features are homologous. A
recent renewed interest in convergence (e.g. Conway Morris 1998a) should not lead to
176 DATING THE ORIGIN OF BILATERIA