We have argued previously (Budd and Jensen 2000) that the likelihood of the lack of
convincing Precambrian crown-group bilaterian fossils cannot be explained by recourse to
a step-change in preservation potential (e.g. a rapid increase in body size or development
of hard parts). Ancestral bilaterians were of some reasonable size, in order to need
muscles and other important organ systems that are almost universally considered to be
present in the ur-bilaterian. Furthermore, the total silence of any sort of convincing trace
fossil record before about 555–560 Ma argues against the hard part hypothesis, as does the
plentiful preservation of non-mineralized, non-bilaterian organisms throughout much of
the Proterozoic, especially in the terminal Proterozoic. Similarly, any possible argument
that the lack of bilaterians in the Proterozoic is because of lack of possible host rocks could
not (especially on its own) explain a total absence for so long. Almost every Cambrian
sedimentary rock yields a few bilaterian fossils of some sort, even the unpromising basal
sandstones; no Proterozoic sedimentary rock before the late terminal Proterozoic has ever
yielded any such thing. If brachiopods and trilobites had been as abundant in the
Proterozoic as they are in the Cambrian, they surely would have been found.
The possible explanation must therefore come down to problems of diversity or
abundance. We have argued here that diversity, although beloved of modellers, is the
wrong metric to use in assessing fossilization potential; and that abundance should increase
extremely rapidly at the base of a radiation, especially with few ecological impediments to
the spread of the new organisms. Furthermore, if the widespread temporary
disappearance of taxa after mass exinctions is owing to restriction of their abundance, then
some generalized metric is available to estimate how long recovery (considered here as a
proxy to time from origination) takes before appearance in the fossil record. On all of
these grounds, it remains extremely hard to imagine that the time from the origination of
crown-group bilaterians to their widespread appearance in the fossil record was more than
a few million years.
If the fossil record is moderately reliable in this regard, then the suggestion must be
that molecular estimates of the origins of bilaterians are largely inaccurate. Potential
sources of inaccuracy have been widely discussed in the literature (e.g. Bromham in
press). More recent ones include the suggestion that molecular clock estimates have an
upwards bias (Bromham et al. 2000; Rodríguez-Trelles et al. 2002) because the bounds of
the molecular estimates of divergence times are asymmetrical (they are rigidly bound to
be non-negative, but non-rigidly bounded at their upper boundary).
In many cases, dissatisfaction with attempts to prove that clock-like conditions pertain
has led to efforts to relax the clock assumption, perhaps by modelling rates with more
complex distributions than the Poisson distribution of the clock model. Perhaps the most
insightful investigation into the complexities of relaxing the clock assumption is that of
Gillespie (1991), who builds up an ‘episodic clock’ model of molecular evolution. Noting
that the average rate of nucleotide substitution on a site basis is approximately once per
billion years, with protein structure changing on a timescale of tens of millions of years,
he suggests that this sort of timescale is far too long to account for changes that take place
during speciation, especially if the latter is driven by environmental change on the scale of
a few thousand years. The inevitable consequence must be that at least some molecular
evolution takes place in highly concentrated bursts, separated by long periods of
quiescence. Although some tests of this possibility have not revealed any sudden bursts of
186 DATING THE ORIGIN OF BILATERIA