Testing the Cambrian explosion: some approachesWe have argued above that the Cambrian explosion needs to be considered in its many
different guises—especially ecological and biogeographical—if we are to understand its
rapidity. Nevertheless, mere argumentation will not resolve the pressing problems
presented by mismatches between the fossil record and the molecular clock, where they
occur. Can analytical or modelling approaches take us further? We would like to suggest,
sketchily, two ways forward: the use of extinction events for estimating diversification
speeds, and the use of phylogenetic tests of the stratigraphic record.
Extinction events and estimates of fossilization attainmentThe major difficulty that is presented by trying to compare fossil and molecular estimates
of divergence times is that there is rarely, if ever, an independent test of either. In the
case of fossil evidence, the only information available is the first appearance datum of a
clade, combined with the subsequent pattern of diversity. Fundamentally, the best we can
say about the clade’s pre-fossil history is a probabilistic or likelihood statement, but there
are few ways of checking the accuracy of our estimates (one possibility is provided by
finds of new fossils that fall outside the known range, and comparing them with the
estimate of the true range from the previously existing fossils; Marshall 1990). One under-
utilized method in this general context, however, would appear to be provided by mass
extinctions. If a clade vanishes from the record at an extinction event, only to reappear
some time later, one reason may be a sudden decline in diversity/abundance, followed by
a ‘lag’ period of time during which the numbers of the organism are not high enough to
ensure the finding of a fossil record (Wignall and Benton 1999; but see also Twitchett
2000, for a discussion of this model and Gale et al. 2000, for the alternative preservation-
driven model). If an extinction is extremely severe, then this scenario will approximate to
a true origin of the clade and the lag between the extinction and the first appearance in the
fossil record will give some measure of the length of time that is required for a diversification
to be recorded.
Considerable interest has recently been directed at these ‘lag’ or ‘survival’ intervals
(see Erwin 2001 for a helpful review; Harries and Little 1999), which suggests that the
logistic models applied to such diversifications may underestimate the complexity of the
underlying processes. For example, empirical studies sometimes point to prolonged
periods of biotic crisis, followed by extremely rapid positive-feedback enhanced recoveries.
Although there is some variability, the evidence from these recoveries is that restocking—
even with an environmentally disturbed survival period—is unlikely to be more than a few
million years, and sometimes less. This also applies to radiation events whose beginning may
be accurately fixed by other events, such as the monograptid radiation at the beginning of
the Silurian (Koren and Bjerreskov 1999). Although neither of these cases is directly
applicable to the Cambrian explosion, they do give some basis for examining the lag times
in general that exist between the time of origin of a clade (or the time of its severe
restriction) and its subsequent (re)appearance in the fossil record. Both these cases suggest
that, contrary to many molecular studies, the lag time is in general no more than a few
million years. Clearly, in order to develop this theme further would require careful
180 DATING THE ORIGIN OF BILATERIA