This assumes that because the evolutionary process is time-dependent and because the
number of individual clocks (=genes) is vast, it is expected that their results will converge
on average values reflecting the time elapsed since the divergence of species. In practice,
datasets are always filtered for sequences that exhibit deviation from clock-like behaviour.
The implications of such sequences for molecular evolution are not readily apparent and
represent an interesting and important subject for study in their own right (Pawlowski &
Berney, Chapter 6).
However, sources of much of the controversy surrounding the mechanics of molecular
clock analyses lie elsewhere. These include phylogenetic uncertainty, the unreliability of
calibration points, neglect of confidence intervals both on calibration points and on clock
estimates, and problems with accurately relating sequence changes to rate variation
(Fortey et al., Chapter 3). These are further exacerbated by the fact that such potential
errors are cumulative and progressively distant clock analyses will be subject to gross
absolute errors.
In addition to these specific errors, molecular clock estimates, palaeontological
estimates, and attempts to correlate the two, are all hostage to a geological timescale that
is under continual revision and, thus, the choice of timescale is non-trivial. Dating
techniques are advancing at a considerable pace, with large increases in precision, and a
timescale that is a decade old is likely to have been supplanted by more accurate ones—
although the pace of advance is uneven in different parts of the column. In addition, the
accuracy of available dates varies throughout the geological column, with poorly dated
intervals not restricted to the oldest periods of geological history. Unless all of these
sources of error are taken into account, in addition to attempts to correlate fossil
occurrences to the geological timescale, and those errors attendant to molecular clocks
themselves, errors will propagate, potentially beyond the age of the events being
estimated.
This volume presents analyses by experts in the mechanics and application of molecular
clocks as well as analyses of the fossil records of protists (Pawlowski & Berney,
Chapter 6), plants (Wellman, Chapter 7; Wikström et al., Chapter 8), the divergence of
complex animals (Fortey et al., Chapter 4; Budd & Jensen, Chapter 9), chordates
(Donoghue et al., Chapter 10), tetrapods (Ruta & Coates, Chapter 11), and modern birds
(Dyke, Chapter 12). Although there is evidence for creeping increases in palaeontological
estimates, as might be expected, there remain serious disagreements over the time of
origin of major groups such as land plants and animals. Interestingly, disagreement does
not follow disciplinary boundaries suggesting that conference between molecular
biologists and palaeontologists heralds a realizable consilience in the future.
In concluding, we would like to thank all of the contributors to the symposium and the
volume for their enthusiasm and willingness to look anew upon their old datasets and
cherished assumptions. We wish to express our gratitude also to the reviewers for their
hard work and commitment in the face of anonymity, and to Mandy Donoghue for
assisting with production of this volume. Finally, we wish to acknowledge the support of
the Palaeontological Association and Systematics Association for co-sponsoring the
symposium.
Philip Donoghue and Paul Smith
INTRODUCTION 3