species did coexist it is possible that they might be preserved in the fossil record in the
wrong order. Take the conservative view that when species did coexist the fossil record is
unreliable and so we should reject it. We will only accept the fossil record of species
which did not coexist. Then the probability that the fossil record is reliable depends on
the proportion of all fossil species that did not coexist. Assuming five million years for the
average lifetime of a species and about 500 million yearssince the base of the Cambrian,
and making due allowance for changes in biodiversity through the Phanerozoic (e.g. from
Sepkoski 1978; Sepkoski et al. 1981), this suggests that 3–5 per cent of fossil species
coexisted at any given time in the Phanerozoic. Put another way, in 95–97 per cent of
random comparisons between pairs of fossil species there is no chance whatsoever that
they could be preserved in the wrong stratigraphic order. Recently, Peters and Foote
(2001) have questioned the reality of the diversity increase apparent in Sepkoski (1978)
and Sepkoski et al.’s (1981) analyses. If their re-interpretation is correct, this would
reduce the proportion of taxa that coexisted in the Phanerozoic even further and
strengthen the above arguments. Furthermore, this is a conservative estimate of the
reliability of the fossil record. Even when species did coexist, the probabilities are that their
first known occurrences will still be in the correct stratigraphic order. More detailed
arguments to support these conclusions are presented in Paul (1982).
This argument becomes weaker when applied to higher taxa, largely because the
periods of coexistence of higher taxa tend to be large relative to the differences between
their times of origin. For example, most phyla have existed at least from the Cambrian to
the present day. Detecting their order of origin in the Cambrian reliably is significantly
more difficult because sampling this relatively brief interval adequately for all phyla is less
likely than sampling the 500+million years since they evolved. However, first
stratigraphic records of all phyla are based on species. Equally, even cladograms depicting
relationships between higher taxa use one or a few ‘representative’ species to characterize
the higher taxa. To my mind a similar problem exists in both cases. Does the
stratigraphically earliest species in a phylum closely approximate to the real time of
origin, that is, is it truly representative of the time of origin? Similarly, are the species
Figure 5.4 A. Total durations of two fossil taxa that do overlap. B. An example where the first
occurrence of taxon V (solid circle) could occur above (later than) the first occurrence of taxon W
(solid square) (i.e. in the wrong order with respect to their real order of origin). Note, however,
that as soon as an example of V is found in level one (open circle), it is no longer possible for the
first occurrences of these taxa to be in the wrong order with respect to their real order of origin.
GHOST RANGES 97