relatively poorly understood, the lack of understanding arising from poor quality data in
heterostracans and from poorly resolved relationships in both groups (cf. Benton et al.
1999).
Conodonts and heterostracans are important for providing constraints on molecular
estimates for the divergence of the living jawless vertebrates and jawed vertebrates, and it
is therefore a happy coincidence that they appear to possess a fossil record that exhibits
internal consistency. Assessments of the quality of the record of groups that might
constrain the divergence of crown-group jawed vertebrates suggest that implied
divergence dates may be less reliable. Groups such as the osteostracans, placoderms, and
chondrichthyans bracket this diversification event, and potentially provide important
upper bounds on divergence timing. However, cladogram-stratigraphy correlation in
these groups is poor, generally at a level of 50 per cent for the SCI, at ≤0.5 for the RCI
(except placoderms which appear to have a fossil record that is internally more
consistent), and with a GER ≈0.3–0.5; SCI values do not pass a 95 per cent confidence
test to determine whether they are not significantly worse than random, although GER
and RCI values are generally no worse than random at the same confidence level. Positive
correlation between high and low GER and SCI values suggests that poor cladogram-
stratigraphy correlation does not arise solely from cladogram inaccuracy, which would
normally produce an inverse correlation, although cladogram inaccuracy is possibly an
important factor. It is more likely that correlated low GER and SCI values reflect a
Figure 10.5 Cladogram from Figure 10.3 calibrated to time and including 95 and 99 per cent
confidence intervals on stratigraphic range data; medium thickness dark grey lines represent 95 per
cent confidence intervals and thin light grey lines represent 99 per cent confidence intervals.
PHILIP C.J.DONOGHUE ET AL. 207