Diffuse co-evolution over geological timescales: Plants and non-avian dinosaurs

Category Paleontology and Stratigraphy
Group GSI.IR
Location International Geological Congress,oslo 2008
Author Barrett, Paul; Butler, Richard; Kenrick, Paul; Penn, Malcolm
Holding Date 03 September 2008

Although the significance of co-evolution over ecological timescales is well established, it is unclear whether this process also drives macroevolutionary and macroecological changes over geological time. Many deep-time co-evolutionary scenarios have been suggested, but few of these hypotheses have been tested adequately. Some of the most pervasive and intuitively appealing diffuse co-evolutionary hypotheses relate to proposed interactions between Mesozoic floras and non-avian dinosaur herbivores. In particular, it has been suggested that changes in dinosaur diversity and/or feeding behaviour mediated: the radiation of cycads in the Triassic and their subsequent decline in the Cretaceous; the abundance and distribution of various conifer families (e.g. Araucariaceae); and the origin and diversification of angiosperms. However, detailed consideration of the fossil record indicates that the evidence supporting these hypotheses is generally weak. All co-evolutionary scenarios require that: the taxa concerned lived in the same geographical areas and environments at the same time; that there are correlated changes in species-diversity, distributions and/or evolutionary innovations in the groups under consideration; and that there is persuasive direct evidence of interactions between the groups (e.g. trace fossils). In all cases, hypotheses of dinosaur-plant co-evolution fail at least one of these criteria (usually all three) and cannot be supported on the basis of current data. This partly results from the nature of the terrestrial fossil record, which is patchily distributed in both time and space, and from the nature of the fossils themselves, which do not provide unequivocal evidence of definitive biological interactions.
Quantitative tests to assess the degree of co-occurrence between dinosaur and plant taxa, implemented using a Geographical Information System, do not find any significant associations between particular clades, nor any evidence that such co-occurrences changed in a correlated pattern through time. This case study suggests that caution is necessary in the extraction of diffuse co-evolutionary signals from palaeontological data. Although diffuse co-evolutionary processes may operate over geological timescales, such hypotheses need to be firmly grounded in direct evidence of interaction and may be difficult to support given the patchiness of the fossil record.