Pterosaur diversity: Untangling the influence of sampling biases, Lagerstätten, and genuine biodiversity signals
Highlights
► We compare pterosaur richness to sampling using multiple regression and other methods. ► Substantial temporal and environmental heterogeneity exists in the pterosaur record. ► Observed diversity is driven largely by ‘background’ sampling and Lagerstätten. ► Peak richness/disparity may have been driven by the pterodactyloid radiation.
Introduction
Pterosauria, a group of flying Mesozoic reptiles, was a taxonomically (130–150 currently recognised species) and ecologically diverse group with a long evolutionary history (Late Triassic–Late Cretaceous: ~ 145 Myr) (Wellnhofer, 1991, Unwin, 2005, Barrett et al., 2008, Butler et al., 2009, Butler et al., 2012). Pterosaurs are best known from Lagerstätten (sites of exceptional preservation) such as the Late Jurassic lithographic limestones of the ‘Solnhofen’ area of southern Germany (Wellnhofer, 1970, Wellnhofer, 1975), or the Lower Cretaceous Jehol Group of China (Wang and Zhou, 2006), and their thin-walled bones are often poorly preserved in higher-energy depositional settings. This taphonomic bias has led several workers to propose the existence of a ‘Lagerstätten effect’, whereby our understanding of pterosaur macroevolutionary patterns is limited by the highly heterogeneous sampling of their fossil record (e.g. Wellnhofer, 1991, Buffetaut, 1995, Buffetaut et al., 1996, Buffetaut et al., 1997, Butler et al., 2009, Butler et al., 2012, Prentice et al., 2011; but see Dyke et al., 2009).
Over the last decade, there has been vigorous debate about the extent to which sampling biases (e.g. temporal variation in sedimentary rock volume, collecting biases) affect species-richness counts and other biodiversity metrics in the fossil record (e.g. Alroy et al., 2001, Peters and Foote, 2001, Smith, 2001, Smith, 2007, Peters, 2005, Smith and McGowan, 2007, Fröbisch, 2008, Barrett et al., 2009, Butler et al., 2009, Butler et al., 2011, Butler et al., 2012, Marx, 2009, Alroy, 2010, Benson et al., 2010, Peters and Heim, 2010, Benson and Butler, 2011, Benton et al., 2011, Upchurch et al., 2011, Benson and Mannion, 2012, Dunhill, 2012, Lloyd et al., 2012). Pterosaurs were one of the first vertebrate groups for which explicit, quantitative comparisons were made between taxonomic richness and potential sampling biases. Butler et al. (2009) compared the generic- and species-richness of pterosaurs (as well as a partially ‘corrected’ phylogenetic richness estimate that used a stratigraphically calibrated tree to include ghost lineages) to the number of geological formations that have yielded pterosaur remains (pterosaur-bearing formations: PBFs) through time. They recovered strong statistically significant correlations between these metrics. A simple modelling approach (Raup, 1972, Smith and McGowan, 2007) was then applied in an attempt to ‘correct’ their richness counts for uneven fossil record sampling. However, the dominant signal in the resultant ‘corrected’ diversity curve seemed to be simply the presence or absence of Lagerstätten. Butler et al. (2009) therefore suggested that accurate reconstruction of pterosaur richness patterns might prove extremely difficult. Subsequently, Butler et al. (2012) examined pterosaur morphological diversity (disparity) through time using discrete characters, and made comparisons to geological sampling through time. They concluded that measures of disparity based upon estimating the total range of morphospace occupied by a group of taxa (range-based disparity) were strongly biased by sample-size differences resulting from uneven sampling of the fossil record through time, but that measures of disparity that examine the variance of taxa within morphospace (variance-based disparity) were more robust to, although not immune from, sampling biases.
The work of Butler et al. (2009) was critiqued by Benton et al. (2011), who reported that correlations between pterosaur species-richness and PBFs were not robust to data transformations (generalised differencing, in this case) that are designed to remove the influence of trend and temporal autocorrelation. However, Benton et al. (2011) inadvertently miscalculated the statistical results for the comparison of phylogenetic species-richness and PDFs using generalised differenced data (G. T. Lloyd, pers. comm., 22 March 2012), an error subsequently corrected in an erratum (Benton et al., 2012). In addition, Benton et al. (2011, pp. 74–77) suggested that, in any case, any correlation between pterosaur diversity and PBFs was not meaningful because PBFs did not represent a ‘global’ sampling metric, and were not independent of (or, were “redundant with”) taxonomic counts of pterosaurs. In other words, at times when pterosaurs were genuinely rare in Mesozoic ecosystems, PBF counts would also be low, regardless of the actual opportunities available to sample their fossil record. In conclusion, Benton et al. (2011) agreed with Butler et al. (2009) that a strong Lagerstätten effect exists in the pterosaur record, but disagreed with regard to the extent to which this effect is superimposed upon a more general, underlying sampling bias on richness that reflects sedimentary rock availability through the Mesozoic.
Here, we revisit our previous analyses (Butler et al., 2009), using an updated dataset derived from ongoing addition of new data on pterosaur spatiotemporal distributions, combined with more sophisticated methodological approaches, including time series multiple regression modelling. Our aims are to consider further the degree to which the pterosaur fossil record is biased by a global tetrapod fossil record sampling signal, and to assess the importance of Lagerstätten in determining our understanding of pterosaur evolutionary history. In addition, we make explicit comparisons between pterosaur taxonomic diversity and disparity through time, and explore the relationship between sampling of the fossil record of pterosaurs and sampling of the fossil record of dinosaurs, the most abundant group of terrestrial Mesozoic vertebrates.
Section snippets
Paleobiology Database dataset
Previous analyses of pterosaur richness (Butler et al., 2009, Benton et al., 2011) were based upon the pterosaur distribution dataset published by Barrett et al. (2008). This dataset has been added to the Paleobiology Database (PBDB), using the original primary literature to enter collections information. A substantial amount of pterosaur data had already been added to the PBDB by several other contributors (see Acknowledgements), and this data was double-checked and updated as necessary.
Relationship between specific- and generic-richness
Our genus-level taxonomic richness estimate is very strongly correlated with species-level taxonomic richness using stage-level bins (r2 = 0.95). As a result of the strength of this correlation, we use species-richness exclusively in all subsequent results and discussion.
Observed patterns of species-richness
Stage-level bins show a pattern in which species-richness is initially moderately high (during the Norian, nine species) but declines in the Rhaetian before increasing slowly to a low Toarcian peak (Fig. 2A). This Toarcian peak
Relationship between pterosaur disparity and species-richness, and implications for choice of disparity metric in macroevolutionary studies
Disparity and observed species-richness for pterosaurs show qualitatively similar patterns through time, with an increase in all disparity metrics and in observed species-richness occurring from the Late Triassic to a Late Jurassic/Early Cretaceous peak, followed by a Late Cretaceous decline. Quantitatively, correlations are much stronger between range-based disparity metrics and observed species-richness than between variance-based disparity metrics and observed species-richness. The weakest
Acknowledgements
RJB is supported by the DFG Emmy Noether Programme (BU 2587/3-1). We thank Graeme Lloyd for discussion, Mike Benton, David Unwin and one anonymous reviewer for their comments that improved the manuscript, and Paleobiology Database members, particularly Matt Carrano and John Alroy, for their data entry. This is Paleobiology Database official publication 166.
References (64)
Successive approximations of diversity curves: ten more years in the library
Geology
(2000)The shifting balance of diversity among major marine animal groups
Science
(2010)- et al.
Effects of sampling standardization on estimates of Phanerozoic marine diversification
Proceedings of the National Academy of Sciences of the United States of America
(2001) - et al.
Pterosaur distribution in time and space: an atlas
Zitteliana B
(2008) - et al.
Dinosaur diversity and the rock record
Proceedings of the Royal Society B
(2009) - et al.
Uncovering the diversification history of marine tetrapods: ecology influences the effect of geological sampling biases
- et al.
Multivariate models are essential for understanding vertebrate diversification in deep time
Biology Letters
(2012) - et al.
Mesozoic marine tetrapod diversity: mass extinctions and temporal heterogeneity in geological megabiases affecting vertebrates
Proceedings of the Royal Society B
(2010) - et al.
High diversity, low disparity and small body size in plesiosaurs (Reptilia, Sauropterygia) from the Triassic–Jurassic boundary
PLoS One
(2012) - et al.
Assessing the quality of the fossil record: insights from vertebrates