Abstract A correlation exists between the quality of the rock record and the diversity of fossils recorded from that rock record but what drives that correlation, and how consistent that correlation is across different environments, remain to be determined. Palaeontologists wishing to investigate past diversity patterns need to first address issues of geological bias in their data.

Abstract It is now well established that seawater chemistry, as well as influencing non-skeletal marine precipitation (‘calcite’ and ‘aragonite seas’), has affected skeletal mineral secretion in some algal and marine invertebrate groups. Skeletal mineralogy has had a yet more profound consequence on fossil preservation. The realization that the fossil record of marine organisms with an aragonite shell is widely depleted in some shelf settings through early, effectively syn-depositional, dissolution (‘missing molluscs’ effect) has led to a re-evaluation of the composition, diversity, ecological and trophic structure of marine benthic communities. Comparisons of molluscan lagerstätten from ‘calcite’ and ‘aragonite seas’ show a similar pattern of skeletal mineralogical loss, that is, no differences are discernibly linked to changed seawater geochemistry. It is notable that the rare mollusc-rich skeletal lagerstätten faunas in the fossil record include many small individuals. Micromolluscs are quantitatively important among modern shell assemblages, yet small size is a major source of taphonomic and biodiversity loss in the fossil record. In skeletal lagerstätten faunas, micromolluscs contribute variably to mollusc biodiversity but appear particularly significant through at least to Triassic times. They highlight a further ‘missing molluscs’ effect of taphonomic loss through early dissolution.

Abstract Correlations between sedimentary rock and fossil records may involve a combination of rock-record sampling bias and common response to external forcing. Quantifying their relative importance from incomplete and uncertain proxy data is not trivial given the potential complexity of interactions among the underlying processes. This paper shows how a non-parametric method can be used to detect causal interactions directly from incomplete and irregular time series, by quantifying directional information transfer between variables. A numerical experiment illustrates how estimates of the relative strength, scale, and directionality of coupling can correctly distinguish a common-cause variable from a spurious relationship, even in cases where correlations are misleading. With a joint analysis of Phanerozoic rock and fossil records pending, the method is applied to oxygen, carbon, and sulphur stable isotope records from marine carbonates, identifying complex interactions between climatic changes and the cycling of carbon and sulphur over the Phanerozoic.

Abstract Many published cladograms report measures of stratigraphic congruence. Strong congruence between cladistic branching order and the order of first fossil occurrences is taken to support both the accuracy of cladograms and the fidelity of the record. Poor congruence may reflect inaccurate trees, a misleading fossil record, or both. However, it has been demonstrated that most congruence indices are logically or empirically biased by parameters that are not uniformly distributed across taxa or through time. These include tree size and balance, mean ghost range duration (gap size) and the range and distribution of origination dates. This study used 650 published cladograms to investigate the influence of these variables on the Gap Excess Ratio (GER). In a range of multivariate models, factors other than congruence per se explained up to 74.5% of the observed variance in GER amongst trees. Arthropods typically have poorer GER values than other groups, but the residual differences from our models are much less striking. The models also show no clear residual trend in GER through the Phanerozoic. Because the GER is strongly influenced by parameters related to cladogram size, balance and duration, comparisons across trees should be made with caution. Supplementary material: Data legends are available at http://www.geolsoc.org.uk/SUP18484

Abstract Studies of biodiversity through time have primarily focused on the marine realm which is generally considered to have a more complete record than terrestrial environments. Recently, this assumption has been challenged, and it has been argued that the record of life on land is comparable or even more robust than that of the shallow oceans. Moreover, it has been claimed that terrestrial successions preserve an exponential rise in diversity, even when corrected for sampling biases such as changes in continental rock volume through time. We evaluate relations between terrestrial diversity and exposed areas of terrestrial sediments using our compiled data on areas of global continental outcrops and generic diversity from the Paleobiology Database. Terrestrial global generic diversity and terrestrial outcrop area are highly correlated following a linear relation. No significant correlation is observed between habitat area and either outcrop area or biodiversity, suggesting that the observed relation between diversity and outcrop area is not driven by a common cause, such as eustasy. We do find evidence for a small residual increase in diversity through time after removing the effect of outcrop area, but caution that this may be driven by an increased proportion of terrestrial fauna with high preservation potential.

Abstract Assessing the quality of the fossil record is notoriously hard, and many recent attempts have used sampling proxies that can be questioned. For example, counts of geological formations and estimated outcrop areas might not be defensible as reliable sampling proxies: geological formations are units of enormously variable dimensions that depend on rock heterogeneity and fossil content (and so are not independent of the fossil record), and outcrop areas are not always proportional to rock exposure, probably a closer indicator of rock availability. It is shown that in many cases formation counts will always correlate with fossil counts, whatever the degree of sampling. It is not clear, in any case, that these proxies provide a good estimate of what is missing in the gap between the known fossil record and reality; rather they largely explore the gap between known and potential fossil records. Further, using simple, single numerical metrics to correct global-scale raw data, or to model sampling-driven patterns may be premature. There are perhaps four approaches to exploring the incompleteness of the fossil record, (1) regional-scale studies of geological completeness; (2) regional- or clade-scale studies of sampling completeness using comprehensive measures of sampling, such as numbers of localities or specimens or fossil quality; (3) phylogenetic and gap-counting methods; and (4) model-based approaches that compare sampling as one of several explanatory variables with measures of environmental change, singly and in combination. We suggest that palaeontologists, like other scientists, should accept that their data are patchy and incomplete, and use appropriate methods to deal with this issue in each analysis. All that matters is whether the data are adequate for a designated study or not. A single answer to the question of whether the fossil record is driven by macroevolution or megabias is unlikely ever to emerge because of temporal, geographical, and taxonomic variance in the data.

Abstract Quantitative patterns in the sedimentary rock record predict many different macroevolutionary patterns in the fossil record, but the reasons for this predictability remain uncertain. There are two competing, but non-mutually exclusive, hypotheses: (1) similarities reflect a sampling bias imposed by variable and incomplete sampling of fossils, and (2) similarities reflect environmental perturbations that influence both the patterns of sedimentation and macroevolution (i.e., common-cause). Macrostratigraphy, which is based on the quantitative analysis of hiatus-bound rock packages, permits variation in the rock record to be expressed in terms of rock quantity and, more importantly, spatiotemporal continuity. In combination with spatially-explicit fossil occurrence data in the Paleobiology Database, it is now possible to more rigorously test alternative hypotheses for similarities in the rock and fossil records and to start distinguishing between geologically-controlled sampling bias and the common-cause hypothesis. Here we summarize results from measuring the intersection of Macrostrat and the Paleobiology Database. Our results suggest that patterns in the fossil record are not dominated by large-scale stratigraphic biases. Instead, they suggest that linkages between multiple Earth systems are driving both spatiotemporal patterns of sedimentation and macroevolution.

Abstract There is increasing evidence to suggest that drivers of bias in the fossil record have also affected actual biodiversity history, so that controls of artefact and true pattern are confounded. Here we examine the role of spatial structuring of the environment as one component of this common cause hypothesis. Our results are based on sampling standardized analyses of the post-Middle Eocene record of shelf molluscs from New Zealand. We find that spatial structuring of the environment directly influenced the quality of the fossil record. Contrary to our expectations, however, we find no evidence to suggest that spatial structuring of the environment was a strong or direct driver of taxic rates, net diversity, or spatial structuring in mollusc faunas at the scale of analysis. Stage-to-stage variation in sampling standardized diversity over the past 40 Ma exhibits two superficially independent dynamics: (a) changes in net diversity were associated primarily with changes in origination rate; and (b) an unknown common cause related extinction rate to the quality of the fossil record and, indirectly, to spatial structuring of the environment. We suggest that tectonic drivers, manifest as second-order sequence stratigraphic cycles, are likely to have been a key element of this common cause.

Abstract Changes in molluscan diversity across the 3rd order sequence boundary from the Lower to the Middle Miocene of the Paratethys were evaluated in the context of environmental bias. Taken at face value, quantitative data from nearshore and sublittoral shell beds suggest a transition from low-diversity Karpatian (Upper Burdigalian) to highly diverse Badenian (Langhian and Lower Serravallian) assemblages, but environmental affiliation of samples reveals a strong facies shift across the sequence boundary. Ordination methods show that benthic assemblages of the two stages, including 4 biozones and four 3rd order depositional sequences over less than four million years, are developed along the same depth-related environmental gradient. Almost all samples are from highstand systems tracts, but Karpatian faunas are mostly from nearshore settings, and Badenian faunas are strongly dominated by sublittoral assemblages. This study emphasizes the importance of highly resolved stratigraphic and palaeoenvironmental frameworks for deciphering palaeodiversity patterns at regional scales and highlights the effort required to reach the asymptote of the collector's curve. Abundance data facilitate the recognition of ecological changes in regional biota and it is suggested that in second and higher order sequences the facies covered within systems tracts will drive observed diversity patterns.

Abstract The deep-sea planktonic microfossil record (foraminifera, coccolithophores, diatoms, radiolaria and dinoflagellates) provides a unique resource for palaeobiology. Despite some geographical gaps due to poor regional preservation, and intermittant time intervals lost to erosion, most time periods for each Cenozoic planktonic biogeographical province are preserved. Vast numbers of specimens and numerous deep-sea cores provide abundant material and the opportunity to tightly integrate macroevolutionary and palaeoenvironmental data. Current documentation of this record is mixed. Catalogues for foraminifera and coccolithophores offer nearly complete species-level clade histories, but taxonomy for siliceous microfossils is incomplete. Published occurrence data is primarily stratigraphic and covers only a fraction of the total preserved diversity. Age models for some sections are excellent (accuracy c . 100 kya) but for many other sections are still poor. Taxonomic errors, age model errors and reworking displace fossil occurrences in time, complicating palaeobiological analysis. With additional taxonomic work, careful collection of whole fauna/floral assemblage occurrence data, improved age models, and the development of better data filtering and analysis tools to deal with data outliers the deep-sea microfossil record can deliver its promise of providing the most complete, detailed record of macroevolutionary change available to science. Supplementary material: Supplementary Appendix is available at http://www.geolsoc.org.uk/SUP18485

Abstract While many studies show a correlation between observed taxonomic richness and various measures of geological sampling, all have been based on the same record of terrestrial and marine sediments collected from the land. Here we present the first analyses of how rock and fossil records vary in the deep-sea. We have developed a novel database of species occurrences of coccolithophores sampled during major drilling programs of the North Atlantic, including the Mediterranean and Caribbean. Our sampling proxy, the number of deep-sea sites sampled – perhaps the most direct measure of sampling used so far – shows an exponential rise towards the Recent. Over the same period species-richness has grown in an approximately linear fashion, but genus-level richness shows a sharp initial increase followed by a much slower decline. However, correlations between both richness measures and sampling are extremely strong and a model assuming true diversity to be constant accurately predicts much of observed richness. We conclude that the deep-sea fossil record, like its land-based counterpart, bears a rock record bias.

Abstract Using data from two palaeontological databases, MIOMAP and FAUNMAP (now linked as NEOMAP), we explore how late Quaternary species loss compared in large and small mammals by determining palaeospecies-area relationships (PSARs) at 19 temporal intervals ranging from c. 30 million to 500 years ago in 10 different biogeographical provinces in the USA. We found that mammalian diversity of both large and small mammals remained relatively stable from 30 million years ago up until both crashed near the Pleistocene–Holocene transition. The diversity crash had two components: the well-known megafaunal extinction that amounted to c. 21% of the pre-crash species, and collateral biodiversity loss due to biogeographical range reductions. Collateral loss resulted in large mammal diversity regionally falling an additional 6–31% above extinction loss, and small mammal diversity falling 16–51%, even though very few small mammals suffered extinction. These results imply that collateral losses due to biogeographical range adjustments may effectively double the regional diversity loss during an extinction event, substantially magnifying the ecological ramifications of the extinctions themselves. This is of interest in forecasting future ecological impacts of mammal extinctions, given that c. 8% of USA mammal species, and 22% of mammal species worldwide, are now considered ‘Threatened’ by the IUCN.

Abstract Mesozoic terrestrial vertebrates gave rise to sea-going forms independently among the ichthyosaurs, sauropterygians, thalattosaurs, crocodyliforms, turtles, squamates, and other lineages. Many passed through a shallow marine phase before becoming adapted for open ocean life. This allows quantitative testing of factors affecting our view of the diversity of ancient organisms inhabiting different oceanic environments. We implemented tests of correlation using generalized difference transformed data, and multiple regression models. These indicate that shallow marine diversity was driven by changes in the extent of flooded continental area and more weakly influenced by uneven fossil sampling. This is congruent with studies of shallow marine invertebrate diversity and suggests that ‘common cause’ effects are influential in the shallow marine realm. In contrast, our view of open ocean tetrapod diversity is strongly distorted by temporal heterogeneity in fossil record sampling, and has little relationship with continental flooding. Adaptation to open ocean life allowed plesiosaurs, ichthyosaurs and sea turtles to ‘escape’ from periodic extinctions driven by major marine regressions, which affected shallow marine taxa in the Late Triassic and over the Jurassic–Cretaceous boundary. Open ocean taxa declined in advance of the end-Cretaceous extinction. Shallow marine taxa continued diversifying in the terminal stages due to increasing sea-level. Supplementary material: The data series and full analytical results are available at http://www.geolsoc.org.uk/SUP18486

Abstract Dinosaurs provide excellent opportunities to examine the impact of sampling biases on the palaeodiversity of terrestrial organisms. The stratigraphical and geographical ranges of 847 dinosaurian species are analysed for palaeodiversity patterns and compared to several sampling metrics. The observed diversity of dinosaurs, Theropoda, Sauropodomorpha and Ornithischia, are positively correlated with sampling at global and regional scales. Sampling metrics for the same region correlate with each other, suggesting that different metrics often capture the same signal. Regional sampling metrics perform well as explanations for regional diversity patterns, but correlations with global diversity are weaker. Residual diversity estimates indicate that sauropodomorphs diversified during the Late Triassic, but major increases in the diversity of theropods and ornithischians did not occur until the Early Jurassic. Diversity increased during the Jurassic, but many groups underwent extinction during the Late Jurassic or at the Jurassic/Cretaceous boundary. Although a recovery occurred during the Cretaceous, only sauropodomorphs display a long-term upward trend. The Campanian–Maastrichtian diversity ‘peak’ is largely a sampling artefact. There is little evidence for a gradualistic decrease in diversity prior to the end-Cretaceous mass extinction (except for ornithischians), and when such decreases do occur they are small relative to those experienced earlier in dinosaur evolution. Supplementary material: The full data set and details of analyses are available at www.geolsoc.org.uk/SUP18487 The same materials (in the form of an Excel workbook) are also available from the first author on request.

Abstract The past decade has witnessed a major revival in attempts to separate biodiversity signals from biases imposed by sampling and the architecture of the rock record. How large a problem this poses to our understanding of biodiversity patterns remains debatable, and new approaches are being developed to investigate this question. Here palaeobiologists with widely differing approaches and interests explore the problems of extracting reliable information on biodiversity change from an imperfect geological record. Topics covered range from the application of information-theoretic approaches that identify directional causal relationships to an in-depth study of how geological biases could influence our understanding of dinosaur evolution. A wide range of new insights into the links between the land, shallow-marine and deep-sea rock and fossil records are presented, making this volume invaluable to anyone in the Earth or life sciences who wishes to remain abreast of this dynamic and rapidly evolving research area.