index fossil research paper

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Proc Natl Acad Sci U S A
• v.112(16); 2015 Apr 21

From the Cover

The future of the fossil record: paleontology in the 21st century, david jablonski.

a Department of Geophysical Sciences and

Neil H. Shubin

b Department of Organismal Biology and Anatomy, University of Chicago, Chicago, IL, 60637

Author contributions: D.J. and N.H.S. wrote the paper.

In the past two decades, great progress in the biological sciences has come from the incorporation of history in understanding basic mechanisms in fields ranging from ecology and conservation biology to physiology and developmental biology. This conceptual expansion has been promoted in large part by theoretical, methodological, and empirical advances in two seemingly disparate fields. The first field is molecular biology, which opens powerful new windows on phylogenetic relationships, genome structure and function, and developmental mechanisms. The second field is paleontology, which affords a unique, direct, and expanding source of information into the anatomies, ecologies, physiologies, and spatial and temporal dynamics of past life. The fossil record is certainly rich in incident and rife with bizarre players, but an extensive body of research now treats the fossil record as a biological laboratory for rigorously framing and testing hypotheses at the intersection of paleontology with diverse disciplines across the full range of timescales encompassed by the earth and life sciences. This Special Feature collects some of the exciting and important new directions and insights, from the beginnings of life on Earth to the immediate precursor to the present-day biota ( Fig. 1 ).

Distribution through geologic time of the contributions to this Special Feature. Ga, billion (10 9 ) y ago; Ka = thousand (10 3 ) y ago; Ma, million (10 6 ) y ago.

Paleontology informs the natural sciences by providing unique sources of data on important phenotypes and on the spatial and temporal dynamics of biological events and processes. Accordingly, we organize these contributions in terms of the important phenotypes contained in the fossil record, and the analysis of dynamics of species, clades, and communities in both space and time.

At its most fundamental level, the fossil record is a narrative of changes to phenotypes and their functions: the origin, persistence, and demise of biological form ( 1 – 4 ), along with changes in behavior, physiology, and life history of vertebrates, invertebrates, plants, fungi, and protists ( 5 – 14 ). The contributions of paleontology to the study of the rates and pattern of phenotypic evolution are legion. At the species level, the fossil record has famously shown that the evolutionary responsiveness of local populations on decadal or centennial timescales usually translates at the 1- to 10-million-y timescale into stasis or nondirectional random walks rather than sustained, directional evolutionary transformation ( 15 , 16 ). For higher taxa, the quantification of form within a multidimensional morphospace was developed in paleontology ( 17 ) and this rich literature continues to find new ways to explore evolutionary diversification, from formal visualizations of evolutionary convergence and parallelism (e.g., ref. 18 ) to direct analyses of the relation between ontogeny and phylogeny ( 19 – 21 ).

The biological world we see around us today is a highly pruned version of a rich and ancient tree of life. Consequently, reliance solely on recent taxa in analyses of origins can definitively mislead. The diversity and disparity of many extant clades, from elephants ( 22 ) to horses ( 23 ), and for that matter hominins ( 24 ), has demonstrably declined in the latest Cenozoic. Those clades were hardly unique, so that fossil data are essential for a fuller understanding of the rates and patterns of phenotypic change within and among many clades (e.g., refs. 25 and 26 ). Of course, some clades are inaccessible or sparsely represented as fossils, but even in such groups, from onychophorans to ants to penguins, fossils alter our picture of the timing and extent of morphological diversification. Formal incorporation of sparse fossil morphologies in phylogeny-based analyses of extant species is a promising interdisciplinary growth area (e.g., ref. 27 ).

Extinction can generate a false signal regarding the origin of evolutionary novelties when only extant taxa are analyzed ( 28 , 29 ). For example, phylogenetic analysis of extant chordates suggests that bone evolved following the split between sharks and bony fishes, but the fossil record shows that bone was present well below that evolutionary node, and that modern sharks represent a derived state for the clade, having lost bony structures widely distributed in their ancestors ( 30 , 31 ). Paleontological data are invaluable for inferring ancestral character states and the assembly of character complexes, and can now be used to test hypotheses drawn from developmental or phylogenetic analyses.

The discovery and analysis of fossils from key intervals in the history of life can inform the sequence, pattern, and phylogenetic dynamics underlying the origin of major functional and anatomical novelties. Such “intermediate forms” in the fossil record can serve as tests of genetic, developmental, and biomechanical hypotheses based on extant taxa. Indeed, some of the most fundamental discoveries of fossil stem taxa have only happened in recent decades. Among vertebrates alone, fossils have illuminated evolutionary pathways leading to the origin of vertebrates ( 32 ), tetrapods ( 33 ), turtles ( 34 ), snakes ( 35 ), mammals ( 36 , 37 ), birds ( 38 ), horses ( 23 ), whales ( 39 , 40 ), hominids ( 24 ), and many other groups. These fossils are valuable on several counts: ( i ) they provide data on the rate and pattern of character acquisition, which in turn illuminates the factors underlying dramatic transformations (e.g., feathers as insulation and display before flight in birds); ( ii ) they inform mechanistic hypotheses on underlying changes in development, both constraining and inspiring experimental tests of rival scenarios (e.g., developmental hypotheses on the origin of the wrist and ankle joints of tetrapods); and ( iii ) they place transitions in their environmental context (e.g., whales on the shoreline of a tropical sea ( 39 ) and hominids arising with bipedal gaits in woodland settings ( 41 ). Moreover, many major transitions occurred in ecosystems lacking close modern analogs, and so can be understood ecologically only by paleontological analysis of the peculiar ecosystems of their times. Examples include the origin of vertebrates, arthropods, mollusks, and echinoderms (among others) in late Proterozoic and Cambrian seas with the initial establishment of macrofaunal foodwebs ( 42 ), the Paleozoic-early Mesozoic diversification of major insect clades in a world before flowering plants ( 43 ), and the protracted evolution of oceanic ecosystems lacking the mineralized phytoplankton that only became major factors in the late Mesozoic and Cenozoic ( 44 ).

Rates and timing of taxonomic diversification within and among lineages can sometimes be inferred using data on extant organisms alone, and the vast comparative biology literature has detected many linkages between intrinsic biological traits and diversification rates in phylogenies of extant taxa (reviewed in refs. 45 and 46 ). However, decomposing diversification rate into its origination and extinction components is crucial for a wide range of issues, from niche conservatism to diversity-dependence ( 47 – 49 ). Such data are difficult to retrieve robustly from extant species because, for example, correlations between rates and individual or clade-level traits may undermine parameter estimates from phylogenies ( 50 , 51 ), and extinction can mask true evolutionary rates or trends relative to those inferred from extant species (e.g., refs. 29 and 52 – 55 ). The fossil record can get closer to the mechanistic underpinnings of present-day biodiversity by providing direct observations on how and when clades differ in origination and extinction rates, and how both variables, and their relation to intrinsic traits, can change over time (e.g., refs. 56 – 59 ). This paleontological access to long-term evolutionary dynamics allows empirical evaluation of multilevel evolutionary processes, and the relative contribution of traits at organismic, species, and even clade level in the waxing and waning of clades through time ( 2 , 7 , 60 , 61 ). A major area of ongoing work on this general topic involves the integration of paleontological and molecular phylogenetic data, particularly when one or the other is sparse, and many new approaches and results are coming online (e.g., refs. 27 and 62 – 66 ).

Regarding extinction, the fossil record reveals that complex, seemingly robust ecological systems can collapse and take millions of years to recover. It also shows that when major extinction events occur on a global scale, recoveries are far more heterogeneous than expected: some surviving clades rediversify prolifically, whereas others linger at low diversities or slip into extinction long after the initial bottleneck ( 67 , 68 ). Abrupt biotic transitions also occur at regional scales, and far more frequently than the attention focused on the Big Five mass extinctions of the geologic past might suggest. The fossil record associated with, for example, the late Cenozoic aridification of the North American interior ( 69 ), the uplift of the Himalayas ( 70 ), and the uplift of Panama ( 71 ) provides opportunities for comparative analyses of the biotic response to changing temperature, moisture, and productivity regimes, quantifying timescales and among-clade dynamics of diversity loss and recovery, and testing alternative drivers. The fossil record is as much a document of clade failure as success, and provides a direct observational window on a wide range of natural experiments, in deep time and directly preceding the present biological moment.

At the same time, another key finding from the fossil record is a lack of extinction where it might be expected. For example, one of the great contributions of fossil data is the demonstration that species tend to respond to climate changes by individualistic range shifts, despite the web of positive and negative interactions seen in every present-day site. Thus, biotic associations have been disbanded and assembled during repeated glacial cycles without significant extinction or evident disruption of energy flow through novel (“non-analog”) systems ( 72 ). The addition of anthropogenic pressures—not the least migration barriers in the form of highways, cultivated land, and cities and suburbs—has the potential to overturn such resiliency, and paleontological data on what biotas “want to do” in response to climatic and other drivers provide a valuable baseline perspective from which to view the likely acceleration of future species movements. Larger-scale biotic interchanges, which were often asymmetric between donor and recipient regions in the geologic past as they are today ( 73 ), further allow comparative evaluation of the major determinants and long-term consequences of successful range shifts of different magnitudes. At least one marine interchange through an ice-free Arctic Ocean has already occurred ( 74 ), and further analyses of that event in advance of its likely re-enactment in the near future ( 75 ) would be valuable.

The temporal perspective afforded by the fossil record has additional applications for predicting change and managing biodiversity. For example, biological baselines used to set conservation or restoration targets generally rely on the earliest sustained record-keeping in a target region, but this approach is severely limited relative to the trajectory of anthropogenic change at most locations. The nascent field of conservation paleobiology is demonstrating more profound shifts in natural systems from prior removal of key species and disruption of habitats and biogeochemical cycles by agriculture, fisheries, and industrialization ( 76 , 77 ). In addition to community-level perspectives, close study of the environmental and biogeographic history of individual species is often feasible. For example, the European bison ( Bison bonasus ) has been managed as a forest specialist, but the fossil record suggests that its woodland distribution was only recently created by the loss of its open grassland habitat ( 78 ). As in all of the topics touched upon here, the exchange of methods, data, and ideas should be a two-way street, and conservation paleobiology will benefit by the infusion of molecular methods and the continuing development of ecological and evolutionary models ( 48 , 79 ).

Clades are dynamic not only over time but through space. Although spatial variation in sampling can be as large a challenge in the fossil record as it is for many components of the modern biota ( 80 ), the fossil record of even sparsely preserved clades can significantly inform the biogeographic and environmental history of lineages and major groups. The geologically recent presence of horses, proboscidians, and rhinos in the New World ( 81 ) and, in the early Cenozoic, monotremes in Argentina ( 82 ), ratites in the Northern Hemisphere ( 83 ), hummingbirds in the Old World ( 83 ), mousebirds in North America ( 84 ), Acropora reef corals in Britain ( 85 ), and the huge (up to 1 m) campaniloid snails of southwest Australia in California, the Caribbean, Africa, and northwest Europe ( 86 ), are just a sampling that demonstrates the pervasive history of biotic expansion and regional extinction that would be difficult or impossible to infer from present-day diversity and distributions. The paleontological data can now feed into new models for the spatial and temporal dynamics in their respective clades, and the history of biodiversity hotspots and coldspots can become much clearer ( 87 – 90 ). These data also suggest that macroecological and macroevolutionary analyses of present-day regional biotas may be better informed by incorporating geologically recent extinction and origination into the equation rather than under an equilibrium assumption a priori ( 48 , 91 – 93 ).

Just as the fossil record has provided rich evidence for the highly clumped distribution of originations and extinctions through geologic time (e.g., refs. 42 and 47 ), spatially explicit paleontological data have shown that evolutionary novelties and major clades preferentially originate at low latitudes ( 94 , 95 ), and perhaps certain environments in both marine and terrestrial settings ( 95 – 97 ). Large-scale diversity trends along latitudinal and other environmental gradients are thus shaped not just by in situ origination and extinction but also by the export of lineages into novel habitats, in violation of the general tendency toward evolutionary niche conservatism seen in most clades. This is an especially active area for interdisciplinary work: spatially explicit phylogenetic analyses incorporating fossil and extant taxa ( 98 ) will go far in elucidating the mechanism and significance of these patterns.

Analyses that combine phenotypes with temporal dynamics in a spatially explicit context are rare but are likely to increase with new methods and datasets. In a pathfinding study, Hellberg et al. ( 99 ) showed that present-day populations of a marine snail diverged significantly in form from its late Pleistocene predecessors, generating a variety of novel shell forms with postglacial range expansion. This exemplary study combined phylogeographic analysis with comparative morphometric analyses within a single species before and after range expansion. Only by coupling analyses of phenotypic diversification in both space and time will we be in a position to achieve a mechanistic understanding of ecological and evolutionary changes across scales, from local to global, and hierarchical levels, from molecules through bodies to species and clades ( 2 ).

Introduction to the Papers in This Special Feature

The papers assembled in this Special Feature, which range in time from the earliest vestiges of life in Archean sediments to the fossil record accumulating from extant species communities ( Fig. 1 ), can be roughly organized as mainly concerned with origins or dynamics.

Brasier et al. ( 100 ) discuss the technical and conceptual advances in the search for the earliest fossil traces of life. Microfossils have long been reported from the Archean Eon [3.5–2.5 Ga (= billion y)], but new techniques for imaging and nanoscale elemental analysis, growing knowledge of present-day prokaryote phylogeny and diversity, and close attention to the environments and processes of microorganism preservation are pushing this field forward. Brasier et al. discuss refined criteria for separating preservational artifacts from fossilized prokaryotes, and revisit three classic microbiotas. The authors reject the biogenicity of many specimens from the Apex Chert (a controversial assemblage dated at 3.46 Ga), and document a microbiota from the roughly contemporary Strelly Pool Sandstone (dated 3.43 Ga), unexpectedly preserved between sand grains from the earliest known shoreline. These shore deposits seem unlikely hosts for cellular preservation, and so these finds open a new search window for the early fossil record and expand our ecological picture of early microbial life. High-resolution imaging in one of the classic younger microbiotas, the Gunflint Chert (1.88 Ga), confirms that the problematic microfossil Eosphaera represents a form unknown in modern biotas, likely an extinct experiment in microbial multicellularity.

Droser and Gehling ( 101 ) synthesize data on the Precambrian assemblage of multicellular fossils known as the Ediacarian Biota. With a near-global distribution from >40 localities in rocks 575–541 My, these fossils offer a picture of early multicellular life about 30 My before the famous Cambrian Explosion. An enigmatic array of forms, the Ediacarian Biota have spawned numerous controversial hypotheses about their relationship to extant clades, and their significance for understanding the rise of multicellular life in general. Droser and Gehling show that the Ediacaran forms, whatever their phylogenetic affinities, present many attributes of later animal life, including mobility, heterotrophy, skeletonization, and participation in complex ecosystems, and so these forms offer an informative comparator to later animal evolution. Moreover, because the biota includes, in addition to many phylogenetically problematic taxa, early relatives of cnidarians, poriferans, and bilaterians, it sheds light on the roots of the metazoan radiations of the Cambrian and later times.

Pieretti et al. ( 102 ) demonstrate the power of combining the growing morphological and phylogenetic records from the fossil record with novel technologies in development and genomics to provide a richer understanding of major evolutionary transitions. With continued discovery of new fossils and fresh perspectives on others, and the expansion of developmental and genomic analyses beyond standard model organisms, this multidisciplinary approach can be applied to many of the long-standing questions of anatomical evolution. Focusing on the evolution of vertebrate appendages, Pieretti et al. first present evidence that the origin of paired, lateral appendages in vertebrates involved a redeployment of the developmental program for a single median fin, initially generating a paired set of pectoral fins and only later giving rise to the pelvic appendages. The transition from fins to limbs is well-documented in the fossil record, and we can see that the defining feature of limbs, the wrist, and digits, arose in a series of evolutionary steps in an extinct lineage of sarcopterygian fishes. The wrist and digits have no clear morphological counterparts in extant fishes, and the underlying developmental and genetic transitions are obscured in model species of teleost fishes because of genome duplication and reorganization in that clade. Experiments using gar (which separated from the teleosts before the duplication event) show that its regulatory enhancers can drive digit development in mice, suggesting that the wrist–digit complex and its regulatory pathways did indeed have homologs in the distal bones of early bony fishes. Changes in gene regulation can also be traced through diversification of limb design, and analysis of mammalian digit reduction, a common evolutionary theme, shows that convergence on reduced digit number has occurred by two very different genetic and developmental mechanisms, giving insight into the remarkable malleability of vertebrate limbs. The next step in the integration of paleontology and developmental biology may well be the experimental modulation of gene expression to test specific hypotheses on transitional forms and their underlying genetic basis against fossil phenotypes close to those transitions.

White et al. ( 41 ) show that a priori assumptions about what evolutionary ancestors should look like can be misleading in our understanding of human evolution. Biologists have long assumed that the immediate ancestors of the human clade were chimpanzee-like. The discovery of Ardipithecus challenges these assumptions on multiple levels. By revealing both unexpected morphological conditions and ecological provenance, this discovery reveals an informative complexity to hypotheses of morphological transformation and adaptation during the acquisition of traits unique to humans.

The empirical analysis of evolutionary tempo and mode has been one of the great contributions of paleontology. Phenotypic stasis has proven to be far more prevalent than expected, and the challenge is now to rigorously test alternative evolutionary models for long-term phenotypic evolution and seek explanatory mechanisms for the different trajectories exhibited in fossil timeseries. Hunt et al. ( 103 ) fit likelihood models to a large compilation of evolutionary studies ranging in scope from 5,000 y to >50 My. The authors again find that stasis and random walks best account for temporal patterns, rather than directional change, but also show that more complex models—particularly stasis plus punctuational change, and shifts from a random walk to stasis—tend to be increasingly supported in a maximum-likelihood framework with increasing numbers of samples within a study. Stasis is more prevalent in marine than in terrestrial settings, and in macroinvertebrates and vertebrates than in microfossils. Evaluating the role of external drivers, Hunt et al. simulate phenotypic sequences using a model in which traits track an empirical long-term climate curve. The frequencies of evolutionary modes in the simulated sequences are quite similar to those in the real data, suggesting that the bounded, often oscillatory, nature of many physical environmental changes contributes significantly to the nondirectional dynamic observed in many fossil sequences.

Goswami et al. ( 104 ) explore the ways that morphological characters are integrated developmentally into a unified, functioning organism, and how this integration influences and is influenced by evolution. The authors argue for the importance of a deep-time perspective to the analysis of integration, modularity, and the origins of morphological variation. Taking advantage of the large mammal populations preserved in Late Pleistocene tar pits, Goswami et al. analyze the evolutionary signal of patterns of phenotypic integration in dire wolves and saber-toothed cats from different pits spanning 27,000 y of evolution at Rancho La Brea. These samples reveal decreasing levels of phenotypic integration as climate changed, indicating that developmental interactions created—or failed to damp—increasing phenotypic variation in responses to stresses associated with environmental change.

A long-standing issue in biodiversity dynamics is the potential operation of negative feedbacks: Is the rate and level of taxonomic or phenotypic diversification itself diversity-dependent? This question is difficult to resolve definitively using comparative phylogenetic data on extant taxa because extinct species cannot enter into diversity or disparity estimates through time. Slater ( 105 ) tests for diversity-dependent morphological diversification in the dog family, Canidae, a group with an excellent fossil record that exhibits three sequential diversifications through their 40-My history in North America. Diversification slowdowns are generally viewed in terms of declining ecological opportunities as the “ecological barrel” gets filled, and fossil canids can be studied in terms of dietary diversity based on dentition and body size. Slater finds support for neither an early burst of evolution and subsequent diversity-dependent slowdown, nor for unconstrained diversification fitting a Brownian motion model. Instead, the North American canids best fit a multipeak Ornstein–Uhlenbeck model, where clades gravitate toward specific trait values, in this case three different body-mass and tooth-area values, one for each of the major dietary categories within the Canidae. These results can thus be interpreted as a set of replicated diversifications, occurring in and among three stable, bounded adaptive zones. Thus, apparent early bursts in phenotypic evolution could derive not only from an evolutionary slowdown in an increasingly crowded world, but from constant evolutionary rate within a bounded morphospace. Those bounds might be set by intrinsic, developmental constraints, the presence of phylogenetically unrelated but ecologically similar competitors, or by climatic and other environmental changes that continually shift the ecological space that can accommodate different dietary groups over time. Diversity ceilings may exist but are moving targets on macroevolutionary timescales, as others have suggested in various contexts (e.g., refs. 47 , 106 , and 107 ).

Regional biotas are shaped by the interaction of origination, extinction, and immigration, so that present-day diversity and its relation to current environmental factors can only tell part of the story. S. Huang et al. ( 108 ) use the rich Pliocene record (2–5 My ago) on the warm, temperate coasts of North America to evaluate the dynamics leading up to the modern distribution of marine biodiversity. With marine bivalves as a model system, the authors show that overall regional diversities were shaped by extinction and subsequent recovery through origination and immigration, with clades (here, taxonomic families) shared by the two coastlines showing a variety of divergent, convergent, and parallel trajectories in species richness. Thus, similar diversities for a given clade in different regions today is no guarantee of a shared history or similar diversification rates, and differences might be geologically recent rather than stemming from deep-seated evolutionary differences. S. Huang et al. find that the contribution of regional extinction to today’s species-richness patterns can be predicted by the geographic range sizes of species during the Pliocene, a variable difficult to extract from present-day data but likely to play a role in extinction risk with the onset of large-amplitude glacial cycles near the Plio-Pleistocene transition. Past distributional shifts (i.e., local or regional extinction and immigration) are crucial to interpreting present-day biogeography.

Extinction does more than remove phenotypes and their ecological roles: it erases evolutionary history (EH). This effect is of interest in its own right; for example, the marine extinctions that accompanied mid-Cenozoic polar cooling were more evenly distributed phylogenetically in the Arctic than in the Antarctic, so that similar regional extinction intensities removed significantly more EH in the Southern Ocean than in the Arctic ( 89 ). However, EH can also be a basis for conservation decisions, with regions or clades prioritized in part by the amount of EH they represent. D. Huang et al. ( 109 ) address for the first time one of the major conceptual and methodological gaps between molecular and phylogenetic approaches to origination and extinction, and assess how these contrasting approaches affect estimates of EH and its loss. Molecular phylogenies, which depict the splits created by differentiation of gene pools in extant species, necessarily yield bifurcations. Paleontological phylogenies, which reflect the phenotypic stasis and nondirectional change that pervades the fossil record, often incorporate a budding evolutionary topology, with lineages persisting after new species have split from ancestral populations or clades. Using simulations, D. Huang et al. find that estimated losses of EH in major extinction events are qualitatively similar when extinctions are random with respect to clade age, although EH is lost more rapidly with increasing extinction intensity in paleontological data than in molecular trees. When extinction focuses on older lineages, as appears to be happening today, both approaches capture the disproportionate erosion of EH, although the molecular signal is damped relative to the paleontological one. Using a phylogeny of living and extinct scallops from California, which have an exceptional fossil record, the authors find a preferential loss of young species in the Plio-Pleistocene extinction pulse, although this signal proves difficult to retrieve from molecular data, consistent with simulation results. Nonetheless, the encouraging outcome of this study is that extinction selectivity measured by these different approaches to EH are broadly comparable, indicating that the fossil record can provide a useful natural laboratory for anticipating future losses of EH caused by anthropogenic extinctions.

At a finer scale, ecological communities are also shaped by origination, local extinction, and immigration. Jackson and Blois ( 110 ) consider how these factors shape terrestrial community composition and how an understanding of the incessant change in communities documented in the Quaternary fossil record—the past 2.6 My—can inform both ecological theory and the management and conservation of terrestrial biodiversity. Climate changes can be gradual and directional, but are often punctuated by episodic events and by rapid-state transitions, and detailed pollen records show that some communities accordingly disassemble gradually via decline and replacement of dominant species, whereas others undergo rapid collapse and turnover. Similar data for insects and vertebrates reinforce this perspective, that local biological communities are “passing manifestations of ecological and biogeographic processes in a world of ceaseless environmental change” ( 110 ). Rates of change have varied greatly over time and among communities through time the past 20,000 y, and the implications of these strong variations are only beginning to be explored and appreciated. While highlighting the abundant paleoecological support for environmental drivers for community composition, Jackson and Blois emphasize that even the broadest spatial patterns are spatial aggregations of local interactions among populations of competitors, predators, mutualists, and parasites. Such interactions govern community outcomes of environmental change, and more work is needed that takes into account interactions—and the traits that mediate them—in relation to environmental drivers. Still unclear, for example, is whether communities can display stability in the functional attributes of their components even as species composition ebbs and flows. Integrated study of paleoecological case studies would test fundamental ecological theory and provide insights for how and where anthropogenically driven activities might be overriding past controls on community dynamics.

In our final paper, Kidwell ( 111 ) examines the many ways in which very young fossil records from the last few decades to millennia can provide unique insights into the current status of extant species, communities, and biomes, emphasizing marine and coastal systems. The fossil record of the Anthropocene—the informal term encompassing the human domination of a majority of natural processes on a global scale—includes information from still-unburied organic remains as well as shallow-cores, thus linking directly to older fossil records. Such approaches provide baseline ecological data otherwise inaccessible by direct observations and written records, and permit the roles of natural and anthropogenic drivers to be assessed in situations both outside and within the timeframe of diverse human impacts. Kidwell notes that virtually all biological systems, even in the most remote areas, today operate around anthropogenically driven environmental trends rather than stationary means, and argues that the interaction of climatic with other human stressors, rather than climate change alone, is creating the apparently unique dynamics in today’s biota. Analyses of shifts in abundance, diversity, and ecosystem function in fossil records document unprecedented biological change within the last few centuries relative to the past 2 My, linked to human expansion and technological advances. Kidwell calls for an integration of paleontological data on the dynamics of extant species, communities, and ecosystems on recent decadal to millennial timescales, which had once been viewed as outside the paleontologist’s purview, with ecological theory and management of the biological world.

The Future of the Fossil Record

The future of the fossil record is as many-faceted as the disciplines that it impacts. A wide range of research questions involving biological systems can benefit greatly by incorporating data from the fossil record. Many exciting directions could not be included in this Special Feature, from ancient DNA ( 112 , 113 ) to biomechanics ( 114 – 116 ), to positive and negative feedbacks among clades ( 117 – 119 ) and between biological systems and the global environment ( 5 , 42 , 120 ). By bringing together papers at the forefront of the integration of paleontology with other disciplines, we hope this Special Feature will help define a future agenda for work at these scientific interfaces.

Sadly, as this volume went to press we learned of the tragic death of one of our contributors, Martin Brasier. Martin was an integrative paleobiologist par excellence , who made fundamental contributions to many aspects of paleontology, most recently regarding the early evolution of life, from the interpretation of the oldest fossil evidence of living things to the biology of the enigmatic Ediacaran Biota. He will be missed.

Acknowledgments

We thank D. Bapst, S. Huang, S. Kidwell, and G. Slater for valuable reviews of this introduction; the authors and reviewers of the contributions to this Special Feature; and John Westlund for assistance with Fig. 1. Support was provided by the Brinson Foundation (N.H.S.) and the National Aeronautics and Space Administration (D.J.).

The authors declare no conflict of interest.

Loading metrics

Open Access

Peer-reviewed

Research Article

Tooth morphology elucidates shark evolution across the end-Cretaceous mass extinction

Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Validation, Visualization, Writing – original draft, Writing – review & editing

* E-mail: [email protected] , [email protected] (MB); [email protected] (BPK)

Affiliation Subdepartment of Evolution and Development, Department of Organismal Biology, Uppsala University, Uppsala, Sweden

Roles Conceptualization, Funding acquisition, Methodology, Supervision, Writing – review & editing

Affiliation Palaeoscience Research Centre, School of Environmental and Rural Science, University of New England, Armidale, New South Wales, Australia

Roles Conceptualization, Funding acquisition, Supervision, Writing – review & editing

Roles Conceptualization, Supervision

Roles Conceptualization, Funding acquisition, Supervision, Validation, Writing – original draft, Writing – review & editing

Affiliation Museum of Evolution, Uppsala University, Uppsala, Sweden

• Mohamad Bazzi, 
• Nicolás E. Campione, 
• Per E. Ahlberg, 
• Henning Blom, 
• Benjamin P. Kear

• Published: August 10, 2021
• https://doi.org/10.1371/journal.pbio.3001108
• Reader Comments

Sharks (Selachimorpha) are iconic marine predators that have survived multiple mass extinctions over geologic time. Their prolific fossil record is represented mainly by isolated shed teeth, which provide the basis for reconstructing deep time diversity changes affecting different selachimorph clades. By contrast, corresponding shifts in shark ecology, as measured through morphological disparity, have received comparatively limited analytical attention. Here, we use a geometric morphometric approach to comprehensively examine tooth morphologies in multiple shark lineages traversing the catastrophic end-Cretaceous mass extinction—this event terminated the Mesozoic Era 66 million years ago. Our results show that selachimorphs maintained virtually static levels of dental disparity in most of their constituent clades across the Cretaceous–Paleogene interval. Nevertheless, selective extinctions did impact apex predator species characterized by triangular blade-like teeth. This is particularly evident among lamniforms, which included the dominant Cretaceous anacoracids. Conversely, other groups, such as carcharhiniforms and orectolobiforms, experienced disparity modifications, while heterodontiforms, hexanchiforms, squaliforms, squatiniforms, and †synechodontiforms were not overtly affected. Finally, while some lamniform lineages disappeared, others underwent postextinction disparity increases, especially odontaspidids, which are typified by narrow-cusped teeth adapted for feeding on fishes. Notably, this increase coincides with the early Paleogene radiation of teleosts as a possible prey source, and the geographic relocation of disparity sampling “hotspots,” perhaps indicating a regionally disjunct extinction recovery. Ultimately, our study reveals a complex morphological response to the end-Cretaceous mass extinction and highlights an event that influenced the evolution of modern sharks.

Citation: Bazzi M, Campione NE, Ahlberg PE, Blom H, Kear BP (2021) Tooth morphology elucidates shark evolution across the end-Cretaceous mass extinction. PLoS Biol 19(8): e3001108. https://doi.org/10.1371/journal.pbio.3001108

Academic Editor: Tiago Bosisio Quental, Universidade de São Paulo, BRAZIL

Received: January 7, 2021; Accepted: July 5, 2021; Published: August 10, 2021

Copyright: © 2021 Bazzi et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All data and R scripts are available from Data Dryad under DOI: https://doi.org/10.5061/dryad.c866t1g5n .

Funding: This work was supported by the Royal Swedish Academy of Sciences (GS2017-0018) to M.B., and a Wallenberg Scholarship from the Knut and Alice Wallenberg Foundation to P.E.A. B.P.K. also acknowledges funding from a Swedish Research Council Project Grant (2020-3423), and N.E.C. is funded by an Australian Research Council Discovery Early Career Research Grant (DE190101423). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Abbreviations: EAR, Eastern Atlantic rim; EEB, European epicontinental basin; FDR, false discovery rate; GPA, generalized Procrustes analysis; IQR, interquartile range; K/Pg, Cretaceous/Paleogene; Ma, million years; PCA, principal components analysis; PV, Procrustes variance; RRPP, residual randomization permutation procedure; TPS, thin plate spline; WAR, Western Atlantic rim; WIB, Western Interior basin; 2B-PLS, 2-block partial least-squares

Introduction

Fossils provide the only direct evidence of interplay between organisms and their environments over vast evolutionary timescales [ 1 – 3 ]. They are, therefore, crucial for exploring the drivers of past biodiversity change and can shed light on the origins of modern ecosystems [ 3 ]. However, the analytical challenge is to discern a genuine biological signal from the combined obfuscations of geologic, taphonomic, sampling, taxonomic, analytical, and interpretive biases [ 4 – 7 ]. While these may be impossible to overcome in entirety, the fossil records of some widely distributed and chronostratigraphically extended clades provide exceptional opportunities to characterize macroevolutionary processes through deep time.

Sharks constitute one such group because their dental remains are abundant in Mesozoic and Cenozoic marine deposits—a timeframe covering approximately 250 million years (Ma) [ 8 , 9 ]. Extant shark species are also ecologically disparate, encompassing a spectrum of macrophagous to microphagous predators that account for nearly half (42%) of all the currently documented chondrichthyan biodiversity ( N = 1,193 species) [ 10 , 11 ]. Nevertheless, the various biological and environmental factors that have shaped shark evolution remain obscure. In particular, their capacity to survive mass extinctions is relevant for understanding the dramatic decline of shark populations observed in our modern oceans [ 11 – 16 ].

The end-Cretaceous mass extinction (approximately 66 Ma), which marks the Cretaceous/Paleogene (K/Pg) chronostratigraphic boundary, is especially pertinent because it profoundly disrupted marine ecosystems but has disputed implications for shark species diversity and morphological disparity. Indeed, contrasting interpretations advocate either limited [ 17 ] or complex interrelationships of biotic and abiotic drivers that seemingly influenced shark evolution from before, during, and after the K/Pg mass extinction event [ 18 – 21 ]. Here, we explore these contentions via a comprehensive assessment of shark dental morphology across the end-Cretaceous mass extinction. Our study expands on previous studies that targeted either geographically localized [ 22 , 23 ] or clade-specific [ 21 ] assemblages. We use a dataset of 1,239 fossil shark teeth, representing 9 major selachimorph clades sampled at global and regional scales. These groups include the following: the Galeomorphii orders Carcharhiniformes, Heterodontiformes, Lamniformes, Orectolobiformes; the Squalomorphii orders Echinorhiniformes, Hexanchiformes, Squaliformes, Squatiniformes; and the extinct [†] Synechodontiformes. Our approach uses geometric morphometrics to compare disparity and morphospace patterns across a constrained 27.6-million-year interval spanning the Campanian and Maastrichtian ages of the Late Cretaceous (83.6 to 66 Ma) to the Danian, Selandian, and Thanetian ages (= Paleocene epoch) of the early Paleogene (66 to 56 Ma). We test the following hypotheses that: (1) selachimorph disparity was in decline before marine ecosystem disruption during the end-Cretaceous mass extinction, resulting from major marine transgressions during the Maastrichtian [ 23 , 24 ]; (2) selachimorph taxonomic richness depletion at the K/Pg boundary [ 18 , 19 ] was coupled with eco-morphological turnover; (3) apex predator shark lineages were disproportionately impacted, consistent with similar losses in marine tetrapods and osteichthyans [ 18 , 21 , 25 ]; and (4) extinction and recovery patterns were consistent at global and regional scales [ 21 – 23 ].

Materials and methods

Dataset assembly.

Our dataset was compiled using photographs and graphic images derived from first-hand observations or the literature ( S1 Data , Fig A and Tables A and B in S1 Text ). Following recommended best practices [ 21 , 26 , 27 ], we screened the raw image data to include only those depicting complete tooth crowns with adequate resolution to determine the crown–root junction. The global sampling includes nearly all major selachimorph orders (Fig A in S1 Text , Fig B in S1 Text , panel A), except for Pristiophoriformes (Sawsharks), which are sparsely documented [ 8 , 28 ]. We also elevated Echinorhinidae to Echinorhiniformes [ 29 – 31 ] and employed a sensitivity analysis to test the morphospace occupation and disparity effects of †Synechodontiformes, which has been classified as either a clade within Galeomorphii or a neoselachian sister lineage (Fig B in S1 Text , panel A) [ 32 – 35 ]. Finally, most images were captured in the labial aspect, unless only a lingual view was available, and with the tooth apex directed to the left (Fig C in S1 Text ). The equivalency of labial and lingual views [ 21 ] was tested using ordinary least-squares linear models based on a subsample of the labial and lingual view data ( N = 866).

Time-binning encompassed 5 geochronological ages spanning the immediate K/Pg interval: Campanian and Maastrichtian/Danian, Selandian, and Thanetian. However, we also implemented an alternative 4-age time-binning scheme that pooled temporally ambiguous specimens assigned to the Danian and Selandian (Tables C and D in S1 Text ). In addition, we carried out analyses using a subsample of the dataset ( N = 659) for which “early” and “late” sub-ages could be defined for the Campanian and Maastrichtian and “early,” “middle,” and “late” for the Danian (Table E in S1 Text ). Numerical age values, in Ma, were taken from the International Chronostratigraphic Chart v2020/03 [ 36 ] and used as references for plotting.

Acquisition of geometric shape data

Landmark-based geometric morphometrics quantifies biological shapes as a series of evolutionary homologous points in Cartesian space [ 26 , 27 , 37 – 42 ]. However, evolutionary homology [ 37 , 38 ] between landmarks cannot be assumed because of the inherent morphological variability in shark teeth (e.g., the location and number of cusplets). As a result, we designated our landmark placements based on topological rather than evolutionary homology [ 41 ].

Landmark digitization was carried out in tpsDig2 v. 2.31 [ 43 ] with resampling to a standard number of equidistant semilandmarks using customized code in R v. 4.0.5 [ 44 ]. The resulting scheme comprised 2 open curves defined by semilandmarks but anchored by 3 fixed landmarks: Type 1 landmarks delimited the mesial and distal crown–root junctions, and a Type 2 landmark pinpointed the tooth apex (sensu [ 37 ]; Fig B in S1 Text , panel B, Table F in S1 Text ). The number of semilandmarks (k) was determined using resampling of the mesial and distal curves at equal spacings of k = 40, 60, 80, 100, 120, 140, and 160. Qualitative observations (Figs D and E in S1 Text ) found that k = 160 best-captured tooth shape complexity and included distal and mesial curves of 78 and 79 sliding semilandmarks, respectively. Tooth serrations were not digitized because of limited image resolution, but we acknowledge that these structures are functionally important [ 45 ].

Lastly, to screen for possible digitization errors, we extracted a random subsample of 30 tooth images (Table G in S1 Text ) and used a one-way ANOVA to calculate the intraclass correlation coefficient ( R ) [ 26 , 46 , 47 ] (also see S1 Text ). We also ran a 2-block partial least-squares (2B-PLS) analysis to infer covariation between the landmark datasets. Other exploratory procedures were employed to assess measurement error: (1) a manual survey of digitized images to confirm landmark placement accuracy; (2) screening of outliers visualized in the morphospace plots and associated thin plate spline (TPS) deformation grids, as well as being identified using the plotOutliers search function in geomorph v. 4.0.0 [ 48 ].

Morphometric analysis

To standardize our digitized specimens for unit size, position, and rotation, we used a generalized Procrustes analysis (GPA) [ 49 , 50 ] that minimizes the bending energies to optimize the positions of the sliding semilandmarks [ 50 , 51 ]. Because large numbers of semilandmarks can impinge on GPA convergence [ 26 , 38 ], we varied the iteration frequency by arbitrarily increasing the max . iter argument in gpagan to compare convergence criteria Q-values (= Procrustes sum of squares). The resulting consensus shape configurations were then inspected (Fig F and Table H in S1 Text ).

The aligned Procrustes coordinates were ordinated via a principal components analysis (PCA) based on the singular value decomposition of the variance–covariance matrix. Shape variation was depicted as both TPS deformation grids and deformation isolines to generate a concentration “heat map” [ 52 ]. Morphospace was visualized using back-transformation [ 53 , 54 ]. All analyses were carried out in R v. 4.0.5 [ 44 ] with the geomorph v. 4.0.0 [ 48 ] and RRPP v. 1.0.0 [ 55 ] packages; visualization used the ggplot2 package [ 56 ]. All data and R scripts are available from Data Dryad under DOI: https://doi.org/10.5061/dryad.c866t1g5n [ 57 ].

Temporal analyses of morphospace

Multivariate normality was assessed using a Henze–Zirkler test [ 58 ] ( HZ = 4,956, p = 0) (Fig G in S1 Text ). Statistical comparisons between time-bins used a nonparametric Procrustes analysis of variance implemented in the RRPP package [ 55 ].

Temporal analyses of disparity

All SSW n values in a given time-bin t are then summed and divided by the sample size at that time ( N t ) to measure PV across all observations [ 48 ]. Note that the following equation was erroneously presented in Bazzi and colleagues [ 21 ].

Computationally, these equations are solved in geomorph [ 48 ], with the expectation that additive partial disparities [ 60 ] for sampling within each time-bin approximate the total PV given t .

H 0 assumes that pairwise absolute differences between PVs across 2 given time-bins (e.g., t1 and t2 ) will be zero. H A alternatively stipulates that the difference will be greater than zero.

We also applied nonparametric bootstrap resampling to estimate confidence intervals around disparity. All post hoc pairwise comparisons of group means were subject to false discovery rate (FDR) adjustments of p -values to mitigate the increased risk of Type I errors associated with multiple comparisons [ 61 ].

Geographic distribution in the fossil record

We accommodated for sample size biases inherent in the fossil record [ 4 , 62 – 64 ] via rarefaction comparisons of time-scaled PV [ 21 ]. These involved subsampling (999 iterations) of all time-bins to a minimum size (Tables A, C, and D in S1 Text ), after which 95% prediction intervals were calculated. Geographic subsampling focused on the UNESCO World Heritage fossil locality at Stevns Klint in Denmark (Fig H in S1 Text ), which preserves exceptionally rich selachimorph assemblages [ 65 , 66 ] spanning the K/Pg succession [ 67 ]. However, we also calculated partial disparities for each time-bin based on marine depositional basins, designated i in (6). Although the precise boundaries of these basins are ambiguous, they do provide a convenient proxy for comparing regional versus global disparity signals across a broader subsampled series.

Influences of heterodonty

Because developmental [ 70 , 71 ] and ontogenetic factors [ 70 , 72 – 74 ] also affect selachimorph tooth morphology, our analyses are presented with the caveat that adequate intraspecific coverage was assumed for each order-level clade.

Digitization measurement error

Visual comparison of the computed consensus (mean) tooth shapes (Fig I in S1 Text ) indicates consistent digitization across landmark datasets. Accordingly, an intraclass correlation coefficient ( R ) of 2% (or 1 in 50) was calculated based on the aligned Procrustes coordinates and their error replicate counterparts ( N = 30) (Table I in S1 Text ). Pearson product–moment correlation (t = 1,874.7, df = 9598, p -value << 0.001, R = 0.99) and a 2B-PLS test (r-pls = 0.997, p -value = 0.001, Z = 7.136) unambiguously demonstrate dataset compatibility (Fig J in S1 Text ).

PCA visualization

PC1 to PC4 explain 89.28% of the shape variance ( Fig 1A–1C , Fig K in S1 Text ), with the remaining PC axes collectively describing 10.71% (Fig K and Table J in S1 Text ). PC1 (62%) captures tooth height and width variation from apicobasally tall and narrow teeth to mesiodistally broad and low crowns (Fig L in S1 Text , panel A). PC2 (12%) alternatively represents distally recurved teeth with low “heels” versus upright triangular teeth with lateral cusplets (Fig L in S1 Text , panel B). PC3 (11%) tracks tall and conical to distally wide and recurved teeth with pronounced lateral cusplets (Fig L in S1 Text , panel C). PC4 (5%) depicts a spectrum of short triangular teeth with reduced cusplets to tall crowns with prominent cusplets commensurate in height with the main cusp (Fig L in S1 Text , panel D).

• PPT PowerPoint slide
• PNG larger image
• TIFF original image

(A) Multivariate shape space for the N = 1,239 global selachimorph tooth sample. Theoretical backtransform tooth shapes (gray) indicate shape variability across morphospace as defined by PC1/PC2. (B, C) Box-and-whisker plots indicating average morphospace occupation along (B) PC1 and (C) PC2. Error bars represent 95% confidence intervals. Proportion of variance and group sample sizes are listed in the axis labels. The data used in this analysis can be accessed online at https://doi.org/10.5061/dryad.c866t1g5n .

https://doi.org/10.1371/journal.pbio.3001108.g001

While there is a substantial overlap in tooth morphologies between selachimorph clades ( Fig 1A–1C ), a general Procrustes one-way ANOVA found significant differences between the group means, both generally and along specific axes of variance ( Table 1 ). Such results are supported by visible segregation between the clades in morphospace ( Fig 1A–1C ) and are further illustrated by measures of central tendency, distribution symmetries, and normality and multimodality tests (Table K in S1 Text ).

Coefficient estimation via OLS. Type I (sequential) sums of squares were used to calculate the sums of squares and cross-products matrices. Effect sizes (Z) are based on the F distribution.

https://doi.org/10.1371/journal.pbio.3001108.t001

Global and regional disparity

We find overall stability in selachimorph global disparity across the Campanian–Thanetian interval ( Fig 2 , Table L in S1 Text ). The only exception was a significant decline within the Selandian time-bin (PV Danian = 0.082; PV Selandian = 0.042; p = 0.033), which may be a product of sampling ( N = 28) and/or uneven clade representation. Disparity during the Thanetian exceeded ( Fig 2 , Fig M and Table L in S1 Text ) that of the preextinction Campanian (PV Campanian = 0.069; PV Thanetian = 0.085; p = 0.033) and was unaffected by pruning of †Synechodontiformes as a possible selachimorph stem group ( Fig 2 , Table M in S1 Text ). Comparisons with the Stevns Klint regional subsample found no significant change in tooth disparity (PV lateMaastrichtian = 0.091; PV earlyDanian = 0.069; p = 0.106) across the K/Pg boundary ( Fig 2 , Table N in S1 Text ). However, a significant disparity increase occurred after the extinction event from the early to middle Danian (PV earlyDanian = 0.069; PV middleDanian = 0.123; p = 0.005).

The 5-age time-binning scheme ( N = 1,156) utilized sample rarefaction on the lowest sampled Selandian time-bin ( N = 28). The 4-age time-binning scheme ( N = 1 198) used the Thanetian ( N = 187). The Stevns Klint regional subsample (3-age time-bin) used the middle Danian ( N = 25). Significant FDR-adjusted p -values for multiple comparisons are indicated (* p < 0.05, ** p < 0.005). Silhouette graphics created by MB. The data used in this analysis can be accessed online at https://doi.org/10.5061/dryad.c866t1g5n . FDR, false discovery rate; K/Pg, Cretaceous/Paleogene; Ma, million years; Sel, Selandian.

https://doi.org/10.1371/journal.pbio.3001108.g002

Superorder-level clade disparity

Relative stasis characterized the disparity of both galeomorphs and squalomorphs across the Maastrichtian–Danian-Selandian ( Fig 3A and 3B , Tables O, P, and Q in S1 Text ). This result is consistent even after the exclusion of †synechodontiforms ( Fig 3A , Table P in S1 Text ). Conversely, independent testing of the Stevns Klint regional subsample produces a significant disparity increase among galeomorphs (PV earlyDanian = 0.069; PV middleDanian = 0.109; p = 0.022) in the early to middle Danian ( Fig 3A , Tables R and S in S1 Text ) and a corresponding reduction in squalomorph disparity ( Fig 3B , Table T in S1 Text ) across the late Maastrichtian to early Danian (PV lateMaastrichtian = 0.077; PV earlyDanian = 0.041; p = 0.005).

(A, B) Global/regional raw and rarefied dental disparity trajectories for Galeomorphii (including and excluding †Synechodontiformes) and Squalomorphii. Sample rarefaction: Galeomorphii, N = 169; Squalimorphii, N = 18. Significant FDR-adjusted p -values for multiple comparisons are indicated (* p < 0.05, ** p < 0.005). Silhouette graphics created by MB. The data used in this analysis can be accessed online at https://doi.org/10.5061/dryad.c866t1g5n . FDR, false discovery rate; K/Pg, Cretaceous/Paleogene; Ma, million years; Sel, Selandian.

https://doi.org/10.1371/journal.pbio.3001108.g003

Order-level clade disparity

Lamniform and carcharhiniform ( Fig 4A and 4B , Tables U and V in S1 Text ) disparities were highest during the Campanian and Maastrichtian interval ( Fig 4A and 4B ). Despite these high levels, global lamniform disparity was demonstrably stable across the Maastrichtian–Danian-Selandian time-bins (PV Maastrichtian = 0.070; PV Danian-Selandian = 0.059; p = 0.396; Fig 4A ) and showed no significant change by the Thanetian (PV Maastrichtian = 0.070; PV Thanetian = 0.055; p = 0.264; Fig 4A ).

(A–D) Temporal resolution varies between groups but broadly tracks the time-binned K/Pg interval. Clade-level subsampling is shown in Tables C and D in S1 Text . Silhouette graphics created by MB and Julius Csotonyi ( https://csotonyi.com/ ). The data used in this analysis can be accessed online at https://doi.org/10.5061/dryad.c866t1g5n . K/Pg, Cretaceous/Paleogene; Ma, million years.

https://doi.org/10.1371/journal.pbio.3001108.g004

Carcharhiniforms otherwise exhibited a marked disparity decline from the Campanian to Danian-Selandian when using the global sample, but not in the Stevns Klint regional subsample ( Fig 4B , Table V in S1 Text ). A distinct disparity drop across the Campanian–Maastrichtian (PV Campanian = 0.085; PV Maastrichtian = 0.056; p = 0.024) was followed by stasis from the Maastrichtian–Danian-Selandian (PV Maastrichtian = 0.056; PV Danian-Selandian = 0.041; p = 0.125) and a subsequent disparity increase by the Thanetian (PV Danian-Selandian = 0.041; PV Thanetian = 0.061; p = 0.020) ( Fig 4B ).

Heterodontiform disparity was uniform across the Campanian–Thanetian interval ( Fig 4C , Table W in S1 Text ). On the other hand, orectolobiforms showed a significant disparity increase (PV Campanian = 0.024; PV Danian-Selandian = 0.064; p = 0.016) between the Campanian and Danian-Selandian ( Fig 4D , Table X in S1 Text ). We recover no significant change in orectolobiform disparity for any time-bin comparisons, including the Maastrichtian–Danian-Selandian (PV Maastrichtian = 0.041; PV Danian-Selandian = 0.064; p = 0.096), which is consistent with the results based on the Stevns Klint regional subsample (PV lateMaastrichtian = 0.109; PV earlyDanian = 0.066; p = 0.294) ( Fig 4D ).

Hexanchiform disparity was not statistically differentiated across the Maastrichtian–Danian-Selandian (PV Maastrichtian = 0.018; PV Danian-Selandian = 0.124; p = 0.104) nor between the Danian-Selandian and Thanetian (PV Danian-Selandian = 0.124; PV Thanetian = 0.054; p = 0.072) ( Fig 4E , Table Y in S1 Text ). Squaliforms similarly maintained stable disparity from the Maastrichtian–Danian-Selandian (PV Maastrichtian = 0.054; PV Danian-Selandian = 0.032; p = 0.084), both within the global and Stevns Klint regional samples ( Fig 4F , Table Z in S1 Text ). Finally, squatiniform and †synechodontiform disparities were also stable, but their small sample sizes make interpretations equivocal ( Fig 4G and 4H , Tables AA and BB in S1 Text ).

Global partial disparity

PVs indicate that selachimorph global disparity was driven by lamniforms during the Campanian and Maastrichtian ( Fig 5A ). Unsurprisingly, lamniforms were also the most numerically abundant taxa during this time (Fig A in S1 Text ). Lamniform contribution to global disparity subsequently decreased in the Danian ( Fig 5A ) but increased again in the Selandian, although the sample size is small (Table C in S1 Text ). Overall, lamniforms accounted for almost all of the global selachimorph variance calculated across the K/Pg time interval ( Fig 5A ). Carcharhiniforms and hexanchiforms otherwise contributed to a global disparity increase from the Danian to the Thanetian ( Fig 5A ).

(A, B) Grouped bar plot depicting clade-level contributions to overall disparity in each time-bin. (B) Families with sample sizes of N < 5 were omitted. Resulting data total = 26 families and 1,039 specimens with the global 4-age time-binning scheme. The data used in this analysis can be accessed online at https://doi.org/10.5061/dryad.c866t1g5n .

https://doi.org/10.1371/journal.pbio.3001108.g005

Despite uneven sampling, family-level partial disparities ( Fig 5B ) reveal that archaeolamnids, mitsukurinids, and anacoracids provided the primary sources of selachimorph disparity during the Campanian. Odontaspidids sequentially increased their disparity from the Danian-Selandian to the Thanetian ( Fig 5B ). Correspondingly, triakids, hexanchids, and squalids contributed markedly to non-lamniform disparity in the Danian-Selandian, with scyliorhinids increasing disparity in the Thanetian. Other clades, such as ginglymostomatids and hexanchids, also increased disparity from the Danian-Selandian–Thanetian ( Fig 5B ).

Geographic distribution of disparity

Samples from the North American Western Interior (WIB) and European epicontinental (EEB) basins accounted for most of the disparity during the Campanian ( Fig 6 ). Alternatively, the Maastrichtian incorporated a substantial contribution from the Ouled Abdoun and Tarfaya basins (= Eastern Atlantic rim [EAR]) of Morocco ( Fig 6 ). In general, geographic regions with small representative subsamples yielded comparatively low disparities during the Maastrichtian, including the North African Mediterranean Tethys and North American Western Atlantic rim (WAR) ( Fig 6 ). By contrast, intensive sampling of Stevns Klint in Denmark and the Limhamn Quarry in Sweden skewed the Danian-Selandian geographic distribution toward these regions and revealed the inordinate influence of lagerstätten deposits on our global disparity signal ( Fig 6 ). Lastly, the Thanetian showed a return to more geographically widespread sampling across a continuous trans -Atlantic belt, spanning the African epicontinental basins and EAR to the EEB and WAR ( Fig 6 ).

Mesozoic marine depositional basins are based on Wretman and colleagues [ 75 ]. Basins with sample sizes of N < 10 were omitted. Pie graphs depict sample proportions. The data used in this analysis can be accessed online at https://doi.org/10.5061/dryad.c866t1g5n .

https://doi.org/10.1371/journal.pbio.3001108.g006

Effects of heterodonty on disparity

Application of the monognathic heterodonty model (Tables CC and DD in S1 Text ) produced a significant disparity increase across the Maastrichtian–Danian-Selandian (PV Maastrichtian = 0.066; PV Danian-Selandian = 0.082; p = 0.032). The combined monognathic and dignathic heterodonty model similarly indicates a disparity increase from the Campanian to Maastrichtian ( p = 0.016). However, a dignathic heterodonty and multivariate monognathic*dignathic heterodonty interaction model yields no comparable disparity shifts (Fig N and Tables CC and DD in S1 Text ), and additional model comparisons found consistent disparity across the Campanian–Thanetian interval, despite some inflation of absolute values (Fig N in S1 Text ).

Global morphospace

Pairwise comparisons along all PCs found significant differences (F 4, 1,155 = 8.6269, p = 0.001) in morphospace distribution between time-bins ( Fig 7 , Table 2 ). The PC1 distributions are platykurtic (kurtosis < 3) and positively skewed, except during the Thanetian, which was negatively skewed ( g 1Thanetian = −0.0336). Distribution-specific interquartile ranges (IQRs) showed minimal morphospace dispersion along PC2 (Tables EE and FF in S1 Text ). A Hartigan dip test for multimodality did not reject a unimodal distribution on PC1 or PC2; however, a significant positive morphospace shift did occur from the Maastrichtian–Danian on PC1 ( Fig 7A , Fig O and Tables EE and FF in S1 Text ). Notably, there was no corresponding change in modal shape configurations, although a reduction in positively loaded morphospace accompanied shortening of the minimum value ranges ( Fig 7A , Fig O in S1 Text ).

(A, B) Jittered box plots visualizing the distribution of time-bins along PC1 and PC2. Graph depicts patterns of overall shape change using the 5-age time-binning scheme. Measures of central tendency, including the median (black line) and arithmetic mean (pink), are shown with computed 95% nonparametric bootstrap confidence intervals and potential outliers (closed black points). TPS grids indicate changes in modal value (green line) between time-bins. The data used in this analysis can be accessed online at https://doi.org/10.5061/dryad.c866t1g5n . Ma, million years; TPS, thin plate spline.

https://doi.org/10.1371/journal.pbio.3001108.g007

Coefficient estimation via OLS. Type I (sequential) sums of squares were used to calculate sums of squares and cross-products matrices. Effect sizes (Z) based on F distribution.

https://doi.org/10.1371/journal.pbio.3001108.t002

The average value along PC1 shift positively across the Selandian–Thanetian ( Fig 7A ). Comparisons between the Campanian–Maastrichtian ( p = 0.025) and Maastrichtian–Thanetian ( p = 0.013) likewise yield significant morphospace shifts (Table GG in S1 Text ).

The distributions for the Campanian and Danian on PC2 are leptokurtic (high kurtosis > 3), but low kurtosis prevails in the other time-bins ( Fig 7B ). A positive shift from the Maastrichtian ( g 1Maastrichtian = −0.0806) to Danian ( g 1 Danian = 0.2833) ( Fig 7B , Table EE in S1 Text ) coincided with a loss of negatively loaded tooth morphologies ( Fig 7B ). Nonetheless, the positively loaded morphologies diminished significantly from the Campanian to the Danian ( p = 0.008) and Maastrichtian–Danian ( p = 0.015) (Table GG in S1 Text ). A comparable morphospace shift across the K/Pg interval was detected using the 4-age time-binning scheme (Fig O in S1 Text ) and incorporated subtle changes in modal shape within the Campanian and Maastrichtian–Danian-Selandian time-bins (Fig O in S1 Text ).

Regional morphospace

The Stevns Klint regional subsample reveals no substantial shifts along PC1 or PC2 during the late Maastrichtian to early Danian ( Fig 8A and 8B , Table GG in S1 Text ). However, a significant positive sub-age shift in mean morphology occurred along PC1 from the late Maastrichtian to middle Danian ( p = 0.005), as well as between the early and middle Danian ( p = 0.005) ( Fig 8A , Table GG in S1 Text ). Both the late Maastrichtian ( g 1 late Maastrichtian = 0.2640) and early Danian ( g 1 early Danian = 0.2957) are characterized by positively skewed distributions, but with negative skewing during the middle Danian ( g 1 middle Danian = −0.1614). An increase in negatively loaded morphologies is associated with the early Danian on PC1 and is further reflected in the modal shape configuration ( Fig 8A ).

(A, B) Jittered box plots visualizing the distribution of time-bins along PC1 and PC2 ( n = 153). Measures of central tendency, including the median (black line) and arithmetic mean (pink), are shown with computed 95% nonparametric bootstrap confidence intervals and potential outliers (closed black points). TPS grids indicate changes in modal value (green line) between time-bins. The data used in this analysis can be accessed online at https://doi.org/10.5061/dryad.c866t1g5n . Ma, million years; TPS, thin plate spline.

https://doi.org/10.1371/journal.pbio.3001108.g008

The late Maastrichtian distribution on PC2 is negatively skewed, while the early to middle Danian exhibits positive skewing ( Fig 8B , Table GG in S1 Text ). Modal shape changes are more pronounced along PC2 ( Fig 8B ), gradually shifting from negatively to positively loaded values ( Fig 8B ).

Superorder-level clade morphospace

Galeomorphs and squalomorphs occupied comparable regions of morphospace on PC1 throughout the entire Maastrichtian–Thanetian interval ( Fig 9A ); although, some differences were evident in their mean and median values. On average, galeomorphs are characterized by tall and narrow teeth, whereas squalomorph teeth are typically low crowned. The most pronounced shift on PC1 is a reduction in negative values among galeomorphs at the K/Pg boundary ( Fig 9A ) and mirrored on PC3 and PC4 (Table HH in S1 Text ). Squalomorphs likewise exhibited a significant positive shift on PC2 ( p = 0.016) from the Maastrichtian–Danian-Selandian ( Fig 9A and 9B , Table II in S1 Text ).

(A–D) Jittered grouped box plots visualizing the distribution of time-bins along PC1 and PC2. Graph depicts patterns of overall shape change using the 4-age time-binning scheme. Measures of central tendency, including the median and arithmetic mean, are shown with computed 95% nonparametric bootstrap confidence intervals and potential outliers (closed black points). The data used in this analysis can be accessed online at https://doi.org/10.5061/dryad.c866t1g5n . MA, million years.

https://doi.org/10.1371/journal.pbio.3001108.g009

Order-level clade morphospace

Lamniforms displayed a significant positive shift in mean morphology along PC1 from the Maastrichtian–Danian-Selandian ( p = 0.002) ( Fig 10A , Table JJ in S1 Text ); this is concurrent with an overall positive to negative shift in their distribution ( g 1 Maastrichitan = 0.5035; g 1 Danian-Selandian = −0.3193). The frequency of negatively loaded PC1 morphologies is otherwise reduced during the Thanetian and is significantly different from both the Campanian ( p = 0.002) and Maastrichtian ( p = 0.002).

(A, B) Patterns along PC1 and PC2. Sampling was insufficient to estimate reliable disparity for Echinorhiniformes. Measures of central tendency, including the median and arithmetic mean, are shown with computed 95% nonparametric bootstrap confidence intervals and potential outliers. Silhouette graphics created by MB and Julius Csotonyi ( https://csotonyi.com/ ). The data used in this analysis can be accessed online at https://doi.org/10.5061/dryad.c866t1g5n .

https://doi.org/10.1371/journal.pbio.3001108.g010

Carcharhiniforms maintained stable morphospace occupation along PC1, but their distribution visibly contracted during the K/Pg transition. This is evidenced by a significant negative shift from the Campanian–Danian-Selandian time-bins ( p = 0.024) and an overall gradual reduction in positive PC1 values ( Fig 10A , Table KK in S1 Text ).

Heterodontiforms, orectolobiforms, hexanchiforms, squaliforms, squatiniforms, and †synechodontiforms all produce no significant deviations along PC1 across the Campanian–Thanetian interval ( Fig 10A , Tables LL to QQ in S1 Text ).

Lamniforms exhibited a significant ( p = 0.004) depletion of negative values from the Maastrichtian–Danian-Selandian on PC2 ( Fig 10B ). Pairwise comparisons also recovered a significant difference between the Campanian and the Danian–Selandian ( p = 0.004). By contrast, carcharhiniforms showed no major changes along PC2, except for a slight increase in positive skewness from the Danian-Selandian–Thanetian; this might imply an exploration of new morphospace ( Fig 10B ). We additionally detected significant differences between the Campanian and Danian-Selandian ( p = 0.031) and between the Campanian and Thanetian ( p = 0.007), although this signal reduced after FDR adjustment of the p -values (Table KK in S1 Text ).

No shifts were recorded for heterodontiforms or orectolobiforms on PC2 ( Fig 10B ). Yet, these clades did incline toward positive skewness. Squaliforms, squatiniforms, and †synechodontiforms likewise experienced no significant changes along PC2 (Tables OO to QQ in S1 Text ).

Disparity dynamics across the K/Pg boundary

Fundamentally, our results indicate that the global dental disparity of selachimorphs was largely stable across the end-Cretaceous mass extinction, even though some aspects of variance can be attributed to heterodonty (see Fig 2 , Fig N and Table CC in S1 Text ). Notably, this static disparity concurs with previous analyses of selachimorph disparity [ 21 ] and relative abundance patterns in the elasmobranch fossil record [ 17 ]. Conversely, diversity-based assessments advocate up to an approximately 50% species loss among selachimorphs over the same time interval [ 18 , 19 , 76 ].

Taxic decline [ 18 ] with no apparent ecological change is expected for a “nonselective” extinction model [ 77 ]. Yet, lamniforms as a dominant group did undergo a “selective” extinction of Cretaceous anacoracids [ 18 , 21 , 25 ], followed by the radiation of odontaspidids in the Paleocene ( Fig 10A and 10B ). Selective extinction among lamniforms is documented here through the loss of triangular, blade-like tooth morphologies typical of anacoracids versus the subsequent expansion of apicobasally tall, laterally cusped odontaspidid dental morphospace ( Fig 10A and 10B ). Such a transition would equate to a “shift” extinction model [ 21 ] and coincides with the ecological diversification of carcharhiniforms [ 21 ], which additionally invaded mesiodistally broad, low-crowned dental morphospace during the Paleocene ( Fig 10A and 10B ).

By comparison, squalomorphs were unaffected by the end-Cretaceous mass extinction, although a minor diversity decline is recognized among squalids [ 18 ]. We suggest that this group experienced a “nonselective” extinction [ 77 ], but with the caveat that their fossil record is poorly sampled (Table A in S1 Text ) and should be interpreted cautiously.

Moreover, we emphasize that the proliferation of apicobasally tall dental morphotypes in selachimorphs during the Paleocene does not evince an ecological turnover, unlike the appearance of novel dental grades [ 78 ] among coeval actinopterygians [ 17 , 20 , 79 – 82 ]. Rather, we propose that while selachimorphs withstood the end-Cretaceous mass extinction, they did suffer sufficient disturbance to trigger a compositional transformation in the morphology of their constituent clades [ 21 , 78 ]—an effect that is best illustrated by the documented fossil histories of lamniforms and carcharhiniforms [ 21 ].

Regional versus global extinction dynamics

Adolfssen and Ward [ 65 ] reported a 33% decline in chondrichthyan species richness across the K/Pg boundary based on fossils from Stevns Klint. Notably, this is substantially less than both the approximately 96% species loss calculated from K/Pg boundary deposits in Morocco [ 83 ] and the approximately 84% estimate derived from global sampling [ 18 ]. Stevns Klint preserves a succession of largely endemic Boreal assemblages [ 65 ], yet their static dental disparity is indistinguishable from that of the global sample ( Fig 2 ). Nonetheless, we recognize a minor disparity increase ( Fig 2 ) from the early to middle Danian that, although impacted by sampling as evident from rarefaction, corresponds with alterations in absolute disparity values among galeomorphs ( Fig 3A ). This slight increase, if real, suggests a region-specific postextinction recovery in the early Paleocene.

A more marked geographic transition occurs in regional sampling “hotspots” ( Fig 6 ), which shift from the WIB of North America during the Campanian and Maastrichtian, to the Scandinavian epeiric basins of Denmark and Sweden by the Danian, and, finally, to the Mediterranean Tethys and Atlantic shelf margins of North Africa and Morocco by the Thanetian. While these results undoubtedly capture preservational biases [ 84 ], they correlate with the changing depositional contexts of epicontinental environments across the K/Pg time frame. For example, the late Maastrichtian Western Interior Seaway regression [ 85 ] coincides with a regional disparity decline in North America, followed by disparity peaks in the transgressive Tethyan peripheries of Europe and North Africa during the Paleocene. Therefore, we suggest that, like other post-mass extinction marine ecosystems [ 86 – 88 ], the recovery of selachimorphs after the K/Pg event might have been geographically localized and centered on epeiric refugia [ 65 ] that provided high-productivity habitats conducive to rediversification.

Ecological implications

The end-Cretaceous mass extinction disproportionately affected larger-bodied lamniforms [ 18 , 19 , 21 , 25 ]. Body size is also implicated in the extinction of coeval marine reptiles, including mosasaurid lizards and plesiosaurians [ 89 – 91 ]. However, unlike these aquatic tetrapods, Cretaceous lamniform apex predators were supplanted by equally large-bodied hexanchiforms in some earliest Paleocene ecosystems [ 92 ]. This accords with our morphospace overlap of anacoracids and hexanchids (Fig P in S1 Text ) and is further compatible with some modern shark communities, in which the hexanchid, Notorynchus cepedianus , is known to invade apex predator niches once vacated by the lamnid, Carcharodon carcharias [ 93 ]. Unfortunately, shark body size is difficult to estimate accurately from fossil shark teeth [ 94 , 95 ], and the paleoecology of Cretaceous anacoracids and hexanchids is poorly understood. Irrespectively, the persistence of larger-bodied selachimorphs across the K/Pg boundary infers size-based niche continuity, a pattern consistent with static dental disparity (this study and [ 21 ]), and, thus, the likelihood that other drivers, including diet, habitat preference, and reproductive strategy [ 96 ], were influential in selecting for selachimorph extinction susceptibility. Indeed, the Paleocene radiation of piscivorous odontaspidids, triakids, and scyliorhinids coincided with the diversification of teleosts as an emerging food resource [ 17 ]. Correspondingly, we posit that feeding “specialization” might have been key to selective lineage loss [ 21 ], as well as the differential survival of more adaptable selachimorph “generalist” predators in the end-Cretaceous mass extinction aftermath.

Conclusions

Understanding the dynamics of shark evolution across the end-Cretaceous mass extinction event [ 18 , 19 , 21 , 22 , 76 ] has lagged analytically behind assessments of other marine vertebrate groups, such as teleost fishes [ 17 , 20 , 79 – 82 ] and aquatic reptiles [ 90 , 91 , 97 , 98 ]. Consequently, we present the first multiclade geometric morphometric evaluation of selachimorph disparity based on their extremely abundant dental fossil record. Our principal discovery of overall static disparity indicates that selachimorphs experienced no demonstrable preextinction decline or eco-morphological turnover as postulated for other vertebrate groups [ 89 , 90 , 99 , 100 ]. Nevertheless, the dominant Cretaceous anacoracids suffered a selective extinction, captured here by the loss of triangular, blade-like tooth morphologies traditionally associated with apex predator lifestyles (e.g., feeding on larger-bodied aquatic tetrapods [ 21 ]). Furthermore, we show that these extinctions are recognizable on both global and regional scales, although geographic shifts in disparity sampling “hotspots” could implicate changing epeiric habitat availability as another delimiting factor.

From a postextinction perspective, while anacoracids disappeared, other lamniform and carcharhiniform groups ecologically proliferated during the Paleocene. Most notably, this affected odontaspidids, triakids, and scyliorhinids, which are characterized by apicobasally tall, laterally cusped teeth. We interpret this as an extinction-mediated ecological “shift” involving significant changes in dental morphology without substantial modifications to selachimorph overall disparity [ 21 ]. Coincidentally, the Paleocene diversification of teleosts offers a potential driver, coupled with the dietary adaptability of selachimorphs as opportunistic “generalist” predators capable of exploiting emergent food resources.

Finally, our analyses underscore the utility of morphospace-disparity analyses to complement traditional taxonomy-based assessments of selachimorph diversity [ 101 , 102 ]. We advocate for similar approaches in future assessments of shark evolution but call attention to heterodonty and differential spatiotemporal sampling as sources of variation that likely mask much of the last 250 Ma of selachimorph eco-morphological evolution.

Supporting information

S1 text. supplementary methods and results..

Supplementary Tables: Tables A–VV. Supplementary Figures: Figs A–DD.

https://doi.org/10.1371/journal.pbio.3001108.s001

S1 Data. Global occurrence database.

https://doi.org/10.1371/journal.pbio.3001108.s002

S2 Data. 2D-landmark coordinate data.

https://doi.org/10.1371/journal.pbio.3001108.s003

S3 Data. Sliders file.

https://doi.org/10.1371/journal.pbio.3001108.s004

S4 Data. Measurement error landmark coordinate data.

https://doi.org/10.1371/journal.pbio.3001108.s005

S5 Data. Annotated R markdown file.

https://doi.org/10.1371/journal.pbio.3001108.s006

S6 Data. ReadMe file.

https://doi.org/10.1371/journal.pbio.3001108.s007

S1 Code. R markdown file containing reproducible R code for running geometric morphometric and statistical analyses of shark dental morphology.

https://doi.org/10.1371/journal.pbio.3001108.s008

S2 Code. R function used to compute a backtransform morphospace.

https://doi.org/10.1371/journal.pbio.3001108.s009

S3 Code. R function used to compute confidence and prediction intervals of the mean, based on a numeric vector.

https://doi.org/10.1371/journal.pbio.3001108.s010

S4 Code. R function used to compute univariate descriptive statistics.

https://doi.org/10.1371/journal.pbio.3001108.s011

S5 Code. R function used to compute and plot minimum convex hulls for a series of specified points.

https://doi.org/10.1371/journal.pbio.3001108.s012

S6 Code. R function used to compute Procrustes variance, bootstrapping, and rarefaction for GPA-aligned data.

https://doi.org/10.1371/journal.pbio.3001108.s013

S7 Code. R function used to generate evenly spaced points from point matrix.

https://doi.org/10.1371/journal.pbio.3001108.s014

Acknowledgments

We thank Mikael Siversson (Western Australian Museum) and Jesper Milan (Geomuseum Faxe) for discussions and access to material. Daniel Snitting (Uppsala University) assisted with processing of image data. Göran Arnqvist (Uppsala University) and Dean Adams (Iowa State University) contributed expertise on the morphometric and statistical analyses.

• View Article
• Google Scholar
• PubMed/NCBI
• 8. Cappetta H. Handbook of Paleoichthyology Volume 3E: Chondrichthyes Mesozoic and Cenozoic Elasmobranchii: Teeth. Munich, Germany: Verlag Dr. Friedrich Pfeil; 2012.
• 31. Ebert DA, Fowler SL, Compagno LJ. Sharks of the World: A Fully Illustrated Guide. Wild Nature Press; 2013.
• 37. Bookstein FL. Morphometric Tools for Landmark Data. Cambridge University Press; 1991. https://doi.org/10.1017/cbo9780511573064
• 38. Dryden IL, Mardia KV. Statistical Shape Analysis, with Applications in R: Second Edition. John Wiley & Sons Ltd.; 2016. https://doi.org/10.1002/9781119072492
• 43. Rohlf FJ. tpsDig. 2016. Available from: http//life.bio.sunysb.edu/morph/ . https://doi.org/10.1093/jac/dkf017
• 44. Team RC. R: A Language and Environment for Statistical Computing. Vienna, Austria. 2019.
• 68. Huber D, Wilga C, Dean M, Ferry L, Gardiner J, Habegger L, et al. Feeding in Cartilaginous Fishes: An Interdisciplinary Synthesis. Springer, Cham; 2019. pp. 231–295. https://doi.org/10.1007/978-3-030-13739-7_8

Academia.edu no longer supports Internet Explorer.

To browse Academia.edu and the wider internet faster and more securely, please take a few seconds to  upgrade your browser .

Enter the email address you signed up with and we'll email you a reset link.

• We're Hiring!
• Help Center

Index fossils

2006, Resonance

RELATED TOPICS

•   We're Hiring!
•   Help Center
• Find new research papers in:
• Health Sciences
• Earth Sciences
• Cognitive Science
• Mathematics
• Computer Science
• Academia ©2024

Encyclopedia of Scientific Dating Methods pp 1–2 Cite as

Index Fossil

• Peter Harries 3  
• Living reference work entry
• First Online: 01 January 2014

156 Accesses

Guide fossils ; Range fossils ; Zone fossils

A primary goal of the Earth sciences is to obtain as highly resolved correlations between different regions as possible. One of the most useful tools in this pursuit is the numerous fossils contained in most sedimentary units. The concept of using index fossils was initiated by the work of William Smith in the early 1800s; he used the fossil content of the units he examined as a critical component in the development of the first geologic map of the United Kingdom (Smith 1815 ). His observations on the distribution of fossils through time was formalized in his principle (or law) of faunal succession and served as the basis for determining the relative age of the rocks containing specific fossils. Building on Smith’s concept, later workers used the temporal intervals defined by index fossils to represent range zones that form the basis of biostratigraphy, the science of correlation using fossils. Thus, index fossils form...

This is a preview of subscription content, log in via an institution .

Bibliography

Kauffman, E. G., and Hazel, J. F. (eds.), 1977. Concepts and Methods of Biostratigraphy . Stroudsburg: Dowden, Hutchinson & Ross, p. 658.

Google Scholar  

Smith, W., 1815. A Delineation of the Strata of England and Wales with a Part of Scotland: Exhibiting the Collieries and Mines, the Marshes and Fen Lands Originally Overflowed by the Sea and the Varieties of Soil According to the Variations in the Substrata . London: John Cary.

Download references

Author information

Authors and affiliations.

School of Geosciences, University of South Florida, 4202 E. Fowler Avenue, CPR107, 33620, Tampa, FL, USA

Peter Harries

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Peter Harries .

Editor information

Editors and affiliations.

School of Geography and Earth Sciences, McMaster University, Hamilton, Ontario, Canada

W. Jack Rink

Department of Medical Physics and Applied Radiation Sciences, McMaster University, Hamilton, Ontario, Canada

Jeroen Thompson

Rights and permissions

Reprints and permissions

Copyright information

© 2013 Springer Science+Business Media Dordrecht

About this entry

Cite this entry.

Harries, P. (2013). Index Fossil. In: Rink, W., Thompson, J. (eds) Encyclopedia of Scientific Dating Methods. Springer, Dordrecht. https://doi.org/10.1007/978-94-007-6326-5_77-3

Download citation

DOI : https://doi.org/10.1007/978-94-007-6326-5_77-3

Received : 22 October 2013

Accepted : 22 October 2013

Published : 07 January 2014

Publisher Name : Springer, Dordrecht

Online ISBN : 978-94-007-6326-5

eBook Packages : Springer Reference Earth and Environm. Science Reference Module Physical and Materials Science Reference Module Earth and Environmental Sciences

• Publish with us

Policies and ethics

• Find a journal
• Track your research

• Engineering Geology
• Field Methods
• Historical Geology
• Hydrogeology
• Earth Interior
• Geologic Time Scale
• Plate Tectonics
• Geologic Lists
• Geological Wonders
• Igneous Rocks
• Extrusive Igneous Rocks
• Intrusive Igneous Rocks
• Metamorphic Rocks
• Foliated Metamorphic Rocks
• Sedimentary Rocks
• Metamorphic rocks
• Borate minerals
• Carbonates Minerals
• Halide Minerals
• Native Mineral
• Physical Properties
• Optical Properties
• Earthquakes
• Volcanic Eruption
• Privacy Policy

• Paleontology

Index Fossils

Index fossils play a crucial role in the field of geology, aiding scientists in dating and correlating rocks . These fossils are valuable indicators of specific time periods and help geologists reconstruct Earth’s history. By studying the distribution of index fossils in different rock layers, scientists can establish relative ages and understand the sequence of events that have shaped the Earth over millions of years.

Definition of Index Fossils:

Index fossils are the remains of once-living organisms that are particularly useful for dating and correlating the age of rocks. These fossils are distinctive, widespread, and existed for a relatively short period in geologic time. The key characteristic of index fossils is that they are associated with a specific time interval, making them reliable markers for dating rocks.

Importance in Geology:

• Stratigraphic Correlation: Index fossils help geologists correlate and match rock layers from different locations. When the same index fossil is found in distinct areas, it suggests that the rock layers containing these fossils formed during the same time period. This correlation is crucial for reconstructing the geologic history of a region.
• Relative Dating: Index fossils are essential for establishing the relative ages of rock layers. By identifying the presence of specific index fossils in a particular sequence of rock strata, geologists can determine which layers are older or younger relative to one another.
• Biostratigraphy: The study of the vertical distribution of fossils in rock layers, known as biostratigraphy, relies heavily on index fossils. This approach helps create a biostratigraphic column, allowing scientists to divide Earth’s history into distinct zones based on the types of fossils present.

Purpose in Dating and Correlating Rocks:

• Time Calibration: Index fossils provide a means of calibrating the geologic time scale . By associating certain fossils with specific time periods, scientists can assign numerical ages to rock layers, allowing for a more precise understanding of Earth’s history.
• Fossil Assemblages: The presence of specific groups of index fossils, known as fossil assemblages, aids in the identification of certain geologic time periods. Assemblages help geologists recognize the characteristics of different epochs, ages, and eras.
• Environmental Reconstruction: Index fossils can also provide insights into past environmental conditions. By studying the types of organisms preserved as index fossils, scientists can infer information about ancient ecosystems, climate, and habitats.

In summary, index fossils are invaluable tools for geologists, facilitating the dating and correlation of rocks and contributing to our understanding of Earth’s geologic history. Their distinctive characteristics and temporal significance make them essential for unraveling the mysteries of our planet’s past.

Criteria for Index Fossils

Examples of index fossils, role in relative dating, methods of index fossil dating, applications in stratigraphy.

The criteria for index fossils include characteristics that make them particularly useful for dating and correlating rocks. Here are the key criteria:

• Definition: Index fossils should have a broad geographic distribution. They should be found in different locations around the world.
• Importance: A widespread distribution ensures that the fossil is not limited to a specific locality, making it a reliable marker for correlating rock layers on a regional or even global scale.
• Definition: Index fossils should have existed for a relatively brief period in Earth’s history.
• Importance: Fossils with a short geologic range are helpful for pinpointing specific time intervals. Their presence in a rock layer can be indicative of a particular age, allowing for accurate dating of the associated rocks.
• Definition: Index fossils should be relatively abundant in the rock layers where they are found.
• Importance: Abundance increases the likelihood of finding the fossil in various locations and ensures that there are enough specimens to provide a robust basis for correlation. Rare fossils are less reliable as indicators because their scarcity makes it harder to establish correlations.
• Definition: Index fossils should possess unique and easily recognizable physical characteristics.
• Importance: The distinctive morphology of these fossils makes them readily identifiable, reducing the likelihood of confusion with other species. This characteristic is essential for accurate correlation and dating of rocks.

These criteria collectively make index fossils powerful tools for stratigraphic correlation and relative dating. The combination of widespread distribution, a short geologic range, abundance, and distinctive morphology enhances the reliability of these fossils as markers for specific time periods in Earth’s history.

Several organisms have been identified as index fossils due to their widespread distribution, short geologic range, abundance, and distinctive morphology. Here are a few examples:

• Characteristics: These extinct marine arthropods had a hard exoskeleton and segmented body.
• Geologic Range: Trilobites were abundant throughout the Paleozoic Era, with various species existing during specific time intervals.
• Characteristics: Ammonites were coiled, chambered marine cephalopods with a distinct spiral shell.
• Geologic Range: They were prevalent in the oceans from the Devonian to the Cretaceous Periods.
• Characteristics: Brachiopods are marine organisms with two shells, similar to bivalve mollusks, but with a different internal anatomy.
• Geologic Range: They were abundant in the Paleozoic and Mesozoic Eras.
• Characteristics: Microscopic marine protists with calcareous or siliceous shells.
• Geologic Range: Foraminifera have been present since the Cambrian Period and are still extant today. Different species are associated with specific time intervals.
• Characteristics: Belemnites were squid-like cephalopods with a straight, bullet-shaped shell.
• Geologic Range: Abundant in the Jurassic and Cretaceous Periods.
• Characteristics: Colonial marine animals with a distinctive fossilized branching structure.
• Geologic Range: Graptolites were abundant from the Ordovician to the Devonian Periods.
• Characteristics: Dinosaurs were diverse reptiles with various body shapes and sizes.
• Geologic Range: Dinosaurs were predominant during the Mesozoic Era, from the Triassic to the Cretaceous Periods.
• Characteristics: Large, tusked mammals related to modern elephants.
• Geologic Range: Mammoths and mastodons were present during the Pleistocene Epoch.

These examples demonstrate the diversity of organisms that have served as index fossils, covering different time periods and environments. The specific choice of index fossils can vary depending on the region and the geological context being studied.

Index fossils play a crucial role in relative dating, which is a method used by geologists to determine the chronological order of events in Earth’s history without assigning specific numerical ages to the rocks. Here’s how index fossils contribute to relative dating:

• Index fossils are used to correlate rock layers (strata) across different geographic locations. When the same index fossil is found in two or more areas, it indicates that the rock layers containing those fossils were likely deposited during the same time period. This correlation helps geologists create a consistent and interconnected stratigraphic record.
• Biostratigraphy is a branch of stratigraphy that uses the distribution of fossils to establish the relative ages of rock layers. Index fossils are essential in biostratigraphy because they allow geologists to divide the rock record into biozones or biostratigraphic units. These units are characterized by the presence of specific index fossils and help organize the geological timeline.
• By examining the vertical sequence of rock layers, geologists can infer relative ages based on the presence or absence of index fossils. For example, if a layer contains an index fossil known to have a short geologic range, it can be inferred that the rock layer is relatively young. Conversely, if a layer lacks certain index fossils but contains others, it suggests a different relative age.
• The study of fossil assemblages, which involves analyzing the combination of fossils present in a particular rock layer, helps geologists determine the relative age of that layer. Index fossils, with their distinctive characteristics, aid in identifying specific time periods and constructing a relative chronological sequence.
• Index fossils are often used as zonal markers, helping geologists define specific zones or intervals within the rock record. Each zone is characterized by the presence of a particular index fossil or assemblage, allowing for a detailed and nuanced understanding of the relative ages of different parts of the geological column.

In summary, index fossils are invaluable tools in relative dating because they provide recognizable markers tied to specific time intervals. By studying the distribution and characteristics of these fossils, geologists can establish the relative order of rock layers and construct a framework for understanding the sequence of events in Earth’s history.

Index fossil dating, a form of relative dating, involves the use of distinctive fossils to establish the relative ages of rock layers and the events they represent. Here are some common methods used in index fossil dating:

• Definition: Biostratigraphy is the primary method of index fossil dating. It involves the study of the vertical distribution of fossils in rock layers.
• Procedure: Geologists identify specific index fossils or assemblages within rock layers. These fossils are known to have short geologic ranges, meaning they existed for a specific and relatively brief period. By analyzing the presence, absence, and sequence of these fossils, geologists can establish the relative ages of the rocks.
• Definition: Zonal fossil assemblages are groups of fossils associated with specific time intervals. Different zones are defined by the presence of particular index fossils or assemblages.
• Procedure: Geologists divide the rock record into zones based on the types of fossils present. Each zone corresponds to a particular time period. The presence of a specific index fossil within a zone helps date the rocks associated with that fossil.
• Definition: Fossil range charts provide a visual representation of the temporal distribution of various fossils.
• Procedure: Geologists create charts that show the known ranges of different fossils over time. Index fossils are highlighted, indicating the time periods during which they were present. By comparing the fossil assemblage in a rock layer to the range chart, geologists can estimate the relative age of the rocks.
• Definition: Faunal succession is the concept that fossil assemblages change over time in a predictable manner.
• Procedure: Geologists observe the progression of fossil assemblages in rock layers. Certain index fossils are associated with specific stages of evolution or environmental conditions. The relative position of these fossils in the rock sequence helps establish the chronological order of events.
• Definition: Correlating rock sequences involves matching and aligning similar rock layers from different locations.
• Procedure: Geologists identify common index fossils in rock layers from different regions. The presence of the same index fossil in corresponding layers suggests contemporaneous deposition. This correlation helps create a broader understanding of the relative ages of rocks on a regional or global scale.
• Definition: The Principle of Faunal Succession states that fossils succeed each other in a definite and recognizable order over geological time.
• Procedure: By applying this principle, geologists can use the distinctive characteristics of index fossils to determine the relative ages of rock layers. The presence of specific fossils in a sequence follows a predictable pattern.

These methods collectively contribute to the accurate dating and correlation of rock layers, allowing geologists to construct a detailed relative chronological framework for Earth’s history.

Stratigraphy, the study of rock layers (strata) and their arrangement, relies heavily on the use of index fossils for dating and interpreting Earth’s history. Index fossils have several applications in stratigraphy, contributing to our understanding of the temporal and spatial relationships within the Earth’s crust. Here are some key applications:

• Index fossils are fundamental for establishing the relative ages of rock layers. By identifying the presence of specific index fossils in different strata, geologists can determine which layers are older or younger relative to one another. This aids in constructing a chronological sequence of events over geological time.
• Index fossils play a crucial role in correlating rock layers across different geographic locations. When the same index fossil is found in separate areas, it suggests contemporaneous deposition. This allows geologists to correlate and link rock formations, creating a comprehensive understanding of regional and global stratigraphy.
• Biostratigraphy involves the use of fossils to subdivide and correlate rock sequences. Index fossils are essential in this process. By identifying and studying the distribution of specific fossils, geologists can establish biozones and create detailed stratigraphic charts that help organize the geological timeline.
• Index fossils often serve as zonal markers, defining specific zones or intervals within the rock record. Each zone corresponds to a particular time period characterized by the presence of a distinct index fossil or fossil assemblage. Zonal markers contribute to the precise subdivision of stratigraphic sequences.
• Sequence stratigraphy involves the study of depositional sequences and their bounding surfaces within sedimentary rocks . Index fossils are used to identify key surfaces and transitions between different depositional environments. This helps geologists understand the changing conditions and events that influenced sedimentation over time.
• Index fossils aid in facies analysis, the study of lateral changes in sedimentary rock characteristics. By correlating the occurrence of specific fossils with variations in lithology, geologists can discern changes in environmental conditions, such as shifts in sea level or depositional environments, within a stratigraphic sequence.
• Event stratigraphy involves identifying and correlating specific geologic events recorded in the rock layers. Index fossils can be used to mark significant events, such as mass extinctions or evolutionary radiations. These events serve as important stratigraphic markers and help refine the stratigraphic framework.
• Index fossils provide valuable information for reconstructing past environments. The types of organisms found in a particular stratum can indicate the environmental conditions prevalent during that time, contributing to the broader understanding of Earth’s paleoenvironments.

In summary, the applications of index fossils in stratigraphy are diverse and multifaceted, ranging from establishing relative ages to correlating rock sequences and understanding past environmental conditions. These applications collectively contribute to the development of a comprehensive and detailed stratigraphic framework.

In conclusion , index fossils play a crucial role in the field of geology, particularly in stratigraphy and relative dating. Key points regarding index fossils include their widespread distribution, short geologic range, abundance, and distinctive morphology. These characteristics make them reliable markers for correlating rock layers, establishing relative ages, and reconstructing Earth’s history.

Summary of Key Points:

• Definition: Index fossils are distinctive remains of organisms that are useful for dating and correlating rocks due to their specific characteristics and temporal significance.
• Criteria: Index fossils should have a widespread distribution, a short geologic range, abundance, and a distinctive morphology.
• Role in Relative Dating: Index fossils are essential for stratigraphic correlation, biostratigraphy, and establishing the relative ages of rock layers without assigning specific numerical ages.
• Applications in Stratigraphy: Index fossils contribute to relative age dating, stratigraphic correlation, biostratigraphy, zonal markers, sequence stratigraphy, facies analysis, event stratigraphy, and paleoenvironmental reconstruction.

Significance in Earth Sciences:

Index fossils provide a unique window into Earth’s past, allowing scientists to decipher the chronology of events, changes in ecosystems, and shifts in environmental conditions. They are fundamental tools for understanding the history of life on our planet, the evolution of species, and the geological processes that have shaped the Earth’s surface.

Future Research Directions:

• Refinement of Chronostratigraphy: Continued research aims to refine chronostratigraphy by improving the accuracy of dating methods and expanding the database of index fossils. Advances in technology, such as more precise dating techniques, can contribute to a more detailed understanding of Earth’s timeline.
• Integration of Multi-disciplinary Approaches: Future research may involve integrating multiple scientific disciplines, such as paleontology , geochronology, and geochemistry, to enhance the reliability and precision of stratigraphic correlations.
• Exploration of Extinct Ecosystems: The study of index fossils can provide insights into past ecosystems and biodiversity. Future research may focus on reconstructing and understanding extinct ecosystems using a combination of fossil data and environmental proxies.
• Global Correlations: As technology and data-sharing capabilities advance, researchers can work towards establishing more robust global correlations of rock sequences. This could lead to a more comprehensive understanding of Earth’s geological and biological history on a global scale.
• Application in Extraterrestrial Stratigraphy: With ongoing exploration of other planets and celestial bodies, the principles of stratigraphy and the concept of index fossils could be applied to understand the geological histories of these extraterrestrial environments.

In essence, the study of index fossils will continue to be a dynamic and evolving field, contributing to our expanding knowledge of Earth’s history and potentially shedding light on the geological histories of other celestial bodies in the future.

RELATED ARTICLES MORE FROM AUTHOR

Stromatolites

Foraminifera

Orthoceras Fossil

10 Stunning Fossils That Provide Insights into Earth’s History

Banded Iron Formations (BIFs)

Full List of Index Fossils

Recent posts.

Turquoise Howlite

Atlantasite (Tasmanite)

Boji Stones (Shaman Stones)

Search Google Appliance

• Online Book Collections
• Online Books by Topic
• Biodiversity Heritage Library
• Library Catalog (SIRIS)
• Image Gallery
• Art & Artist Files
• Caldwell Lighting
• Trade Literature
• All Digital Collections
• Current Exhibitions
• Online Exhibitions
• Past Exhibitions
• Index of Library & Archival Exhibitions on the Web
• Research Tools and OneSearch
• E-journals, E-books, and Databases
• Smithsonian Research Online (SRO)
• Borrowing and Access Privileges
• Smithsonian Libraries and Archives on PRISM (SI staff)
• E-news Sign Up
• Internships and Fellowships
• Work with Us
• About the Libraries
• Library Locations
• Departments
• History of the Libraries
• Advisory Board
• Annual Reports
• Adopt-a-Book
• Ways to Give
• Gifts-in-Kind

You are here

Paleontology research guide, smithsonian resources, paleontology.

The Smithsonian Libraries and Archives' Paleontology Research Guide is a select list of resources for students, teachers, and researchers to learn about paleontology. 

• National Museum of Natural History Department of Paleobiology : Website for the Smithsonian division which studies of fossil animals, plants, and protists that provides information about Smithsonian research and collections.
• Cerro Ballena : Website exploring how paleontologists studied Cerro Ballena, a unique paleontological site in Chile, and what they have found about the fossils found there, featuring images, maps, and 3-D models of their finds.
• Pyenson Lab : Blog from the Pysenson Lab, a research group studying the evolution and paleobiology of marine vertebrates, with posts about their field work, research, and marine paleobiology more generally.
• Vertebrate Paleontology Library : Smithsonian library which focuses on the physical geography and paleozoology and vertebrates of the Paleozoic, Mesozoic, Tertiary, and Quaternary periods with over 1,800 volumes.
• Natural History Research Guide : Includes resources from the National Museum of Natural History and its libraries, links to other museums and libraries, and directories and databases related to biodiversity.​
• Science Teaching Resources (Life Science ): Collection of Life Science related lessons, activities, literacy resources, and videos from Q?rius at the National Museum of Natural History.
• Integrated Taxonomic Information System : Taxonomic database searchable by scientific and common names maintained through a partnership of U.S., Canadian, and Mexican federal agencies including the Smithsonian.
• Databases for Science Research : List of science research databases from the Smithsonian Libraries. Many of the databases are free access, but others do require users to be onsite at a Smithsonian library or have Smithsonian network access.
• Biodiversity Heritage Library : Online library featuring open access legacy literature from the Smithsonian Libraries and a consortium of other natural history and botanical libraries.
• Unearthed! A Digital Dig into the Smithsonian Libraries' Paleobiology Collection : Online collection created by Smithsonian Libraries to showcase the Biodiversity Heritage Library's holdings of books pertaining to paleobiology.
• The Dino Directory : Directory of known dinosaur species from the British Natural History Museum searchable by name, geologic age, and location featuring an illustration, taxanomy, and dietary information for each species.
• The Dinosauria : Educational website from the University of California - Berkeley featuring information about dinosaurs, their fossil record, their lives, and their taxonomy.
• Natural History Museum - Dinosaurs : Website from the British Natural History museum featuring information about their dinosaur collection, dinosaur crafts, and articles about dinosaur origins, evolution, and extinction.
• The Paleobiology Database : Database providing fossil occurrence and taxonomic data across geological ages. The includes an interactive map to display data, which can also be filtered by geological age.
• Berkeley Paleo Collections : Digitized image collections of microfossil, vertebrate, plant, and invertebrate specimens from the University of California Museum of Paleontology.
• Index to Organism Names : Searchable index of millions of scientific names, both fossil and current, from the scientific literature.
• A Guide to the Orders of Trilobites : An online guide featuring information about the morphology, classification, and paleobiology of trilobites, as well as many images and fact sheets about the ancient arthropods.
• NMITA : Taxonomic database of marine life from the Neogene period, with maps, identification guides, and photos for selected species.
• Paleoclimatology Data : Collection of resources about Paleoclimatology from the National Oceanic and Atmospheric Administration. The resources include datasets, current projects, and presentations.
• Geology, Paleontology & Theories of the Earth : Digitized collection of early and influential books in paleontology and geoscience from the Linda Hall Library.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• NEWS ROUND-UP
• 19 January 2022

Fossil-record bias and huge research database

Fossil record skewed by rich countries

Access options

Access Nature and 54 other Nature Portfolio journals

Get Nature+, our best-value online-access subscription

24,99 € / 30 days

cancel any time

Subscribe to this journal

Receive 51 print issues and online access

185,98 € per year

only 3,65 € per issue

Rent or buy this article

Prices vary by article type

Prices may be subject to local taxes which are calculated during checkout

Nature 601 , 303 (2022)

doi: https://doi.org/10.1038/d41586-022-00103-9

• Palaeontology

A guide to the Nature Index

Nature Index 13 MAR 24

Decoding chromatin states by proteomic profiling of nucleosome readers

Article 06 MAR 24

‘All of Us’ genetics chart stirs unease over controversial depiction of race

News 23 FEB 24

Fossils document evolutionary changes of jaw joint to mammalian middle ear

Article 03 APR 24

Jurassic shuotheriids show earliest dental diversification of mammaliaforms

Adaptive foraging behaviours in the Horn of Africa during Toba supereruption

Article 20 MAR 24

Postdoc Fellow / Senior Scientist

The Yakoub and Sulzer labs at Harvard Medical School-Brigham and Women’s Hospital and Columbia University

Boston, Massachusetts (US)

Harvard Medical School and Brigham and Women's Hospital

Postdoc in Computational Genomics – Machine Learning for Multi-Omics Profiling of Cancer Evolution

Computational Postdoc - Artificial Intelligence in Oncology and Regulatory Genomics and Cancer Evolution at the DKFZ - limited to 2 years

Heidelberg, Baden-Württemberg (DE)

German Cancer Research Center in the Helmholtz Association (DKFZ)

Computational Postdoc

The German Cancer Research Center is the largest biomedical research institution in Germany.

PhD / PostDoc Medical bioinformatics (m/f/d)

The Institute of Medical Bioinformatics and Systems Medicine / University of Freiburg is looking for a PhD/PostDoc Medical bioinformatics (m/w/d)

Freiburg im Breisgau, Baden-Württemberg (DE)

University of Freiburg

Postdoctoral Research Associate position at University of Oklahoma Health Sciences Center

Postdoctoral Research Associate position at University of Oklahoma Health Sciences Center   The Kamiya Mehla lab at the newly established Departmen...

Oklahoma City, Oklahoma

University of Oklahoma Health Sciences Center

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

AI Index Report

Welcome to the seventh edition of the AI Index report. The 2024 Index is our most comprehensive to date and arrives at an important moment when AI’s influence on society has never been more pronounced. This year, we have broadened our scope to more extensively cover essential trends such as technical advancements in AI, public perceptions of the technology, and the geopolitical dynamics surrounding its development. Featuring more original data than ever before, this edition introduces new estimates on AI training costs, detailed analyses of the responsible AI landscape, and an entirely new chapter dedicated to AI’s impact on science and medicine.

Read the 2024 AI Index Report

The AI Index report tracks, collates, distills, and visualizes data related to artificial intelligence (AI). Our mission is to provide unbiased, rigorously vetted, broadly sourced data in order for policymakers, researchers, executives, journalists, and the general public to develop a more thorough and nuanced understanding of the complex field of AI.

The AI Index is recognized globally as one of the most credible and authoritative sources for data and insights on artificial intelligence. Previous editions have been cited in major newspapers, including the The New York Times, Bloomberg, and The Guardian, have amassed hundreds of academic citations, and been referenced by high-level policymakers in the United States, the United Kingdom, and the European Union, among other places. This year’s edition surpasses all previous ones in size, scale, and scope, reflecting the growing significance that AI is coming to hold in all of our lives.

Steering Committee Co-Directors

Ray Perrault

Steering committee members.

Erik Brynjolfsson

John Etchemendy

Katrina Ligett

Terah Lyons

James Manyika

Juan Carlos Niebles

Vanessa Parli

Yoav Shoham

Russell Wald

Staff members.

Loredana Fattorini

Nestor Maslej

Letter from the co-directors.

A decade ago, the best AI systems in the world were unable to classify objects in images at a human level. AI struggled with language comprehension and could not solve math problems. Today, AI systems routinely exceed human performance on standard benchmarks.

Progress accelerated in 2023. New state-of-the-art systems like GPT-4, Gemini, and Claude 3 are impressively multimodal: They can generate fluent text in dozens of languages, process audio, and even explain memes. As AI has improved, it has increasingly forced its way into our lives. Companies are racing to build AI-based products, and AI is increasingly being used by the general public. But current AI technology still has significant problems. It cannot reliably deal with facts, perform complex reasoning, or explain its conclusions.

AI faces two interrelated futures. First, technology continues to improve and is increasingly used, having major consequences for productivity and employment. It can be put to both good and bad uses. In the second future, the adoption of AI is constrained by the limitations of the technology. Regardless of which future unfolds, governments are increasingly concerned. They are stepping in to encourage the upside, such as funding university R&D and incentivizing private investment. Governments are also aiming to manage the potential downsides, such as impacts on employment, privacy concerns, misinformation, and intellectual property rights.

As AI rapidly evolves, the AI Index aims to help the AI community, policymakers, business leaders, journalists, and the general public navigate this complex landscape. It provides ongoing, objective snapshots tracking several key areas: technical progress in AI capabilities, the community and investments driving AI development and deployment, public opinion on current and potential future impacts, and policy measures taken to stimulate AI innovation while managing its risks and challenges. By comprehensively monitoring the AI ecosystem, the Index serves as an important resource for understanding this transformative technological force.

On the technical front, this year’s AI Index reports that the number of new large language models released worldwide in 2023 doubled over the previous year. Two-thirds were open-source, but the highest-performing models came from industry players with closed systems. Gemini Ultra became the first LLM to reach human-level performance on the Massive Multitask Language Understanding (MMLU) benchmark; performance on the benchmark has improved by 15 percentage points since last year. Additionally, GPT-4 achieved an impressive 0.97 mean win rate score on the comprehensive Holistic Evaluation of Language Models (HELM) benchmark, which includes MMLU among other evaluations.

Although global private investment in AI decreased for the second consecutive year, investment in generative AI skyrocketed. More Fortune 500 earnings calls mentioned AI than ever before, and new studies show that AI tangibly boosts worker productivity. On the policymaking front, global mentions of AI in legislative proceedings have never been higher. U.S. regulators passed more AI-related regulations in 2023 than ever before. Still, many expressed concerns about AI’s ability to generate deepfakes and impact elections. The public became more aware of AI, and studies suggest that they responded with nervousness.

Ray Perrault Co-director, AI Index

Our Supporting Partners

Analytics & Research Partners

Stay up to date on the AI Index by subscribing to the  Stanford HAI newsletter.

• Share full article

Advertisement

Supported by

Harvard and Caltech Will Require Test Scores for Admission

The universities are the latest highly selective schools to end their policies that made submitting SAT or ACT scores optional.

By Anemona Hartocollis and Stephanie Saul

Harvard will reinstate standardized testing as a requirement of admission, the university announced Thursday, becoming the latest in a series of highly competitive universities to reverse their test-optional policies.

Students applying to enter Harvard in fall 2025 and beyond will be required to submit SAT or ACT scores, though the university said a few other test scores will be accepted in “exceptional cases,” including Advanced Placement or International Baccalaureate tests. The university had previously said it was going to keep its test-optional policy through the entering class of fall 2026.

Within hours of Harvard’s announcement, Caltech, a science and engineering institute, also said it was reinstating its testing requirements for students applying for admission in fall 2025.

The schools had been among nearly 2,000 colleges across the country that dropped test score requirements over the last few years, a trend that escalated during the pandemic when it was harder for students to get to test sites.

Dropping test score requirements was widely viewed as a tool to help diversify admissions, by encouraging poor and underrepresented students who had potential but did not score well on the tests to apply. But supporters of the tests have said without scores, it became harder to identify promising students who outperformed in their environments.

In explaining its decision to accelerate the return to testing, Harvard cited a study by Opportunity Insights , which found that test scores were a better predictor of academic success in college than high school grades and that they can help admissions officers identify highly talented students from low income groups who might otherwise had gone unnoticed.

“Standardized tests are a means for all students, regardless of their background and life experience, to provide information that is predictive of success in college and beyond,” Hopi Hoekstra, dean of the faculty of arts and sciences, said in a statement announcing the move.

“In short, more information, especially such strongly predictive information, is valuable for identifying talent from across the socioeconomic range,” she added.

Caltech, in Pasadena, Calif., said that reinstating testing requirements reaffirmed the school’s “commitment as a community of scientists and engineers to using all relevant data in its decision-making processes.”

Harvard and Caltech join a growing number of schools, notable for their selectivity, that have since reversed their policies, including Brown, Yale, Dartmouth, M.I.T., Georgetown, Purdue and the University of Texas at Austin.

For Harvard, the move comes at a time of transition, and perhaps a return to more conservative policies.

Last June, the Supreme Court struck down race-conscious college admissions in cases involving Harvard and the University of North Carolina, raising fears that with the demise of affirmative action, those schools would become less diverse.

And in January, Harvard’s first Black president, Claudine Gay, resigned under pressure from critics who said she had not acted strongly enough to combat antisemitism on campus after the Oct. 7 attack by Hamas on Israel, and under mounting accusations of plagiarism in her academic work, which she stood by.

The provost, Alan Garber, was named interim president, while the dean of the law school, John Manning, became interim provost, the university’s second-highest administrative position. Mr. Manning is considered a strong potential candidate to replace Dr. Gay. His background stands out for his conservative associations, having clerked for the former Supreme Court justice Antonin Scalia.

In the current climate on campus, a return to test scores could be seen as a return to tradition. It also may address concerns of many parents that the college admissions process, especially in elite institutions, is inscrutable and disconnected from merit.

Applications to Harvard were down by 5 percent this year, while those at many of its peer universities went up, suggesting that the recent turmoil may have dented its reputation. But it still received a staggering number of undergraduate applications — 54,008 — and admitted only 3.6 percent. Requiring test scores could make sorting through applications more manageable.

Critics of standardized tests have long raised concerns that the tests helped fuel inequality because some wealthier students raised their scores through high-priced tutoring. But recent studies have found that test scores help predict college grades, chances of graduation and post-college success, and that test scores are more reliable than high school grades, partly because of grade inflation in recent years .

But Robert Schaeffer, director of public education at FairTest, an organization that opposes standardized testing, said Thursday that the Opportunity Insights analysis had been criticized by other researchers. “Those scholars say that when you eliminate the role of wealth, test scores are not better than high school G.P.A.,” he said, adding that it is not clear whether that pattern is true among the admissions pool at super selective colleges such as Harvard.

Mr. Schaeffer said that at least 1,850 universities remain test optional, including Michigan, Vanderbilt, Wisconsin and Syracuse, which have recently extended their policies. “The vast majority of colleges will not require test scores.” An exception, he said, could be the University of North Carolina system, which is considering a plan to require tests, but only for those students with a G.P.A. below 2.8.

Acknowledging the concerns of critics, Harvard said that it would reassess the new policy regularly. The school said that test scores would be considered along with other information about an applicant’s experience, skills, talents, contributions to communities and references. They will also be looked at in the context of how other students are doing at the same high school.

“Admissions officers understand that not all students attend well-resourced schools, and those who come from modest economic backgrounds or first-generation college families may have had fewer opportunities to prepare for standardized tests,” William R. Fitzsimmons, Harvard’s dean of admissions and financial aid, said in a statement.

Harvard said that in the interest of selecting a diverse student body, it has enhanced financial aid and stepped up recruitment of underserved students by joining a consortium of 30 public and private universities that recruits students from rural communities.

An earlier version of this article misstated Robert Schaeffer’s position. He is the director of public education at FairTest, not the director.

How we handle corrections

Anemona Hartocollis is a national reporter for The Times, covering higher education. More about Anemona Hartocollis

Stephanie Saul reports on colleges and universities, with a recent focus on the dramatic changes in college admissions and the debate around diversity, equity and inclusion in higher education. More about Stephanie Saul