g uly h uv - bioone › journals › paleobiology › volume-43 › issue...herbivorous dinosaur jaw...

Herbivorous dinosaur jaw disparity and its relationshipto extrinsic evolutionary drivers

Authors: MacLaren, Jamie A., Anderson, Philip S. L., Barrett, Paul M.,and Rayfield, Emily J.

Source: Paleobiology, 43(1) : 15-33

Published By: The Paleontological SocietyURL: https://doi.org/10.1017/pab.2016.31

BioOne Complete (complete.BioOne.org) is a full-text database of 200 subscribed and open-access titlesin the biological, ecological, and environmental sciences published by nonprofit societies, associations,museums, institutions, and presses.

Your use of this PDF, the BioOne Complete website, and all posted and associated content indicates youracceptance of BioOne’s Terms of Use, available at www.bioone.org/terms-of-use.

Usage of BioOne Complete content is strictly limited to personal, educational, and non - commercial use.Commercial inquiries or rights and permissions requests should be directed to the individual publisher ascopyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofitpublishers, academic institutions, research libraries, and research funders in the common goal of maximizing access tocritical research.

Downloaded From: https://bioone.org/journals/Paleobiology on 16 Jun 2020Terms of Use: https://bioone.org/terms-of-use

Herbivorous dinosaur jaw disparity and its relationship toextrinsic evolutionary drivers

Jamie A. MacLaren, Philip S. L. Anderson, Paul M. Barrett, and Emily J. Rayfield

Abstract.—Morphological responses of nonmammalian herbivores to external ecological drivers have notbeen quantified over extended timescales. Herbivorous nonavian dinosaurs are an ideal group to test forsuch responses, because they dominated terrestrial ecosystems for more than 155Myr and included thelargest herbivores that ever existed. The radiation of dinosaurs was punctuated by several ecologicallyimportant events, including extinctions at the Triassic/Jurassic (Tr/J) and Jurassic/Cretaceous (J/K)boundaries, the decline of cycadophytes, and the origin of angiosperms, all of which may have hadprofound consequences for herbivore communities. Here we present the first analysis of morphological andbiomechanical disparity for sauropodomorph and ornithischian dinosaurs in order to investigate patterns ofjaw shape and function through time. We find that morphological and biomechanical mandibular disparityare decoupled: mandibular shape disparity follows taxonomic diversity, with a steady increase through theMesozoic. By contrast, biomechanical disparity builds to a peak in the Late Jurassic that corresponds toincreased functional variation among sauropods. The reduction in biomechanical disparity followingthis peak coincides with the J/K extinction, the associated loss of sauropod and stegosaur diversity, and thedecline of cycadophytes. We find no specific correspondence between biomechanical disparity and theproliferation of angiosperms. Continual ecological and functional replacement of pre-existing taxa accountsfor disparity patterns through much of the Cretaceous, with the exception of several unique groups, such aspsittacosaurids that are never replaced in their biomechanical or morphological profiles.

Jamie A. MacLaren. Department of Biology, Universiteit Antwerpen, Campus Drie Eiken, Universiteitsplein,Wilrijk, Antwerp, 2610, Belgium

Philip S. L. Anderson. Department of Animal Biology, University of Illinois at Urbana–Champaign, 515Morrill Hall, 505 S. Goodwin Ave., Urbana, Illinois 61801, U.S.A.

Paul M. Barrett. Department of Earth Sciences, Natural History Museum, London, Cromwell Road, London,SW7 5BD, U.K.

Emily J. Rayfield. * School of Earth Sciences, University of Bristol, 24 Tyndall Avenue, Bristol, BS8 1TQ, U.K.

Accepted: 19 July 2016Published online: 15 December 2016Data available from the Dryad Digital Repository: http://dx.doi.org/10.5061/dryad.c78k5

Introduction

Sauropodomorph and ornithischian dino-saurs were the foremost herbivorous terrestrialvertebrates of the Mesozoic Era in terms ofspecies richness, abundance, and functionaldiversity (Weishampel and Norman 1989; Sereno1999; Weishampel et al. 2004; Barrett 2014).Both groups survived two extinction events—the end-Triassic mass extinction (Tr/J) and asmaller extinction at the Jurassic/Cretaceousboundary (J/K)—and persisted through severalepisodes of floral turnover, including thedecline of cycadophytes and the proliferationof angiosperms (Sereno 1997; Barrett and Willis2001; Lloyd et al. 2008; Butler et al. 2009b).However, relatively few studies have attemptedto quantify the responses of nonavian dinosaursto these extrinsic environmental drivers.

A number of studies have investigated theecological and evolutionary responses ofdinosaurs to the Tr/J mass extinction in termsof diversity analyses, but only a handful ofstudies have quantified morphological dispar-ity (Brusatte et al. 2008a,b) or the evolution ofother traits across this interval (Irmis 2011;Sookias et al. 2012). These studies found thatdinosaur morphospace occupation was notgreatly affected by the Tr/J extinction (Brusatteet al. 2008a,b): dinosaurian disparity remainedessentially unchanged across the Tr/Jboundary,whereas crurotarsans became almostcompletely extinct (Brusatte et al. 2008a). Withrespect to dinosaurs the J/K extinction has beenstudied in terms of diversity analyses (e.g.,Upchurch and Barrett 2005; Barrett et al. 2009;Butler et al. 2010, 2011; Upchurch et al. 2011),and the potential ecological consequences of

Paleobiology, 43(1), 2017, pp. 15–33DOI: 10.1017/pab.2016.31

© 2016 The Paleontological Society.All rights reserved.. This is anOpenAccess article, distributedunder the terms of theCreativeCommons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-se, distribution,and reproduction in any medium, provided the original work is properly cited. 0094-8373/17Downloaded From: https://bioone.org/journals/Paleobiology on 16 Jun 2020

Terms of Use: https://bioone.org/terms-of-use

this event have been discussed qualitatively interms of changes to dinosaur browsing regimesand community composition (Bakker 1978;Barrett and Willis 2001; Barrett and Upchurch2005). Possible associations between paleobota-nical turnovers and dinosaur evolution havebeen proposed (e.g., Bakker 1978; Weishampeland Norman 1989; Tiffney 1992; Mustoe2007), with the suggestion that changes in theprevalent mode of dinosaur herbivory (e.g.,high-browsing vs. low browsing; extensive oralprocessing vs. lack of oral processing) werereciprocally related to changes in the taxonomicand ecological composition of contemporaryplant communities. In particular, it has beensuggested that a decline in sauropodomorphand stegosaur abundance and diversity mightbe associated with a decline in cycadophytediversity during the Early Cretaceous and thatthe ecological radiation of angiosperms duringthe same period may have been fosteredby a coincident taxonomic radiation of low-browsing ornithischian dinosaurswith complexjaw mechanisms (e.g., Bakker 1978; Weishampeland Norman 1989; Tiffney 1992; Mustoe2007). Hypotheses regarding dinosaur–plantcoevolution have been more recently testedquantitatively and qualitatively using spatio-temporal comparisons between the dinosaurand paleobotanical records (Barrett andWillis 2001; Butler et al. 2009a,b, 2010). Thesediversity-based spatiotemporal studies foundno definitive evidence for the coradiation of anyMesozoic plant and dinosaur group, althoughsome temporal correlations were suggestive ofpossible interactions. Physiological limits onsome of these coevolutionary hypotheses havealso been proposed on the basis of the possiblenutritional value of potential food plants (e.g.,Hummel et al. 2008; Gee 2011).

Disparity analyses quantify morphologicaldiversity within a group of organisms, ratherthan merely documenting taxonomic richness(Wills et al. 1994; Ciampaglio et al. 2009).Unlike species-richness estimates, disparityanalyses can be robust to sampling biases anddocument the variation in morphology andpotential function within taxonomic groups(Wills et al. 1994). Assessments of morpholo-gical disparity using either anatomicalmeasurements or cladistic characters have

been conducted on various extinct vertebrategroups, including dinosaurs (Brusatte et al.2008a,b, 2012; Young and Larvan 2010; Butleret al. 2011; Foth and Rauhut 2013; Button et al.2014). By contrast, a new method for assessingthe diversity of biomechanical profiles,multivariate biomechanical disparity (Anderson2009; Anderson et al. 2011, 2013; Stubbs et al.2013), has not been widely applied. Biomecha-nical disparity offers a novel means to quantifyvariation in biomechanically relevant traits andto infer their potential ecological significance:for example, biomechanical traits mightinclude mechanical advantage (the ratio ofmuscle moment arms indicating the efficiencyof force transfer during biting), polar momentof inertia (a proxy for flexural stiffness), andmandibular articulation offset (dictatingsimultaneous occlusion of the entire toothrow, or scissor-like occlusion) (Anderson2009; Anderson et al. 2011, 2013; Stubbs et al.2013). Other studies have explored disparityof individual biomechanical traits such asmechanical advantage (Sakamoto 2010;Brusatte et al. 2012), average maximum stress,or a metric of skull strength (Foth and Rauhut2013). Continuous measurements can beprojected into multivariate “biomechanicalmorphospace.” Previous work in this area hasused two-dimensional (2D) views of mandib-ular elements to investigate the appearanceand diversity of biomechanical profiles duringthe radiation of Paleozoic fishes (Anderson2009; Anderson et al. 2011), the water-to-landtransition in tetrapods (Anderson et al. 2013),theMesozoic diversification of crocodylomorphs(Stubbs et al. 2013), and niche partitioning insauropod dinosaurs (Button et al. 2014).

Despite previous work, the functionalresponses to these potential evolutionarydrivers, and hence how the organism inter-acted with its environment and potentialdrivers of selection, have not been quantified.Without this information we lack a completepicture of how dinosaur communities andclades interacted with and exploited Mesozoicenvironments over time. In addressingthese questions, assessing the morphologicalvariation evident from the fossil record maynot be sufficient, as we do not know whethermorphology and morphological diversity are

16 JAMIE A. MACLAREN ET AL.


reliable predictors of function and functionaldiversity. Therefore, in order to assess therelationship between jaw shape, function, andextrinsic evolutionary drivers, we provide thefirst quantitative assessment of the morpho-logical and biomechanical disparity of anindividual functional unit (the lower jaw) inherbivorous nonavian dinosaurs through time.This approach complements previous attemptsto examine these questions though spatiotem-poral comparisons of species-richness patternsand provides the only rigorous biomechani-cally and functionally based analysis of theseissues attempted to date. We hypothesize thatornithischians and sauropodomorphs willshow distinct morphologies and biomechani-cal profiles (i.e., in both the shape andmechanical capabilities of the jaw). We alsohypothesize that the shift in plant communitystructure after the J/K boundary will triggera corresponding shift in dinosaurian jawbiomechanical profiles, due to the differingphysiognomies, digestibility, and mechanicalproperties of the varied potential food plantclades that were ecologically important atdifferent times throughout the Mesozoic(Bakker 1978; Weishampel 1984; Niklas 1992;Hummel et al. 2008; Gee 2011). We use ageometric morphometric landmark analysis tocompare dinosaur mandibular shape variabil-ity to variation in mandibular biomechanicalprofiles. We then compare these data with thetiming of several extrinsic events (tetrapodextinctions, changes in floral communities) thathave been proposed to influence dinosaurevolutionary history, in order to determinewhether coincident patterns are present.

Materials and Methods

Data for 2D landmark and biomechanical traitanalyses were compiled from 167 sauropodo-morph and ornithischian dinosaur taxa (seeSupplementary Information, Appendix 6). Herbi-vorous nonavian theropods were excluded fromthis data set, as complete mandibular material forthese animals is rare. A mandibular biomechani-cal profile represents a good proxy for character-izing the feeding system, as the mandible isprimarily adapted for feeding, whereas thecranium has multiple functional roles, some of

which are unrelated to feeding, such as housingthe brain and sensory organs (Hylander et al.1991; Hylander and Johnson 1997).

Morphology.—The archosaur mandible is aprimarily planar structure, although itsmorphology does differ between groups, withvarying degrees of inturning and bowing,particularly with respect to its symphysealregion (Romer 1956). However, to include asmany taxa as possible, in order to account forthe greatest amount of biomechanical andmandibular and dental shape variation, weselected a standard lateral view of themandible as the basis for this study. The 2Dlandmarks were applied to homologous andanalogous points on lateral images of dinosaurjaws using tpsDig II software (Rohlf 2004;Zelditch et al. 2012). Six fixed landmarkswere described, identifying biologically andoperationally homologous points on bothsauropodomorph and ornithischian jaws (seeSupplementary Fig.1). The overall morphologyof each jaw was described by a series of slidingsemilandmarks (sLM). Six sLM curves, eachbracketed by two of the fixed landmarks, wereused to define the shape of the jaw. In total, 88landmarks (both fixed and sliding) weredescribed. sLMs were slid using the Chord-d2

technique to minimize Procrustes distancesrather than bending energy (Rohlf 2008); thiswas performed in tpsRelw. Described curveswere appended to landmarks in tpsUtil (Rohlf2004); appended landmarks were thensuperimposed using generalized least-squares(Procrustes) methods in tpsRelw (Rohlf 2008).Procrustes superimposition aligned jaws,eliminating scale, location, and rotationaldifferences between specimens (Rohlf 2004).Consensus models, partial warps, and relativewarps were then calculated using tpsRelwsoftware. Relative warp scores were subjectedto principal components analysis (PCA) toproduce shape-based morphospace plots.

Biomechanics.—Eighteen continuous biome-chanical characters or traits were quantified,many of which have important functionalconsequences in extant organisms (Table 1).Full details of the biomechanical charactersare described in the Supplementary Material.Biomechanical trait measurements werestandardized using a z-transformation

HERBIVOROUS DINOSAUR JAW DISPARITY 17


technique, giving all characters a mean of 0and a variance of 1 (Anderson et al. 2011). Astandardized matrix of biomechanical characterscores was then subjected to principalcoordinates analysis (PCoA), using the Gowermodel to correct for missing data to producebiomechanical morphospace plots. PCoA andcreation of morphospace plots was performedin Past, Version 3 (Hammer et al. 2001).

Significant differences in morphospace occu-pation were tested using nonparametric multi-variate analysis of variance (NPMANOVA) inPast, Version 3 (Hammer et al. 2001). All princi-pal axes accounting for more than 1% of varia-tion were used in the NPMANOVA, resulting in12 axes for shape-based and 15 axes forbiomechanical morphospace. Principal axeswere used to display two types of morphospacecomparisons: overall shape-based and bio-mechanical morphospace between sauropodo-morphs and ornithischians. We also created aseries of morphospace plots representing eight20 Myr time slices. These time slices were con-structed by combining taxa from two adjacent10Myr time bins used for the disparity analyses(see following section). Combining time binsallowed for good sample size and enabled com-parisons across major ecological transitions, forexample, mass extinction events.

Disparity.—Disparity through time wascalculated across sixteen 10Myr time bins. Thelengths of the time bins either side of the Tr/Jboundary were adjusted to accommodate thedate of the boundary as in Butler et al. (2012).Use of 10Myr time bins enables comparisonsacross both the Tr/J and J/K boundaries,standardizes bin length, and provides greatersample sizes per bin than those available forstrict stage-level comparisons. Sauropodomorphdisparity was also analyzed for vertical feedingenvelopes in 3m intervals. Species assignmentto each maximum feeding envelope is listed inthe Supplementary Material. To account forvariation in the published literature, maximumsauropodomorph feeding envelopes weretaken from published works, includingreconstructions from new material (e.g.,Upchurch and Barrett 2000; Apesteguía 2004;Sander et al. 2006; Peyer and Allain 2010;Whitlock 2011; Stevens 2013). Disparityanalyses were carried out using the Morpho-logical Disparity Analysis (MDA) package forMatlab (Navarro 2003). For all disparity tests,two variance-based disparity metrics weretested: the sum of variance and mean pairwisedistance. Both thesemetrics are robust to samplesize variation (Ciampaglio et al. 2009). The sumof variance metric is plotted in the main text.

TABLE 1. Continuous biomechanical characters used in this study.

Code Functional trait Description

C1 Anterior mechanical advantage Ratio of maximum out-lever (on functional tooth row) and jaw musclein-lever moment arms

C2 Posterior mechanical advantage Ratio of minimum out-lever (on functional tooth row) and jaw musclein-lever moment arms

C3 Opening mechanical advantage Ratio of maximum out-lever and opening in-lever moment armsC4 Maximum aspect ratio Proxy for maximum flexural stiffness in the jawC5 Average aspect ratio Proxy for average flexural stiffness across the entire jawC6 Relative adductor fossa length Length of adductor muscle attachment; proxy for jaw muscle sizeC7 Relative dental row length Length of functional tooth row relative to total jaw lengthC8 Relative articular offset Proxy for deviation of biting action from scissor-like mastication.C9 Relative mandibular fenestra Area of mandibular fenestrae relative to total lateral jaw areaC10 Relative dental curvature Curvature of functional tooth row; proxy for shearing vs. compressive

masticationC11 Cheek tooth height:breadth Proxy for maximum tooth size for teeth occluding with maxillary teethC12 Premaxilliary occluding tooth height:

breadthProxy formaximum tooth size for teeth occludingwith premaxillary teeth

C13 Tooth packing Proxy for tooth separation and how closely teeth are packedC14 Predentary tooth procumbancy Proxy for anterior-most tooth procumbancyC15 Tooth height:jaw depth Height of tooth present above deepest section of functional jaw takenC16 Relative symphyseal length Proxy for robustness of anterior jawC17 Mandibular symphysis orientation Proxy for symphyseal resistance to bending during bitingC18 Predentary offset Proxy for predentary curvature in ornithischians



Mean pairwise distance results can be viewedin the Supplementary Material. Data werebootstrapped (1000 replicates), and 95%confidence intervals were calculated andgraphically presented. Significant differencesand likelihood ratios between each time binwere calculated using pairwise t-tests andmarginal-likelihood assessment on sum ofvariance measures (Finarelli and Flynn 2007).A likelihood ratio>8 is considered a likely result(Finarelli and Flynn 2007). Results of t-testsweresubsequently corrected for multiple com-parisons, using Bonferroni corrections whereappropriate (Holm 1979). Results for meanpairwise distance can be found in the Supple-mentary Material.

Results

Shape Morphospace Occupation.—Our resultsdemonstrate that sauropodomorph and

ornithischian jaws occupy significantly differentregions of morphological morphospace (p< 0.01;Fig. 1; Table 2). There is minimal overlapbetween sauropodomorphs and ornithischiansalong PC1, with only seven ornithischian jawmorphologies occupying similar regions tosauropodomorphs. Overlapping ornithischiantaxa represent basal members of theirrespective groups (basal ornithischians:Agilisaurus and Pisanosaurus; thyreophoransEmausaurus and Gigantspinosaurus; and thebasal ceratopsian Yinlong), with the exceptionof Stegosaurus (two species). Regions of overlapare occupied by a wide range of both basaland derived sauropodomorphs; theseinclude: Plateosaurus gracilis, Lamplughsaura,mamenchisaurids, brachiosaurids, and twoSouth American titanosaurids (Antarctosaurusand Bonitasaura). Sauropodomorphs occupymorphospace exclusively in the −PC1 region:this region is characterized by dorsoventrally

FIGURE 1. Patterns of morphospace occupation for herbivorous nonavian ornithischian and sauropodomorphdinosaurs. PC1 and PC2 account for 50.4% of variation. Ornithischian and sauropodomorph taxa occupy significantlydifferent regions of shape-based morphospace (p< 0.05). Filled circles, Sauropodomorpha; open circles, Ornithischia.Silhouettes represent jaw profiles found in that region of morphospace.



narrow jaws and the lack of a prominentcoronoid process. Noneusauropod sauropod-omorphs (e.g., Plateosaurus, Melanorosaurus), forthe most part, account for sauropodomorphoccupation of morphospace in +PC2: thisregion is typified by very narrow anterior jaws.Macronarian and diplodocoid taxa (includingDiplodocus and Tapuiasaurus) primarily occupy−PC2 regions ofmorphospace (Fig. 1). The centerof the morphospace (0.0 PC1; 0.0 PC2) isoccupied by nonhadrosaurid iguanodontians(Parksosaurus, Theiophytalia, and Dryosaurus).Jaws in this region exhibit a greater gapbetween landmarks 1 and 2 than in sauropod-omorphmorphospace (due to the presence of thepredentary in iguanodontians). Disparategroups of nonthyreophoran ornithischiansexpand morphospace occupation into +PC1and +PC2 (hadrosaurids) and −PC2 regions(leptoceratopsids and psittacosaurids). +PC1and +PC2 regions typically contain jaws withprominent coronoid processes and downwardlydeflected predentaries; −PC2 regions containrobust, dorsoventrally broad jaws. Non-ceratopsid marginocephalian jaw morpho-logies, such as those of psittacosaurids andleptoceratopsids, contribute strongly to theexpansion of ornithischian shape morphospace,predominantly into +PC1/−PC2. Taxa areabsent in a region of morphospace around+0.05 PC1/−0.075 PC2.

Biomechanical Morphospace Occupation.—Ourresults demonstrate that sauropodomorph andornithischian taxa also occupy significantlydifferent regions of biomechanical morpho-space (p<0.01; Figs. 2, 3; Table 2). There isgreater overlap in biomechanical morphospaceoccupation than shapemorphospace, with 16–20

ornithischian taxa occupying morphospacethat is shared with sauropodomorphs(Figs. 2, 3). Overlapping ornithischian taxainclude basal ornithischians (Pisanosaurus,heterodontosaurids) and basal members ofThyreophora (Emausaurus, stegosaurs),Marginocephalia (Yinlong), and Ornithopoda(Changchunsaurus, Dysalotosaurus). Sauro-podomorphs occupy regions of +PCo1.Noneusauropod sauropodomorphs (e.g.,Coloradisaurus, Pantydraco) predominate in+PCo1/−PCo2. This region is characterizedby jaws with a high mechanical advantage anda large adductor muscle attachment area.Diplodocids, nonneosauropods, and nontita-nosaurian macronarians (e.g.,Mamenchisaurus,Camarasaurus) stretch sauropodomorphoccupation into +PCo2. Jaws in this regionalso display high mechanical advantages,coupled with high aspect ratios. Manyiguanodontian, ceratopsid, and psittacosauridjaw profiles occupy similar regions of+PCo2 biomechanical morphospace (Fig. 2).Occupation is spread deeper into −PCo1 byleptoceratopsids (e.g., Montanoceratops). Thisregion of functional space is characterized bydeep jaws with short adductor muscleattachment and a high posterior mechanicaladvantage. Expansion into −PCo2 is accountedfor by deep-jawed ankylosaurs (Euoplocephalus,Silvisaurus), with low tooth:jaw depth ratiosand high relative dental length (Fig. 2). Similarpatterns are observed in PCo3, with morebasal sauropodomorphs occupying −PCo3,with a large cluster of iguanodontians andceratopsids occupying regions of centralmorphospace (0.0 PCo1; 0.0 PCo3). Functionalloadings, interpretations for the first fourprincipal axes, and individual speciesplacement in morphospace can be found inthe Supplementary Material.

Morphospace Occupation through Time.—Breakdown of shape and biomechanicalmorphospace into 20Myr time bins highlightspatterns of morphospace occupation by eachclade through time (Figs. 4–6). Initialoccupation during the Late Triassic–MiddleJurassic is dominated by sauropodomorphs,with low numbers of contemporaneous basalornithischians (e.g., heterodontosaurids andthyreophorans). In the bin representing the

TABLE 2. Results of significance testing (NPMANOVA)on morphospace occupation (PC1 and PC2) and bio-mechanical occupation (PCo1 and PCo2; PCo1 and PCo3)between Ornithischia and Sauropodomorpha (at p< 0.05).

Shape–basedmorphospace Sauropodomorpha Ornithischia

Sauropodomorpha — <0.001Ornithischia <0.001 —Biomechanical

morphospaceSauropodomorpha Ornithischia

Sauropodomorpha — <0.001Ornithischia <0.001 —



20 Myr prior to the J/K boundary (165–145 Ma),thyreophorans, ornithopods, marginocepha-lians, and heterodontosaurids all occupysimilar regions of shape morphospace, yet atthis time, the same clades occupy disparateregions of biomechanical morphospace withlittle overlap (Figs. 5, 6; 165–145Ma, Table 3).Sauropodomorphs at this time show signi-ficantly different biomechanical occupationto stegosaurs and ornithopods, but notheterodontosaurids or the basal ceratopsianYinlong (NPMANOVA, p<0.01; Table 3). Thesauropodomorphs are biomechanically diverseprior to the J/K boundary, occupying the regionof morphospace that correlates to high toothheight:base, high mechanical advantages, andlarge mandibular fenestrae. After the J/Kboundary, morphospace and biomechanicalmorphospace plots show a drop in sauropo-domorph morphological and biomechanicalvariation as sample size diminishes and

expansion in disparity by marginocephaliansand, later, ornithopods (Figs. 4–6, 145–65Ma).By the Early Cretaceous, the surviving Jurassicherbivorous dinosaur clades (sauropodo-morphs, marginocephalians, ornithopods, andthyreophorans) are statistically distinct in bothshape and biomechanical morphospace(Table 2). Sauropodomorphs display substan-tially reduced variation, whereas ankylosaurs,ceratopsians, and ornithopods expand intohitherto unoccupied regions of biomechanicalmorphospace. Marginocephalians (e.g.,Psittacosaurus) share areas of biomechanicalmorphospace with iguanodontians but occupyvery different regions of shape space (Fig. 4,145–105Ma).

In the latest Cretaceous, the four cladespresent occupy distinct regions of shape mor-phospace (p< 0.01; Table 2), with the exceptionof onemarginocephalian taxon (Stegoceras) thatplots between nonhadrosaurid ornithopods

FIGURE 2. Patterns of biomechanical morphospace occupation for herbivorous nonavian ornithischian andsauropodomorph dinosaurs. PCo1 and PCo2 account for 25.2% of variation. Ornithischian and sauropodomorph taxaoccupy significantly different regions of biomechanical morphospace (p< 0.05). Filled circles, Sauropodomorpha; opencircles, Ornithischia. Silhouettes represent jaw biomechanical profiles found in that region of biomechanicalmorphospace.



and ankylosaurians (Fig. 4, 85–65Ma). Bio-mechanically, Stegoceras is nested amongornithopods and is closer to sauropods thanmany contemporaneous ceratopsians. Corre-sponding biomechanical morphospace plotsshow a very different trend. Margin-ocephalians overlap with both ornithopodsand thyreophorans. Thyreophorans andornithopods do not overlap, and saur-opodomorphs overlap minimally with orni-thopods (Figs. 5, 6, 85–65Ma). Whereasvariation in marginocephalian jaw shapeand biomechanics increases throughout theCretaceous, ornithopod shape and biomechanicalvariation remains constant throughout the LateCretaceous. Leptoceratopsids (e.g., Udanocera-tops, Montanoceratops) extend biomechanicalmorphospace occupation into the region ofmorphospace characterized by deepmandibleswith short adductor muscle attachmentand high posterior mechanical advantages

(Figs. 5, 6). Full details of the biomechanicalcharacter loadings are described in theSupplementary Appendix 5.

Disparity.—Morphological (shape) andbiomechanical disparity measures aredecoupled through the Mesozoic (Fig. 7).Morphological disparity primarily trackssample diversity (Fig. 7A): it does notfluctuate greatly through the first 80Myr ofdinosaur evolution, begins to increase from theMiddle Jurassic onward, and reaches a peak inthe Late Cretaceous (Fig. 7A). There are nosignificant differences in disparity betweentime bins (p>0.05). By contrast, biomechanicaldisparity undulates through the Mesozoic(Fig. 7B), a decoupling from sample diversityand morphological diversity. Several smallpeaks and troughs (for example the peak in theLate Jurassic) correspond to increased samplesize (Fig. 7B, diamond data points): however,time periods with greatest sample sizes do not

FIGURE 3. Patterns of biomechanical morphospace occupation for herbivorous nonavian ornithischian andsauropodomorph dinosaurs. PCo1 and PCo3 account for 23.9% of variation. Ornithischian and sauropodomorph taxaoccupy significantly different regions of biomechanical morphospace (p< 0.05). Filled circles, Sauropodomorpha; emptycircles, Ornithischia. Silhouettes represent jaw biomechanical profiles found in that region of biomechanicalmorphospace.



correspond to peaks in biomechanical disparity(during the latest Cretaceous, for example). Thepeak in the latest Jurassic also corresponds withthe presence of high-browsing sauropodo-morphs (>9m), which display a higher degreeof biomechanical disparity than some lower-browsing forms (p> 0.05; see SupplementaryFig. 10). There are no significant differencesin disparity between successive time binsfor either biomechanical or morphologicaldisparity curves (at p= 0.05) and no marginal-likelihood values exceed the threshold value

of 8. There are a few instances where disparitydiverges markedly from sample size,suggesting that a trend, albeit nonsignificant,might be observed. For example, morphologicaldisparity rises in the Early Cretaceous,immediately after the J/K extinction, and inthe early Late Cretaceous, while sample sizedrops. Likewise, biomechanical disparity dropsin the Middle Jurassic while sample size risesslightly. Conversely, in the latest Cretaceous,sample size rises sharply while biomechanicaldisparity drops very slightly.

TABLE 3. NPMANOVA significance testing between clade occupations of biomechanical morphospace throughtime. Bold p-values represent significant differences (at p< 0.05). SA, Sauropodomorpha; BO, Basal Ornithischia;TH, Thyreophora; OR, Ornithopoda; MA, Marginocephalia.

Time bin NPMANOVA p-values

Clades SA BO

225–202Ma SA — 0.114BO 0.114 —

Clades SA BO

202–185Ma SA — 0.009BO 0.009 —

Clades SA BO TH

185–165Ma SA — 0.142 1BO 0.142 — 1TH 1 1 —

Clades SA BO TH OR MA

165–145Ma SA — 0.505 0.009 0.015 1BO 0.505 — 0.520 0.124 1TH 0.009 0.520 — 0.158 1OR 0.015 0.124 0.158 — 1

Clades SA OR MA

145–125Ma SA — 0.084 0.003OR 0.084 — 0.016MA 0.003 0.016 —

Clades SA TH OR MA

125–105Ma SA — 0.186 <0.001 <0.001TH 0.186 — 0.003 0.007OR <0.001 0.003 — <0.001MA <0.001 0.007 <0.001 —

Clades SA TH OR MA

105–85Ma SA — 0.164 0.002 0.043TH 0.164 — 0.005 0.037OR 0.002 0.005 — <0.001MA 0.043 0.037 <0.001 —

Clades SA TH OR MA

85–65Ma SA — <0.001 <0.001 <0.001TH <0.001 — <0.001 <0.001OR <0.001 <0.001 — <0.001MA <0.001 <0.001 <0.001 —



Discussion

Impact of Extinction on Herbivorous DinosaurDisparity.—Our results from both morpho-logical and biomechanical disparity curvessupport conclusions from previous studiesexamining dinosaur disparity aroundextinction events (Brusatte et al. 2008a, 2012).Morphological disparity across the Tr/Jboundary increases slightly, likely triggeredby the addition of heterodontosaurid jawprofilesto the morphospace (Fig. 7B). Biomechanicaldisparity decreases from an initial peak in the

Carnian (225 Ma) to the Tr/J boundary, acrosswhich there is a further nonsignificant decrease(Fig. 7B). The placement of taxa in biomechanicalmorphospace suggests that both ornithischianand sauropodomorph taxa share similarbiomechanical profiles immediately before andafter the Tr/J boundary (Figs. 5, 6). By contrast,the transition across the J/K boundary shows adecoupled relationship between biomechanicaland morphological disparity (Fig. 7). Morpho-logical disparity after the J/K boundary increasessharply: this pattern can be attributed to thepresence of novel jaw morphologies such

FIGURE 4. Patterns of morphospace occupation for herbivorous nonavian dinosaurs through the Mesozoic (20Myr timebins), based on PC1 and PC2 (accounting for 50.4% of variation). Sauropodomorpha occupy isolated regions ofmorphospace for the majority of the Mesozoic, with overlap between North American sauropods and thyreophoransbetween 185 and 145Ma.



as those of psittacosaurids and earlyhadrosauroids in combination with those ofnew sauropod clades (Fig. 4, 145–125Ma).It should be noted that this disparity increaseis nonsignificant, likely due to the low taxoncount (n=5). The lack of many dinosaur-bearingformations between the Berriasian and Albianmay partially account for the low speciesrichness observed in this interval, although itcould also be attributed to the J/K extinctionevent (Barrett et al. 2009; Upchurch et al. 2011).Nevertheless, shape variation at this time doesnot track sample diversity. Biomechanical

disparity shows a decrease across the J/Kboundary (Fig. 7B). The majority of thebiomechanical profiles exhibited prior to theJ/K boundary do not persist into the earliestCretaceous (Figs. 5, 6, 145–125Ma), which isconsistent with the fundamental faunal turnoverthat takes place and the proliferation ofmarginocephalian and ornithopod taxa (e.g.,Bakker 1978; Weishampel and Norman 1989;Barrett and Willis 2001). Finally, our resultsconcur with disparity patterns observed in thelatest Cretaceous leading to the Cretaceous/Paleogene (K/Pg)mass extinction (Brusatte et al.

FIGURE 5. Patterns of biomechanical morphospace occupation for herbivorous nonavian dinosaurs through theMesozoic (20Myr time bins), based on PCo1 and PCo2 (accounting for 25.2% of variation). Sauropodomorphspredominantly overlap only with heterodontosaurids (202–145Ma). Aptian–Maastrichtian marginocephalians andornithopods occupy similar regions of morphospace (125–65Ma).



2012): both morphological and biomechanicalcurves show a decrease in disparity from theCampanian to the Maastrichtian, despite anotable increase in sample size.

Patterns of Morphospace Occupation.—Discretemorphospace occupation suggests that, whenconsidered as a single data set, the jaws ofsauropodomorphs and ornithischians aredifferent in both shape and in jaw bio-mechanics (Figs. 1–3). Individual occupation ofmorphospace by each taxon is graphicallyrepresented in Supplementary Figures 2–6.

Limited overlap between these clades suggestslittle competition between ornithischians andsauropodomorphs in feeding function,particularly during the latter part of theMesozoic (see also Barrett and Upchurch 2005).However, where overlap does occur, it tends tobe between the basal members of variousornithischian clades (e.g., heterodontosaurids,basal thyreophorans, and basal ceratopsians)and sauropodomorphs. This suggests that earlyornithischians adopted similar morphologicaland mechanical attributes to their feeding

FIGURE 6. Patterns of biomechanical morphospace occupation for herbivorous nonavian dinosaurs through theMesozoic (20Myr time bins), based on PCo1 and PCo3 (accounting for 23.9% of variation). Sauropodomorphs overlapvery little with contemporaneous taxa before the latest Cretaceous (85–65Ma). Albian–Maastrichtianmarginocephalians and thyreophorans occupy similar regions of biomechanical morphospace (105–65Ma).



FIGURE 7. Comparison of shape-based and biomechanical disparity curves across 10Myr time bins based on sum ofvariance metric. (A) shape-based disparity; (B) biomechanical disparity. Morphological and biomechanical disparitycurves are decoupled, with morphological disparity increasing through the Mesozoic and biomechanical disparitypeaking in the latest Jurassic. Shaded region spans the 95% confidence intervals based on 1000 bootstrap replicates.Disparity (dots) is plotted alongside jaw specimen sample size curve (diamonds). Flower represents earliest fossilangiosperms (Sun et al. 2002; Du and Wang 2015).



apparatus as macronarian sauropodomorphs(Supplementary Fig. 2, a–c). Later groups ofornithischians radiated into distinct areas ofmorphospace (Figs. 4–6). Breakdown of mor-phological and biomechanical morphospaceinto 20 Myr time bins shows that earliersauropodomorphs are, in general, replaced intheir biomechanical profiles by later sauropo-domorphs through the Jurassic and Cretaceous(Figs. 4–6). Sauropodomorph morphospaceoccupation shows a degree of migrationthrough time, with basal sauropodomorphsoccupying different regions of morphospace toJurassic andCretaceous neosauropods (Figs. 4–6,filled circles). Some later sauropods showconvergence in biomechanical profile withother, earlier forms. For example, the macronarian Camarasaurus occupies very similarregions of morphospace to the earlier diverg-ing eusauropod Datousaurus (SupplementaryFig. 2a–c), despite the former existing around10Myr earlier: this pattern supports the results ofanother recent quantitative craniodental study(Button et al. 2014). Similarly, the titanosauridAntarctosaurus occupies biomechanical morpho-space almost identical to the basal macronarianAbrosaurus (Supplementary Fig. 2a–c). Perhapssurprisingly, we find minimal convergentoccupation in biomechanical morphospacebetween titanosaurids (e.g., Antarctosaurus) anddiplodocids (e.g., Diplodocus) (SupplementaryFig. 2a–c: see also Button et al. 2014). Thispattern is in contrast to shape-based morpho-space (this study), in which these groups occupysimilar regions of morphospace (Fig. 1 andSupplementary Fig. 2). Both shape-based andbiomechanical morphospace patterns showextensive overlap between phylogeneticallyseparate groups of sauropodomorphs. Withinthe sauropods, brachiosaurids are found to bebiomechanically intermediate between basalmacronarian sauropods with short snouts andclosely packed tooth rows (such asCamarasaurus)and titanosaurids with longer snouts andpencil-like teeth (such as Antarctosaurus), anddiplodocids are outliers in this biomechanicalmorphospace. This pattern supports quantitativework on sauropodomorph cranial morphologyrelated to feeding, with similar placement of thesame taxa in cranial (Button et al. 2014) andmandibular morphospace (this study). Late

Jurassic sauropods such as Camarasaurus showsome morphological overlap in mandibularshape with stegosaurs. By contrast, these sameclades show minimal overlap in biomechanicalmorphospace: only Gigantspinosaurus(Stegosauria) and Manidens (Heterodontosau-ridae) share occupation of Late Jurassicsauropodomorph biomechanical morphospace(Supplementary Figs. 3,b–c, 4,b–c). This suggeststhat mandibles with similar gross morphologywere biomechanically and functionallydifferentiated by this time. In general, sauropo-domorphs and heterodontosaurids occupysimilar regions of both shape-based andbiomechanical morphospace and do not extendtheir occupation ofmorphospace beyond regionsalready occupied by the end of the Early Jurassic(Figs. 4–6). From the Middle Jurassic onward,there is slight expansion of morphospacealong PC1 by diplodocoid sauropodomorphsand Jurassic ornithopods (e.g., Camptosaurus),which is also reflected in the morphologicaldisparity curve (Fig. 4, 165–145 Ma; Fig. 7A).Morphological disparity shows an increase fromthe latest Jurassic through the Cretaceous withthe evolution of new groups of ornithischiandinosaurs, particularly marginocephalians.

Early Cretaceous marginocephalians (psit-tacosaurids, Archaeoceratops, and Liaoceratops)occupy novel regions of morphological andbiomechanical morphospace: these taxa shareregions of biomechanical morphospace withhadrosauroids until the disappearance of basalmarginocephalians prior to the last 20Myr ofthe Mesozoic (Figs. 4–6, 85–65Ma). Regions ofbiomechanical morphospace formerly occu-pied by psittacosaurids were then occupiedexclusively by derived hadrosaurids andankylosaurs (Figs. 5, 6, 85–65 Ma). However,the morphological profile of psittacosauridswas never replaced. The latest Cretaceous seesan expansion of biomechanical and shape-based morphospace by two distinct groups ofmarginocephalians: ceratopsids (e.g., Tricera-tops) and leptoceratopsids (e.g., Udanoceratops).The biomechanical profiles of ceratopsidsshow no overlap with those of hadrosaurids.This supports the conclusions of Mallon andAnderson (2013) who, in their study of herbi-vores from the Dinosaur Park Formation(Campanian), found that contemporaneous



hadrosaurids, ankylosaurs, and ceratopsidsoccupied different feeding niches based upondiffering cranial and mandibular mechanicsand morphologies. This study also supportsprevious conclusions on niche partitioningbetween hadrosaurs and ceratopsids (Mallonand Anderson 2013). However, this study alsofound that the majority of derived ceratopsidsplot in similar regions of biomechanical mor-phospace to contemporaneous ankylosaurs, incontrast to the conclusions of Mallon andAnderson (2013). In addition, Asian ankylo-saurs show biomechanical morphospace occu-pation more similar to leptoceratopsids than toceratopsids or North American ankylosaurs.It should be noted, however, that neitherleptoceratopsids nor Asian ankylosaurs wereincluded in Mallon and Anderson (2013),which focused solely on the Dinosaur ParkFormation fauna. Leptoceratopsids expandinto regions of shape-based and biomechanicalmorphospace that had no previous occupants:their extreme mandibular morphologiesaccount for the peak in morphologicaldisparity in the latest Cretaceous (Fig. 7A).Contemporaneous taxa include ceratopsidsand ankylosaurs that have similar biomecha-nical profiles to each other (see above). Thisbiomechanical similarity would cause disparityto be low: however, the inclusion of the highlydisparate leptoceratopsids (in addition tohadrosaurids and the rhabdodontid Zalmoxes)leads to an increase in biomechanical disparitylevels from the early Late Cretaceous. Margin-ocephalian, ornithopod, and thyreophoran bio-mechanical morphospace occupation in thelatest Cretaceous suggests that these groups,while varying from each other in mandibularshape, also share a variety of functional andbiomechanical traits relating to feeding. LateCretaceous hadrosaurids and ankylosauridsfilled the biomechanical roles vacated by EarlyCretaceous nonhadrosaurid iguanodontiansand nodosaurids, respectively. Individualoccupation of morphospace by each taxon canbe viewed in Supplementary Figures 2–6.Dinosaur–Plant Coevolution.—Changes in

dinosaur communities and feeding regimesduring the Late Jurassic–Early Cretaceousinterval have been linked to several majorfloristic changes (decline of cycadophytes,

gymnosperms, and pteridophytes; rise ofangiosperms to ecological dominance) (e.g.,Weishampel and Norman 1989; Tiffney 1992;Mustoe 2007). Our results provide quantitativeevidence that the mandibles of sauropodo-morphs and ornithischians evolved differentmorphologies and biomechanical profiles,potentially enabling them to feed on differentplants in different ways. Moreover, theirminimal overlap in biomechanical morpho-space suggests that there was limitedcompetition between ornithischians and sauro-podomorphs when feeding (see also Barrettand Upchurch 2005). Our data demonstratethat there was no significant increase in thebiomechanical disparity of the feeding apparatusof either major herbivorous dinosaur cladethat was coincident with the proliferation ofangiosperms (Fig. 7). Nevertheless, althoughthis novel food source appears to have hadno discernible impact on the mandibularbiomechanical morphospace occupation ofherbivorous dinosaurs, patterns of morphologicaldisparity do show a marked increase coincidentwith the later Cretaceous proliferation ofangiosperms. This coincident increase is notinterpreted as indication of direct causality,but reflects the appearance of the highlydisparate ankylosaurid and leptoceratopsianjaw morphotypes.

Potential links to cycadophyte declinethrough the Late Jurassic–Early Cretaceous areless clear. The Early Cretaceous decline incycadophytes occurred at a time of majorfaunal change affecting dinosaur clades, butprevious analyses of dinosaur and plant dis-tribution have shown that few of the observedchanges in dinosaur faunas could be linkeddirectly with cycadophyte decline (Butler et al.2009b). Although reduced biomechanicalmandibular disparity across the J/K boundarydoes coincide with the onset of this event,direct evidence of dinosaur herbivory oncycads is sparse (Hummel et al. 2008; Butleret al. 2009b; Gee 2011), and other causes relat-ing to the poorly understood J/K extinctionmay also be involved (Butler et al. 2011;Upchurch et al. 2011). In addition, morpholo-gical disparity after this extinction event showsa notable increase, with different clades ofdinosaurs diversifying into new, unexplored



regions of mandibular morphospace (e.g.,psittacosaurids, early titanosaurs). Resultsfrom this study do not support a coevolu-tionary relationship between herbivorousdinosaur mandibular disparity and angios-perm proliferation and show a similarlynegative relationship to the decline of cycado-phytes. Rather, patterns of mandibular shapeand mechanical diversity seem to be mostgreatly affected by the extinction and emer-gence of different dinosaurian clades.

Sampling Issues.—When disparity trackssample diversity closely, as it does in thisstudy for shape-based disparity, sampling biascannot be ruled out. Morphological disparityin this study partly tracks jaw sample size,suggesting a potential bias in the data set forsome features of the disparity curve (e.g., highsample and disparity in latest Cretaceous;Fig. 7A). The use of the sum of variancedisparity measure and bootstrapping the datahas accounted for sample size as best as ispossible for the data set (Foote 1992, 1994;Ciampaglio et al. 2009) (Fig. 7A). Peaks of highshape disparity in the earliest Cretaceous andearly Late Cretaceous do not correlate withpeaks in sample size. Biomechanical variationdisplays a different trend, demonstrating adecoupling of morphological and biomechanicaldiversity through time. A peak in biomechanicaldisparity in the Late Jurassic is coincident withan increase in jaw sample size, but alsocorresponds to the evolution of high-browsing(>9m) sauropods (e.g., Upchurch and Barrett2000). In addition, many of the sauropod taxa inthis time slice are recovered from the MorrisonFormation of thewesternUnited States (n=6 outof a total of 14 sauropods). The exclusion of theMorrison taxa removes the Late Jurassic peak inbiomechanical disparity (Supplementary Fig. 8i).A similar jackknifing of the taxa from theDashanpu Formation (including the “Upperand Lower Shaximiao” formations) yielded atrough in disparity in the Middle Jurassic butretained a strong peak in the latest Jurassic(Supplementary Fig. 8ii). These results suggestthat the data may be sensitive to the inclusion orexclusion of particularly rich fossil-bearing sites.In addition, the lack of available jaw materialfrom North and South American titanosaursseriously underrepresents sauropodomorph

diversity in the Cretaceous. The addition oftitanosaurid taxa to the analysis may increaseboth the disparity and overall morphospaceoccupation of sauropodomorphs, although thetitanosaur jaws sampled in this study alreadyaccount for a broad range of morphologies(Supplementary Fig. a–c, taxon 37–44).

Supplementary analyses of biomechanicaland shape-based disparity within sauropodo-morphs in relation to maximum feeding heightshow higher levels of disparity in high-browsingsauropods (>9m; e.g., Brachiosaurus, Mamenchi-saurus) when compared with mid-browsing taxa(6–9m; e.g., Camarasaurus), and almost equalin disparity to very low-browsing sauropo-domorphs (0–3m; e.g., Pantydraco, Riojasaurus)(see Supplementary Fig. 10). This pattern con-trasts with sample diversity, with the lowestsample size found in the high-browsing feedingenvelope (n = 6) (Supplementary Fig. 10).Unfortunately, low sample sizes within eachfeeding level prevent any significant differencesor definitive conclusions to be made. However,this pattern remains intriguing and the addi-tion of more mandibular remains from high-and mid-browsing taxa to our sample (as andwhen they are discovered) would complementthis study. This is an avenue of study thatrequires more investigation in the future toenable deeper insights into niche partitioningbetween sauropod groups based on maximumbrowse height.

Relatively few Early Cretaceous sauro-podomorph, thyreophoran, or marginocephaliantaxa possess well-preserved mandibular mate-rial (see list of taxa in SupplementaryMaterial).The dip in biomechanical disparity after theJ/K recovered by our analyses may, therefore,be an artefact due to either geological biases oruneven collection effort, underrepresenting thetrue diversity of jaw biomechanical profiles atthis time. Due to the lack of complete mand-ibles from rebbachisaurids, dicraeosaurids,and other clades, it is possible that the latestJurassic and earliest Cretaceous disparitylevels reported herein are currently under-sampling the total diversity of mandiblemorphology and potential function. Suchexclusions cannot be corrected for by our ana-lyses and represent a limitation of the fossilmaterial currently available.



Conclusions

For the first time, we have quantified themorphological and biomechanical variation ofornithischian and sauropodomorph jawsthroughout the Mesozoic and examined howdiversity related to external extrinsic driverssuch as extinction events and the rise ofangiosperms. We find that herbivorous dino-saur clades have jaws that occupy differentregions of morphospace throughout theMesozoic. Furthermore, sauropodomorphs andornithischians have jaws that also function inbroadly different ways, yet there is somepotentially convergent overlap in biomechanicalfunction between different ornithischian cladesin the Cretaceous. Basal members of each cladetend to be more similar in form and function toeach other, while derived taxa are more func-tionally and morphologically divergent. Herbi-vorous dinosaur jaws maintained a numericallysteady diversity of biomechanical traits, with apeak observed in the Late Jurassic triggered bythe diversification of high-browsing sauropods.This is consistent with a rapid evolutionaryradiation in biomechanical diversity amongherbivorous dinosaurs followed by a plateau.The Tr/J extinction had no overall effect onbiomechanical variation among herbivorousdinosaurs, despite fundamental changes infloral and faunal composition across the bound-ary. This consistency suggests that Early Jurassicdinosaurs filled the functional feeding nichesvacated by the extinction of Late Triassic taxa.Similar successive replacement patterns are alsoseen in Devonian gnathostomes and Devo-nian to mid-Pennsylvanian tetrapodomorphs(Anderson et al. 2011, 2013). Biomechanicaldisparity across the J/K boundary suggests thatlarge-scale faunal turnover at this time did affectmandibular disparity, which did not recover topre-J/K disparity levels through the Cretaceous(Fig. 7). A diverse fauna of high-browsingsauropods did not persist into the Early Cretac-eous, and the sauropodomorph contribution tooverall disparity wanes through the Cretaceous,despite a later increase in their Late Cretaceousspecies richness. The highly specializedpsittacosaurids were not replaced in theirbiomechanical profile. However, their role as abiomechanically disparate group in Asia is later

filled by Late Cretaceous leptoceratopsids (e.g.,Udanoceratops), a group that is also present inNorth America. Late Cretaceous hadrosauridsand ankylosaurids filled the biomechanical rolesvacated by Early Cretaceous nonhadrosauridiguanodontians and nodosaurids respectively.Our results imply that, after the establishment ofpeak overall biomechanical variation in thelatest Jurassic, only marginocephalians demon-strated widespread variation in biomechanicalprofiles over time, triggered by the isolatedadaptive radiations of psittacosaurids andleptoceratopsians. The remainder of Cretaceousherbivorous dinosaurs underwent progressiveniche replacement, with successive replacementby related taxa with comparable biomechanicalprofiles.

Acknowledgments

The authors would like to thank D. Buttonand T. Stubbs (both University of Bristol) fortaxonomic and methodological discussion andS. Chapman (NHM) for specimen provision.Funding support was received from the BobSavage Fund of University of Bristol and froma BBSRC grant BB/I011668/1 awarded toE.J.R. Helpful reviews of this contribution werereceived from J. Whitlock and an anonymousreferee. P.S.L.A., E.J.R., and P.M.B. conceivedand developed the project idea. J.A.M. col-lected the data and performed the analyses.J.A.M., along with P.S.L.A., E.J.R., and P.M.B.interpreted the data. All authors contributed towriting the manuscript. J.A.M. designed andprepared the figures.

Literature CitedAnderson, P. S. L. 2009. Biomechanics, functional patterns, anddisparity in Late Devonian arthrodires. Paleobiology 35:321–342.

Anderson, P. S. L., M. Friedman, M. D. Brazeau, and E. J. Rayfield.2011. Initial radiation of jaws demonstrated stability despitefaunal and environmental change. Nature 476:206–209.

Anderson, P. S. L., M. Friedman, and M. Ruta. 2013. Late to thetable: diversification of tetrapod mandibular biomechanicslagged behind the evolution of terrestriality. Integrative andComparative Biology 53:197–208.

Apesteguía, S. 2004. Bonitasaura salgadoi gen. et sp. nov.: a beakedsauropod from the Late Cretaceous of Patagonia. Nat-urwissenschaften 91:493–497.

Bakker, R. T. 1978. Dinosaur feeding behaviour and the origin offlowering plants. Nature 274:661–663.



Barrett, P. M. 2014. Paleobiology of herbivorous dinosaurs. AnnualReview of Earth and Planetary Sciences 42:207–230.

Barrett, P. M., and K. J. Willis. 2001. Did dinosaurs invent flowers?Dinosaur–angiosperm coevolution revisited. Biological Reviews76:411–447.

Barrett, P. M., and P. M. Upchurch. 2005. Sauropodomorph diver-sity through time. Pp. 125–128 in K. A. Curry Rogers, andJ. A. Wilson, eds. The sauropods: evolution and paleobiology.University of California Press, Berkeley.

Barrett, P. M., A. J. McGowan, and V. Page. 2009. Dinosaur diver-sity and the rock record. Proceedings of the Royal Society B276:2667–2674.

Brusatte, S. L., M. J. Benton, M. Ruta, and G. T. Lloyd. 2008a.The first 50 Myr of dinosaur evolution: macroevolutionarypattern and morphological disparity. Biology Letters 4:733–736.

——. 2008b. Superiority, competition, and opportunism in theevolutionary radiation of dinosaurs. Science 321:1485–1488.

Brusatte, S. L., R. J. Butler, A. Prieto-Márquez, and M. A. Norell.2012. Dinosaur morphological diversity and the end-Cretaceousextinction. Nature Communications 3:804.

Butler, R. J., P. M. Barrett, P. Kenrick, and M. G. Penn. 2009a.Diversity patterns amongst herbivorous dinosaurs and plantsduring the Cretaceous: implications for hypotheses of dinosaur/angiosperm co-evolution. Journal of Evolutionary Biology 22:446–459.

——. 2009b. Testing co-evolutionary hypotheses over geologicaltimescales: interactions between Mesozoic non-avian dinosaursand cycads. Biological Reviews 84:73–89.

Butler, R. J., P. M. Barrett, M. G. Penn, and P. Kenrick. 2010. Testingcoevolutionary hypotheses over geological timescales: interac-tions between Cretaceous dinosaurs and plants. BiologicalJournal of the Linnean Society 100:1–15.

Butler, R. J., R. B. J. Benson, M. T. Carrano, P. D. Mannion, andP. Upchurch. 2011. Sea level, dinosaur diversity and samplingbiases: investigating the “common cause” hypothesis in theterrestrial realm. Proceedings of the Royal Society B 278:1165–1170.

Butler, R. J., S. L. Brusatte, B. Andres, and R. B. J. Benson. 2012. Howdo geological sampling biases affect studies of morphologicalevolution in deep time? A case study of pterosaur (Reptilia:Archosauria) disparity. Evolution 66:147–162.

Button, D. J., E. J. Rayfield, and P. M. Barrett. 2014. Cranialbiomechanics underpins high sauropod diversity in resource-poor environments. Proceedings of the Royal Society B 281:20142114.

Ciampaglio, C. N., M. Kemp, and D. W. McShea. 2009. Detectingchanges in morphospace occupation patterns in the fossilrecord: characterization and analysis of measures of disparity.Paleobiology 27:695–715.

Du, Z.-Y., and Q.-F. Wang. 2015. Phylogenetic tree of vascularplants reveals the origins of aquatic angiosperms. Journal ofSystematics and Evolution 54:342–348.

Finarelli, J. A., and J. J. Flynn. 2007. The evolution of encephaliza-tion in caniform carnivorans. Evolution 61:1758–1772.

Foote, M. 1992. Rarefaction analysis of morphological andtaxonomic diversity. Paleobiology 18:1–16.

——. 1994. Morphological disparity in Ordovician–DevonianCrinoids and the Early Saturation of Morphological Space.Paleobiology 20:320–344.

Foth, C., and O. W. M. Rauhut. 2013. The good, the bad, and theugly: the influence of skull reconstructions and intraspecificvariability in studies of cranial morphometrics in theropods andbasal saurischians. PLoS One 8:e72007.

Gee, C. T. 2011. Dietary options for the sauropod dinosaursfrom an integrated botanical and paleobotanical perspective.Pp. 34–56 in N. Klein, K. Remes, C. T. Gee, and P. M. Sander, eds.

Biology of the sauropod dinosaurs: understanding the life ofgiants. Indiana University Press, Indianapolis.

Hammer, Ø., D. A. T. Harper, and P. D. Ryan. 2001. Paleontologicalstatistics software package for education and data analysis.Palaeontologia Electronica 4:9.

Holm, S. 1979. A simple sequentially rejective multiple test proce-dure. Scandinavian Journal of Statistics 6:65–70.

Hummel, J., C. T. Gee, K.-H. Südekum, P.M. Sander, G. Nogge, andM. Clauss. 2008. In vitro digestibility of fern and gymnospermfoliage: implications for sauropod feeding ecology and dietselection. Proceedings of the Royal Society B 275:1015–1021.

Hylander, W. L., and K. R. Johnson. 1997. In vivo bone strain patternsin the zygomatic arch of macaques and the significance of thesepatterns for functional interpretations of craniofacial form. Amer-ican Journal of Physical Anthropology 102:203–232.

Hylander, W. L., P. G. Picq, and K. R. Johnson. 1991. Masticatory-stress hypotheses and the supraorbital region of primates.American Journal of Physical Anthropology 86:1–36.

Irmis, R. B. 2011. Evaluating hypotheses for the early diversificationof dinosaurs. Earth and Environmental Science Transactions ofthe Royal Society of Edinburgh 101:397–426.

Lloyd, G. T., K. E. Davis, D. Pisani, J. E. Tarver,M. Ruta,M. Sakamoto,D. W. E. Hone, R. Jennings, and M. J. Benton. 2008. Dinosaurs andthe Cretaceous terrestrial revolution. Proceedings of the RoyalSociety B 275:2483–2490.

Mallon, J. C., and J. S. Anderson. 2013. Skull ecomorphology ofmegaherbivorous dinosaurs from the dinosaur park formation(upper Campanian) of Alberta, Canada. PloS ONE 8:e67182.

Mustoe, G. 2007. Coevolution of cycads and dinosaurs. CycadNewsletter 30:6–9.

Navarro, N. 2003.MDA: aMATLAB-based program formorphospace-disparity analysis. Computers and Geosciences 29:655–664.

Niklas, K. J. 1992. Plant biomechanics. University of Chicago Press,Chicago.

Peyer, K., and R. Allain. 2010. A reconstruction of Tazoudasaurusnaimi (Dinosauria, Sauropoda) from the late Early Jurassic ofMorocco. Historical Biology 22:134–141.

Rohlf, F. J. 2004. tpsDig. Department of Ecology and Evolution,State University of New York, Stony Brook, N.Y.

——. 2008. tpsRelW. Department of Ecology and Evolution, StateUniversity of New York, Stony Brook, N.Y.

Romer, A. S. 1956. Osteology of the reptiles. University of ChicagoPress, Chicago.

Sakamoto, M. 2010. Jaw biomechanics and the evolution of bitingperformance in theropod dinosaurs. Proceedings of the RoyalSociety of London B 277:3327–3333.

Sander, P. M., O. Mateus, T. Laven, and N. Knótschke. 2006. Bonehistology indicates insular dwarfism in a new Late Jurassicsauropod dinosaur. Nature 441:739–741.

Sereno, P. C. 1997. The origin and evolution of dinosaurs. AnnualReview of Earth and Planetary Sciences 25:435–489.

——. 1999. The evolution of dinosaurs. Science 284:2137–2147.Sookias, R. B., R. J. Butler, and R. B. J. Benson. 2012. Rise ofdinosaurs reveals major body-size transitions are driven bypassive processes of trait evolution. Proceedings of the RoyalSociety B 279:2180–2187.

Stevens, K. A. 2013. The articulation of sauropod necks:methodology and mythology. PloS ONE 8:e78572.

Stubbs, T. L., S. E. Pierce, E. J. Rayfield, and P. S. L. Anderson. 2013.Morphological and biomechanical disparity of crocodile-linearchosaurs following the end-Triassic extinction. Proceedings ofthe Royal Society B 280:20131940.

Sun, G., Q. Ji, D. L. Dilcher, S. Zheng, K. C. Nixon, and X. Wang.2002. Archaefructaceae, a new basal angiosperm family. Science296:899–904.

Tiffney, B. H. 1992. The role of vertebrate herbivory in the evolutionof land plants. Palaeobotanist 41:87–97.



Upchurch, P. M., and P. M. Barrett. 2000. The evolution ofsauropod feeding mechanisms. Pp. 79–122. in H.-D. Sues,ed. Evolution of herbivory in terrestrial vertebrates:perspectives from the fossil record. Cambridge University Press,Cambridge.

——. 2005. Phylogenetic and taxic perspectives on sauropoddiversity. Pp. 104–124 in K. A. Curry Rogers, and J. A. Wilson,eds. The sauropods: evolution and paleobiology. University ofCalifornia Press, Berkeley.

Upchurch, P. M., P. D. Mannion, R. B. J. Benson, R. J. Butler, andM. T. Carrano. 2011. Geological and anthropogenic controls onthe sampling of the terrestrial fossil record: a case study from theDinosauria. Geological Society of London Special Publications358:209–240.

Weishampel, D. B. 1984. Interactions between Mesozoicplants and vertebrates: fructifications and seed predation. NeuesJahrbuch für Geologie und Paläontologie, Abhandlungen167:224–250.

Weishampel, D. B., and D. B. Norman. 1989. Vertebrate herbivoryin the Mesozoic; jaws, plants, and evolutionary metrics. Geolo-gical Society of America Special Paper 238:87–101.

Weishampel, D. B., P. Dodson, and H. Osmolska. 2004. TheDinosauria. University of California Press, Berkeley.

Whitlock, J. A. 2011. Inferences of diplodocoid (Sauropoda:Dinosauria) feeding behavior from snout shape and microwearanalyses. PLoS ONE 6:e18304.

Wills, M. A., D. E. G. Briggs, and R. A. Fortey. 1994. Disparity as anevolutionary index: a comparison of Cambrian and Recentarthropods. Paleobiology 20:93–130.

Young, M. T., andM. D. Larvan. 2010. Macroevolutionary trends inthe skull of sauropodomorph dinosaurs—the largest terrestrialanimals to have ever lived. InA. M. T. Elewa, ed. Morphometricsfor nonmorphometricians: lecture notes in Earth sciences124:259–269. Springer, Berlin.

Zelditch, M. L., D. L. Swiderski, and H. D. Sheets. 2012. Geometricmorphometrics for biologists: a primer. Academic, New York.



g uly h uv - bioone › journals › paleobiology › volume-43 › issue...herbivorous dinosaur jaw...

Documents