ARTICLEReceived11May2015|Accepted3Jul2015|Published11Aug2015Geographicrangedidnotconferresiliencetoextinctioninterrestrialvertebratesattheend-TriassiccrisisAlexanderM.Dunhill1,2&MatthewA. Wills1Rates of extinction vary greatly through geological time, with losses particularly concentratedin mass extinctions. Species duration at other times varies greatly, but the reasons for this areunclear. Geographical rangecorrelateswithlineagedurationamongstmarineinvertebrates,but it is less clear how far this generality extends to other groups in other habitats. It is alsounclearwhetherawidegeographicaldistributionmakesgroupsmorelikelytosurvivemassextinctions. Here we test for extinction selectivity amongst terrestrial vertebrates across theend-Triassic event. We demonstrate that terrestrial vertebrate clades with larger geographicalranges were more resilient to extinction than those with smaller ranges throughouttheTriassicandJurassic. However, thisrelationshipweakenedwithincreasingproximitytothe end-Triassic mass extinction, breaking down altogether across the event itself. Wedemonstratethatthesendingsarenotafunctionofsamplingbiases; aperennial issueinstudiesofthiskind.DOI:10.1038/ncomms89801Milner Centre for Evolution, University of Bath, Claverton Down, Bath BA2 7AY, UK.2School of Earth and Environment, University of Leeds, Leeds LS2 9JT,UK. CorrespondenceandrequestsformaterialsshouldbeaddressedtoA.M.D. (email:a.dunhill@leeds.ac.uk).NATURECOMMUNICATIONS | 6:7980| DOI: 10.1038/ncomms8980| www.nature.com/naturecommunications 1&2015MacmillanPublishersLimited.Allrightsreserved.Is it possible to make generalizations about which cladesandhigher taxaaremost likelytogoextinct ongeologicaltimescales? Geographic range is often claimed to bea determinant of extinction vulnerability16. Groups withrestricted ranges are widely believed to suffer extinctionmore frequently thanthose withbroader ranges1because thelatter are thought to be afforded some resilience to regionalenvironmental perturbations7,8. During global biotic crises, bycontrast, there is reasoned to be no such selectivity2,6sincewidespread environmental disturbances simultaneously affectboth wide- and narrow-ranging taxa on global scales1,3,6.Surprisingly, the effect of geographic range on extinctionrisk has not been tested comprehensively for the terrestrialfossil record, withastrikingpaucityof studies onvertebratesof any kind. Most published studies focus on marineinvertebrates36,810, anddespite geographic range being usedas a major criterion for assessing the extinction risk ofmodernterrestrial species11, itisunclearthatthendingsfromthese fossil taxa can be extended to all groups in all majorhabitats. Moreover, littleisknownaboutthedifferencebetweenintervals with background levels of extinction and thosecharacterized as mass extinctions3. The only way to answersuchquestions is toutilize fossil evidence of past organismaldistributions5,12.The Triassic to Jurassic is a critically important period interrestrial vertebrateevolution13,14. Inparticular, it follows thelargest of all massextinctions, thePermoTriassicgreat dying1315. Many terrestrial vertebrate lineages originated in theaftermathof the PermoTriassic event, but were subsequentlysubjectedtomajorchangesinterrestrial ecosystemsthroughoutthe ensuingTriassic andJurassic. These changes includedthegradual aridication of Pangaea16, as well as its initial rifting andfragmentation17alliedtothe eruptionof the Central Atlanticmagmatic province18,19. This culminated in the end-Triassic massextinction event16,20that saw the demise of numerous amphibianand reptile groups before the subsequent rise to dominance of thedinosaurs13,16,2123.The signicant vertebrate faunal turnover throughout theTriassicandJurassic(lyingeithersideoftheend-Triassicmassextinction) make this an ideal period in which to study extinctionselectivity. We therefore collated palaeobiogeographical andstratigraphic distributional data24for Triassic and Jurassicterrestrial vertebrate clades to ask three questions. (1) Is there isa relationship between palaeobiogeographical distribution and therisk of extinction during normal times? (2) Does any suchrelationshipdisappear across theend-Triassicmass extinction?(3) Can any of these ndings be attributed to sampling biases?We nd that wider geographical range conferred greaterresilience to extinction in terrestrial vertebrate groups throughoutmost of the Triassic and Jurassic. However, this insuranceweakened towards the end-Triassic mass extinction, and wasimperceptible across the event itself. Major sampling biases werediscounted as the cause of these patterns.ResultsandDiscussionGeographic range and diversication rates are correlated.Diversication rates and changes in geographic range at the cladelevel are positivelyandstronglycorrelatedacross all time bins(Figs 1 and 2 and Supplementary Table 1), with weaker (but stillmostly signicant) correlations when the data are subdivided intoepochs (Fig. 2 and Supplementary Table 1). The strong correlationbetween changes in geographic range and diversication rateacross all time intervals (and at epoch level and within stage-leveltime bins) conrms that increasing range size coincides withincreasingdiversity, whilst rangesizereductionstypicallyattenddiversity reductions. Taxa with larger geographic ranges aretherefore more likely toexhibit increasing diversity andlowerextinctionratesthanthosewithsmallerranges. Greaterratesoforiginationmightalsobe expected to resultfrom moreextensiveranges; rst, because large ranges are more likely to be fragmentedinto peripheral isolates, and second, because larger rangesencompassagreaterdiversityofenvironmentsandselectivefor-ces25,26. This patternis the opposite of that proposedfor themarineinvertebratefauna, wheretaxawithnarrowrangesshowhigher origination rates6,27. The results from our geographic rangedata sets, both raw convex hull and standardized mean great circledistances (GCDs; Fig. 1), are similar, demonstrating that ourndings are not contingent on the precise methodology used.100101011012150 175 200 225 250Triassic JurassicEarlyMiddleLateEarlyMiddleLateGeological time (Ma)Geographic range(lnR1lnR0)Diversification(lnD1lnD0)Mean GCD n = 5Mean GCD n = 10Geographic range(lnR1lnR0)Convex hullsMean GCDsFigure1|Meanratesofchangeingeographicrangesizeanddiversicationratesforterrestrial vertebrates,partitionedbytimebin.Rateofchangeingeographicrangesize(DGeographicrange)asrepresented by (a) convex hulls around raw palaeogeographic occurrencesand (b) mean GCD between palaeogeographic occurrences standardized to5 and 10 samples; and (c) mean vertebrate diversication rates of ranged-throughdiversitydata(DDiversication). ThefossilrecordsoftheLadinian12,Toarcian19andmuchofthemiddleJurassic2830areoflowerqualitythantherestofMesozoic,andthismaycontributetosomeofthenegativediversitychangestherein.The dropindiversityobservedthroughthe Rhaetian could also be regarded as a sampling artefact as the Rhaetianis not as well sampled as the preceding Norian. However, the upper Triassicrepresentsoneofthehighest-qualityterrestrialfossilrecords22,30.AlternatinggreyandwhitebarscorrespondtoTriassicJurassicepochs.ARTICLE NATURECOMMUNICATIONS|DOI:10.1038/ncomms89802 NATURECOMMUNICATIONS | 6:7980| DOI: 10.1038/ncomms8980| www.nature.com/naturecommunications&2015MacmillanPublishersLimited.Allrightsreserved.This relationship breaks down across the mass extinction.Strikingly, ratesof geographicrangechangeanddiversicationare not signicantly correlated immediately before theend-Triassic extinction (during the Rhaetian), whereas thiscorrelationissignicantevenintheCarnianandNorianstagesthat immediately precede it (Fig. 2 and Supplementary Table 1).Hence, diversicationratebecomes decoupledfromrangesizechange rate across the extinctionboundaryandthe insuranceagainst extinctionofferedbylarger geographicranges at othertimes disappears. During the end-Triassic mass extinction event,relativelywidespreadgroupsareas likelyto sufferhighlevels ofextinction as groups with narrower geographic ranges. Forexample, Phytosauria andTheropoda have similar, geographicranges in the Rhaetian (Fig. 3). However, phytosaurs suffercomplete extinction at the end Triassic, whereas theropoddiversity remains stable across the boundary and into theHettangian, evenwhilstundergoingsignicant rangeexpansion(Fig. 3). Of the time intervals that do not showsignicantcorrelation between diversication rate and geographic rangechange rate, all have very small sample sizes apart fromtheRhaetian. ThestandardizedmeanGCDrangemetricsshowlesscorrelationwithdiversicationrate thanthe rawconvex hullmetric (Supplementary Table 1). This is unsurprising since areducedsample size leads toa reductioninstatistical power.However, many of these nonsignicant correlations still haverelatively high correlation coefcients and are approachingsignicance, whereas the Rhaetianresults are clearly different(with negative coefcients that are far from signicant)(Supplementary Table 1). Our temporal divisions are very muchlonger (2.018.9 Myrs)28,29thanthe extinctionevent20, whichoccurred in pulses over a period of o1.0 Myr (ref. 19). Effects aretherefore time averaged, meaning that the breakdownof therelationship between diversication rate and the rate ofgeographic range changeintheveryendTriassic (anintervalknown to contain a major biotic crisis) is even more striking.Samplingbiaseshavelittleeffectonourresults. Weobservedseveral signicant bivariate correlations betweendiversicationrate, geographic range change rate and various putative samplingproxies detrended using rst differences (Supplementary Tables 2and 3). However, multiple regression models identied thechanges in geographic range rate as the dominant variableinuencing diversication rate (Table 1 and SupplementaryTables 49).Althoughcertain parts of the TriassicJurassic are reputedto have a poor terrestrial fossil record (that is, Ladinian,ToarcianmidJurassic)13,20,3032, thelateTriassicpossessesoneof the best23,32(Fig. 4). We see positive correlations between landarea and geographic range and a negative correlation between sealevel andrangeintheGCDdata, but alsopositivecorrelationsbetweensealevel andgeographicrangeandbetweensealevelanddiversicationrate inthe convexhull data. This suggeststhatgreaterlandareaandlowersealevelsmayresultingreatergeographicranges amongst terrestrial organisms. However, thenegative correlationbetweenstandardizedrange andsea level(and the lack of correlation between diversication rate and landarea) suggests that while expanding landmasses might beexpected to induce the expansion of terrestrial ranges andincrease diversication, climatic and other effects complicate this21012320 10 0 102101210 5 0 10 15 5 152.01.00.50.00.51.510 15 5 0 5210122 1 0 1 2 3210122 1 0 1 21.00.50.00.51.01.52.01.0 0.0 1.0 2.0n = 5 : rs = 0.34**n = 10 : rs = 0.44***n = 5 : rs = 0.24n = 10 : rs = 0.26n = 5 : rs = 0.22n = 10 : rs = 0.21rs = 0.67*** rs = 0.64*** rs = 0.40Diversification (lnD1lnD0)Convex hull geographic range (lnR1lnR0)Mean GCD geographic range (lnR1lnR0)Figure2|Scatterplotsofdiversicationratesagainstperlineageratesofchangeingeographicrangesize.Rateofgeographicrangesizechange(DGeographicrange)ascalculatedusingconvexhullsaroundrawoccurrencedatafor(a)alltimebins, (b)lateTriassicand(c)Rhaetian.Rateofgeographic range size change calculated as mean GCDs between occurrences standardized to samples of 5 and 10 occurrences for (d) all time bins, (e) lateTriassicand(f)Rhaetian.Spearmansrank correlationcoefcients**signicantatPo0.01,***signicantatPo0.001.NATURECOMMUNICATIONS|DOI:10.1038/ncomms8980 ARTICLENATURECOMMUNICATIONS | 6:7980| DOI: 10.1038/ncomms8980| www.nature.com/naturecommunications 3&2015MacmillanPublishersLimited.Allrightsreserved.CynognathiaRhaetian: 1.337Hettangian: PhytosauriaRhaetian: 6.967Hettangian: ProbainognathiaRhaetian: 13.23Hettangian:10SauropodomorphaRhaetian: 12.503Hettangian:15.935TheropodaRhaetian: 5Hettangian:5AetosauriaRhaetian: 1.832Hettangian: 1000'0" E 600'0" E 200'0" E 200'0" W 600'0" W700'0" N300'0" N100'0" S500'0" S900'0"0km1000'0" E 600'0" E 200'0" E 200'0" W 600'0" W700'0" N300'0" N100'0" S500'0" S900'0"0 km2,500N5,0002,5005,000NFigure 3|Geographic range maps before and after the end-Triassic mass extinction. Convex hull geographic range maps and mean generic diversity ofsix terrestrial vertebrate groups during the (a) Rhaetian and (b) Hettangian. The Aetosauria, Cynognathia and Phytosauria became extinct during the bioticcrisis, despitethewidespreaddistributionofCynognathiaandPhytosauriaintheRhaetian.The Probainognathia,SauropodomorphaandTheropodaallsurvived the biotic crisis and expanded their ranges in the Hettangian, albeit with different diversication trajectories. The diversity of Sauropodomorphaincreased,thediversityofTheropodawasstationary andthatofProbainognathiadecreasedacrosstheboundary.Table 1 | Summary of multiple regression models of diversity change (dependent variable) in terms of geographic range changeandsamplingandenvironmentalproxies.Model Dependent Independents AdjustedR2P AICConvexfull Diversitychange Range change landarea sealevel formations abundance totalrange 0.53 o0.001 98.59Convexbest Diversitychange Range change 0.64 o0.001 100.53GCD5full Diversitychange Range change landarea sealevel formations abundance totalrange 0.22 o0.001 65.5GCD5best Diversitychange Range change totalrange sealevel 0.23 o0.001 68.95GCD10full Diversitychange Range change landarea sealevel formations abundance totalrange 0.32 o0.001 63.66GCD10best Diversitychange Range change totalrange sealevel 0.35 o0.001 68.97AIC, akaikeinformationcriterion.ARTICLE NATURECOMMUNICATIONS|DOI:10.1038/ncomms89804 NATURECOMMUNICATIONS | 6:7980| DOI: 10.1038/ncomms8980| www.nature.com/naturecommunications&2015MacmillanPublishersLimited.Allrightsreserved.relationship. Indeed, the typicallyharshenvironments indeepcontinental interiors may constrainmany ranges. Specically,lethallyhottemperatures, particularlyintheearlyTriassic, mayhavelimitedorevenexcludeddiversityinequatorial regions33.Individuallineagerangescorrelatewithtotalrangesacrosstimebins, and both appear to reect the same underlying pattern. Theconsiderablevariationbetweenindividual lineagerangeswithineach bin (coupled with the fact that the standardized rangemetrics still show some correlation with total range, albeit weakerthan the raw range metrics) indicates that range estimates are notgoverned straightforwardly by sampling intensity (Figs 1 and 3).Moreover, theabsenceof terrestrial vertebratesfromequatorialregions is entirely plausible during the climatic greenhouse of theearlyTriassic33. Groupswithbroadergeographical distributionsare likely to be subject to a wider range of selective pressures andthe peripheral isolation of subgroups; both factors favouringspeciation and increasing diversity25,26. However, the weakcorrelationobservedbetweenchanges infossil abundanceandbothdiversicationrateandgeographic range change mayberepresentativeofsamplingbias. Itisalsopossiblethatagreaterinvestment of research effort in more abundant fossil groups mayhave resulted in increased taxonomic splitting34.We also observed signicant pairwise correlations betweenboth raw and standardized geographic range change anddiversicationrate onone hand, andchanges innumbers offormations on the other. Although formation counts are regardedas effective sampling proxies for terrestrial data sets35,redundancy between sampling proxies and diversity metrics(arising from the probable non-independence of formationand fossil content) remains problematic3638. Inpractice, thelevel ofthisredundancyislikelytobeminimizedbytheuseofall terrestrial vertebrate-bearingformations39,40, rather thanbyadoptingastrictercountofonlythoseformationscontainingaparticulargroupof terrestrial vertebrate fossils31,4143.However,standardization of geographic range data results in the removal ofsignicant correlations between range size and fossil abundance,coupledwithaweakeningofthecorrelationbetweenrangesizeand total range size (that is, standardizing geographic rangecalculationstoaconstantsamplesizeacrossall lineagesineachtime bin appears to remove putative sampling effects). Bycontrast, standardizing range data does not affect the correlationbetweenrangesizeandformationcounts. Thislastrelationshipmay therefore arise fromredundancy36, rather than being atemporal bias resulting from variation in the amount of preservedsedimentary rock (and concomitant intensity of sampling)through geological time.Themultipleregressionmodels showthat geographicrangechange is the dominant driver of diversication rate through theTriassicJurassic, totheexclusionofallthesamplingproxiesinthe modelusing convexhulls, but with total rangeand sealevelshowing some inuence in the standardized mean GCD models.This suggests that, although sampling biases are a perennial issuein fossil data sets, the link between changes in geographic rangeanddiversicationrateappearrobust, despitethepatchynatureof the vertebrate fossil record.Implications for extinction studies. We demonstrate that broadgeographic range conferred insurance against extinction on majorcladesofterrestrial vertebratesduringperiodswithbackgroundlevels of extinction. In line with marine invertebrate studies acrossthe same biotic crisis6and at other times in the Phanerozoic3, thisinsurancedisappearedduringtheend-Triassicmassextinction.However, these results are in marked contrast to patternsreportedfor marineinvertebrates at theCretaceousPaleogeneextinction9,44(where groups with larger geographic rangesretainincreasedresilience to extinctionacross the crisis thanthosewithsmallerranges). Itisreasonablycommontoobservediscretemacroevolutionarypatterns indifferent higher taxaoracross different major habitats. Notable examples include theincongruence between terrestrial and marine Phanerozoicdiversitycurves4547, variationsintheapparent forceofCopesrule sensu lato in different higher taxa4850, and variations in therelationships between body size, population density and fecundityacross clades51,52. However, the differences between patternsobservedatmajorextinctioneventsmayresultfromdifferencesLand area (106 km2)Sea level (m)0102030405060704080120Formations02,0004,000Abundance5.01071.51082.5108Total range (km2)175 200 225 250Triassic JurassicEarlyMiddleLateEarlyMiddle150Late6,000Geological time (Ma)150145140135130Figure4|Samplingandenvironmentalproxy data.(a)Non-marinearea58, (b) average sea level60, (c) terrestrial formation count24, (d) fossilabundance24and (e) total geographic range of all taxa. Alternating grey andwhitebarscorrespondtoTriassicJurassicepochs.NATURECOMMUNICATIONS|DOI:10.1038/ncomms8980 ARTICLENATURECOMMUNICATIONS | 6:7980| DOI: 10.1038/ncomms8980| www.nature.com/naturecommunications 5&2015MacmillanPublishersLimited.Allrightsreserved.betweenthe particular drivingforces of the crises themselves.These differences highlight the need for greater numbers oflarge-scale, comparative studies before attempting to makemacroevolutionary generalizations. Fortunately, the expansionandrenementofresourcessuchasthePaleobiologyDatabase,coupled with quantitative tools such as GIS, have great potentialfor testing temporal and spatial macroevolutionary patterns.Equally, palaeontological dataprovideabroaderperspectiveonthecurrent biodiversitycrisis. Specically, it enablesdeeptimetests of the purported relationships between present dayextinction susceptibility and geographical range size53,54,latitudinal distribution55, niche breadth56and body size57,58.MethodsFossiloccurrencedata. Stage-level occurrence data for TriassicJurassicterrestrial vertebrates were initially downloaded from the Paleobiology Databse24(https://paleobiodb.org) on 4 February 2013 (last accessed 20 April 2015) and, afterpreprocessing, consisted of 3,507 occurrences of 857 genera (see SupplementaryNote 1 for Paleobiology Database download specications). Terrestrial vertebrateoccurrences from marine deposits were not included as they would not give a truerepresentation of geographic range. Ichnogenera and other form taxa where thenremoved from the data set as they could not be assigned accurately to parentgenera. Marine tetrapod taxa recorded in terrestrial formations were also removed.Generic indeterminatetaxa and taxa with uncertain generic assignments (that is,aff., cf., ex gr., sensu lato, ?) were also excluded. Although these exclusionsinevitably resultedin an underestimation of the geographic range of highertaxonomic groups, it would be inappropriate to compare ranges constructed fromtaxa of uncertainafliation with rates of generic extinction, origination anddiversication, which cannot include these indeterminate occurrences.Fossil occurrences were vetted for synonymy and outdated taxonomy, andsorted into higher taxonomic groups according to phylogenetic and ecologicalrelationships within the constraints of reasonable sample sizes (see SupplementaryNote 1 for detailed classicationof taxa). As with all higher taxonomicclassication, groups were not directly comparable. This is an unavoidable problemunless working at the species or, to a lesser extent, the generic level. Two data setswere compiled: data1 and data2 (SupplementaryFig. 1). Data1 (SupplementaryData 1) contained a number of paraphyletic assemblages where basal taxa ofparticular lineages were grouped together to form a paraphyletic stem assemblage(for example, Archosauriformes,basal Cynodontia and Dinosauromorpha;Supplementary Fig. 1). Since the inclusionof paraphyletic groups is arguablyproblematic (they do not represent true evolutionary groups), a second data setexcluding all parapyla was also prepared. In data2, the paraphyletic assemblageswere collapsed into smaller, monophyletic family-level groups wherever possible(Supplementary Fig. 1). The two data sets correlated very closely in terms of bothgeographic range change rate and diversicationrate. All analyses in themanuscript therefore refer exclusively to data1.Fossil occurrences were binned at the stratigraphic stage level. Any occurrencesnot assigned to a stage were attached to the stage, or range of stages, correspondingto the formation from which they were recovered. Fossil occurrences that wereassigned to more than one stage were randomly assigned to a single stage withintheir given range, a process that was repeated 1,000 times to obtain a mean value.This method avoided either the articial ination or deation of taxonomicrichness in any given stage compared with maximum or minimum diversity values.Sampling and environmental proxy data. Non-marine area. A mean estimate ofthe continental landmass for each stage59. It was anticipated that geographic rangewould correlate positively with land area as the area of terrestrial habitat creates anupper limit for the geographic range of terrestrial organisms. These measurementswere derived from an independentsource59, and were subject to differentdenitions of stage-level boundaries than the fossil occurrence data set, which usedthe Geological Time Scale 2012 (ref. 60).Sea level. A mean estimate of relative sea level for each stage61. It was expectedthat geographic range would correlate inversely with mean sea level, as higher sealevel would result in less terrestrial landmass for terrestrial organisms to inhabit. Aswith the non-marine area measurements, the sea level averages were obtained froman independent source61and are subject to different denitions of stage-levelboundaries from the fossil occurrence data set60.Terrestrial formations. Formation counts are widely regarded as effectivesampling proxies for the terrestrial fossil record24,31,35,42,62. It is still unclear if thisis true, as formation counts probably share a common signal with fossil occurrencedata (that is, formations are not independent from their fossil content36,37).However, given the lack of comprehensive data on global sedimentary rock outcroparea, formation counts offer the only possible metric of global rock availability. Inthis analysis, redundancy was minimizedby using a total count of terrestrialtetrapod bearing formations, rather than a strict count of group-specic bearingformations. There is also an argumentfor redundancy between formation countsand geographic range, as a taxon that is genuinely wide ranging is more likely to bepresent in more formations across the globe than a taxon with a small geographicrange. Such possibilities were explored using multiple regression models.Fossil abundance. The fossil abundanceper time period serves as a proxy forhuman sampling and collecting effort24. However, there is danger of circularity, aspalaeontologists will be more likely to collect from formations yielding a highernumber of fossils38,63. Therefore, fossil abundance may be more representative ofpreservational factors or true biological abundance, rather than a measure ofhuman sampling effort.Total geographic range. It is reasonable to assume that vertebrates werenot genuinely absent from large areas of the globe through parts of theTriassicJurassic. Therefore, if total geographic range (that is the total geographicrange of all tetrapod occurrences per time bin) were to correlate strongly with thegeographic ranges of individual fossil groups, it would indicate that the pattern ofgeographic range through the study period is controlled by spatial sampling ratherthan reecting a biological pattern.Analysis. Palaeogeographic ranges were constructedby converting modernfossil occurrence coordinates to palaeocoordinates using PointTracker64. Palaeo-geographic ranges were constructed using two methods: (i) in ArcGIS v10.1 usingconvex hulls around the palaeogeographic occurrences for each group65,66; and (ii)using mean GCDs between palaeogeographic occurrences with sample sizesstandardized to 5 and 10 occurrences per lineage per time bin and replicated 1,000times to obtain a mean value. GCD distances were calculated using the sphericallaw of cosines (as an acceptable approximation of the Haversine formula forterrestrial calculations).Per lineage origination (Or) and extinction (Er) rates were calculated using themethodology of Foote67and modied by Foote68:Or lnNbtNftNbt1Er lnNbtNbLNbt2where Nbtnumber of range-through taxa, Nftnumber of taxa that originatewithin time bin and cross top boundary of time bin and NbLnumber of taxa thatcross bottom boundary of time bin but make their last appearance in time bin.Rates were not expressed relative to time bin duration; although this may causeunderestimation of rates in shorter time bins relative to longer time bins, Foote28demonstrated that both extinction and origination are pulsed rather than spreadthroughout time intervals. All analyses were carried out at the generic level.No signicant correlations were detected between geographic range change andextinction rate or between geographic range change and origination rate(Supplementary Table 10). The absence of signicant correlations betweenorigination/extinction rates and change in geographic range could be regarded assomewhat surprising, but this result is a function of the paucity of data for the ratecalculations. However, the extinction and origination rate calculations rely on taxathat range through three consecutive time bins67, which are scarce in this data setbecause of the patchiness of the terrestrial fossil record and the long durations ofthe time bins. Therefore,it was judged that a metric of diversicationcalculatedfrom generic range data offered a more robust picture of biotic change.Diversication rate (Dr) and geographic change rate (Rr) were calculatedusing ametric modied from Rode and Lieberman69:Drln D1 ln D0 3Rrln R1 ln R0 4where D1ranged-through diversity calculated from rst and last appearances forcurrent time interval, D0ranged-through diversity calculated from rst and lastappearances for the previous time interval, R1geographic range in time intervaland R0geographic range in previous time interval.Relationships between geographic range change and generic diversication rateswithin clades were tested using pairwise Spearmans rankorder correlation tests.Putative sampling biases were investigated using both pairwise correlation andmultiple regression models, with diversication rate as the dependent variable andgeographic range and various sampling proxies as independent variables. Firstdifferencing was used to detrend the sampling proxy data before comparison withthe biotic rate data. All statistical analyses were carried out in R 3.1.1.References1. Barnosky, A. D. et al. Has the Earths sixth mass extinction already arrived?Nature 471, 5157 (2011).2. Jablonski, D. Mass extinctions and macroevolution. Paleobiology 31, 192210(2005).3. Payne, J. L. & Finnegan, S. The effect of geographic range on extinction riskduring background and mass extinction. Proc. Natl Acad. Sci. USA 104,1050610511 (2007).4. Powell, M. G. Geographic range and genus longevity of late Paleozoicbrachiopods. Paleobiology 33, 530546 (2007).5. Rode, A. L. & Lieberman, B. S. Using GIS to unlock the interactions betweenbiogeography, environment, and evolution in Middle and Late DevonianARTICLE NATURECOMMUNICATIONS|DOI:10.1038/ncomms89806 NATURECOMMUNICATIONS | 6:7980| DOI: 10.1038/ncomms8980| www.nature.com/naturecommunications&2015MacmillanPublishersLimited.Allrightsreserved.brachiopods and bivalves. Palaeogeogr.Palaeoclimatol. Palaeoecol. 211,345359 (2004).6. Kiessling, W. & Aberhan, M. Geographical distribution and extinction risk:lessons from Triassic-Jurassicmarine benthic organisms. J. Biogeogr. 34,14731489 (2007).7. Foote, M., Crampton, J. S., Beu, A. G. & Cooper, R. A. On the bidirectionalrelationship between geographic range and taxonomic duration.Paleobiology34, 421433 (2008).8. Nurnberg, S. & Aberhan, M. Habitat breadth and geographic range predictdiversitydynamics in marine Mesozoic bivalves. Paleobiology39, 360372(2013).9. Jablonski, D. Extinction and the spatial dynamics of biodiversity. Proc. NatlAcad. Sci. USA 105, 1152811535 (2008).10. Jablonski, D. & Roy, K. Geographical range and speciation in fossil and livingmolluscs. Proc. R. Soc. B 270, 401406 (2003).11. Mace, G. M. et al. Quantication of extinction risk: IUCNs system forclassifying threatenedspeciesCuanticacion del Riesgo de Extincion: Sistemade la UICN para la Clasicacion de Especies Amenazadas. Conserv. Biol. 22,14241442 (2008).12. Valentine,J. W., Jablonski, D., Krug, A. Z. & Berke, S. K. The sampling andestimation of marine paleodiversity patterns: implications of a Pliocene model.Paleobiology39, 120 (2013).13. Brusatte, S. L., Benton, M. J., Desojo, J. B. & Langer, M. C. The higher-levelphylogeny of Archosauria (Tetrapoda: Diapsida). J. Syst. Palaeontol. 8, 347(2010).14. Ezcurra, M. D. Biogeography of Triassic tetrapods: evidence for provincialismand driven sympatric cladogenesis in the early evolution of modern tetrapodlineages. Proc. Biol. Sci. 277, 25472552 (2010).15. Ruta, M., Botha-Brink, J., Mitchell, S. A. & Benton, M. J. The radiation ofcynodonts and the ground plan of mammalianmorphological diversity. Proc.Biol. Sci. 280, doi:10.1098/rspb.2013.1865 (2013).16. Tanner, L. H., Lucas, S. G. & Chapman, M. G. Assessing the record and causesof Late Triassic extinctions. Earth Sci. Rev. 65, 103139 (2004).17. Whiteside, J. H., Grogan, D. S., Olsen, P. E. & Kent, D. V. Climatically drivenbiogeographic provinces of Late Triassic tropical Pangea. Proc. Natl Acad. Sci.USA 108, 89728977 (2011).18. Whiteside, J., Olsen, P. E., Eglinton, T., Brookeld, M. E. & Sambrotto, R. N.Compound-specic carbon isotopes from Earths largest ood basalt eruptionsdirectly linked to the end-Triassicmass extinction. Proc. Natl Acad. Sci. USA107, 67216725 (2010).19. Blackburn, T. J. et al. Zircon U-Pb geochronology links the end-triassicextinction with the central atlantic magmatic province. Science 340, 941945(2013).20. Toljagic, O. & Butler, R. J. TriassicJurassic mass extinction as trigger for theMesozoic radiation of crocodylomorphs. Biol. Lett. 9, doi:10.1098/rsbl.2013.0095(2013).21. Benton, M. J. in The Fossil Record Vol. 2 (ed. Benton, M. J.) 681715 (Chapmanand Hall, 1993).22. Brusatte, S. L. et al. The origin and early radiation of dinosaurs. Earth Sci. Rev.101, 68100 (2010).23. Langer, M. C., Ezcurra, M. D., Bittencourt, J. S. & Novas, F. E. The origin andearly evolution of dinosaurs. Biol. Rev. 85, 55110 (2010).24. Carrano, M. T. & Butler, R. J. Taxonomic occurrences of Triassicto Jurassic Tetrapoda. Paleobiology Database: http://paleobiodb.org(20132015).25. Roy, K. Effects of the mesozoic marine revolution on the taxonomic,morphologic,and biogeographic evolution of a group: aporrhaid gastropodsduring the Mesozoic. Paleobiology 20, 274296 (1994).26. Budd, A. F. & Coates, A. G. Nonprogressive evolution in a clade of CretaceousMontastraea-like corals. Paleobiology 18, 425446 (1992).27. Jablonski, D. Larval ecology and macroevolution in marine invertebrates. Bull.Mar. Sci. 39, 565587 (1986).28. Foote, M. Pulsed origination and extinction in the marine realm. Paleobiology31, 620 (2005).29. Bambach, R. K. Phanerozoic biodiversity and mass extinctions.Annu. Rev.Earth Planet. Sci. 34, 127155 (2006).30. Stubbs, T. L., Pierce, S. E., Rayeld, E. J. & Anderson, P. S. L. Morphologicaland biomechanical disparity of crocodile-line archosaurs following the end-Triassic extinction. Proc. Biol. Sci. 280, doi:10.1098/rspb.2013.1940 (2013).31. Upchurch, P., Mannion, P. D., Benson, R. B. J., Butler, R. J. & Carrano, M. T.Geological and anthropogenic controls on the sampling of the terrestrial fossilrecord: a case study from the Dinosauria. Geol. Soc. Spec. Pub. 358, 209240(2011).32. Wang, S. C. & Dodson, P. Estimating the diversity of dinosaurs. Proc. NatlAcad. Sci. USA 103, 1360113605 (2006).33. Sun, Y. D. et al. Lethally hot temperatures during the early Triassic greenhouse.Science 338, 366370 (2012).34. Lloyd, G. T., Young, J. R. & Smith, A. B. Taxonomic structure of the fossilrecord is shaped by sampling bias. Syst. Biol. 61, 8089 (2012).35. Benson, R. B. J. & Upchurch, P. Diversity trends in the establishment ofterrestrial vertebrate ecosystems: interactions between spatial and temporalsampling biases. Geology 41, 4346 (2013).36. Dunhill, A. M., Hannisdal,B. & Benton, M. J. Disentangling geologicalmegabias and common-causefrom redundancy in the British fossil record.Nat. Commun. 5, 4818 (2014).37. Benton, M. J., Dunhill, A. M., Lloyd, G. T. & Marx, F. G. Assessing the quality ofthe fossil record: insights from vertebrates. Geol. Soc. Spec. Pub. 358, 6394 (2011).38. Dunhill, A. M., Benton, M. J., Twitchett, R. J. & Newell, A. J. Testing the fossilrecord: sampling proxies and scaling in the British Triassic-Jurassic.Palaeogeogr.Palaeoclimatol. Palaeoecol. 404, 111 (2014).39. Benson, R. B. J. & Butler, R. J. Uncovering the diversication history of marinetetrapods: ecology inuences the effect of geological sampling biases. Geol. Soc.Spec. Pub. 358, 191208 (2011).40. Benson, R. B. J., Butler, R. J., Lindgren, J. & Smith, A. S. Mesozoic marinetetrapod diversity: mass extinctions and temporal heterogeneity in geologicalmegabiases affecting vertebrates.Proc. R. Soc. B 277, 829834 (2010).41. Frobisch, J. Global taxonomic diversityof Anomodonts (Tetropoda,Therapsida) and the terrestrial rock record across the Permian-Triassicboundary. PLoS ONE 3, e3733 (2008).42. Barrett, P. M., McGowan, A. J. & Page, V. Dinosaur diversity and the rockrecord. Proc. R. Soc. B 276, 26672674 (2009).43. Butler, R. J., Barrett, P. M., Nowbath, S. & Upchurch,P. Estimating theeffects of the rock record on pterosaur diversitypatterns: implications forhypotheses of bird/pterosaur competitivereplacement. Paleobiology 35,432446 (2009).44. Jablonski, D. Scale and hierarchy in macroevolution. Palaeontology 50, 87109(2007).45. Benton, M. J. The origins of modern biodiversity on land. Phil. Trans. R. Soc.Lond. B 365, 36673679 (2010).46. Alroy, J. The shifting balance of diversity among major marine animal groups.Science 329, 11911194 (2010).47. Sepkoski, J. J., Bambach, R. K., Raup, D. M. & Valentine,J. W. Phanerozoicmarine diversityand the fossil record. Nature 293, 435437 (1981).48. Jablonski, D. Body-size evolution in Cretaceous molluscs and the status ofCopes rule. Nature 385, 250252 (1997).49. Benson, R. B. J., Frigot, R. A., Goswami, A., Andres, B. & Butler, R. J.Competition and constraint drove Copes rule in the evolution of giant yingreptiles. Nat. Commun. 5, 3567 (2014).50. Raia, P., Carotenuto, F., Passaro, F., Fulgione, D. & Fortelius, M. Ecologicalspecialization in fossil mammals explains Copes Rule. Am. Nat. 179, 328337(2012).51. Savage, V. M., Gillooly, J. F., Brown, J. H., West, G. B. & Charnov, E. L. Effectsof body size and temperature on population growth. Am. Nat. 163, 429441(2004).52. Don, R. L. Inuence of body size and population density on fertilizationsuccess and reproductive output in a free-spawning invertebrate. Biol. Bull. 181,261268 (1991).53. Cooper, N., Bielby, J., Thomas, G. H. & Purvis, A. Macroecology and extinctionrisk correlates of frogs. Glob. Ecol. Biogeogr. 17, 211221 (2008).54. Pimm, S. L. et al. The biodiversity of species and their rates of extinction,distribution, and protection. Science 344,, 987 (2014).55. Cardillo, M. et al. The predictability of extinction: biological and externalcorrelates of decline in mammals. Proc. Biol. Sci. 275, 14411448 (2008).56. Slatyer, R. A., Hirst, M. & Sexton, J. P. Niche breadth predictsgeographical range size: a general ecological pattern. Ecol. Lett. 16, 11041114(2013).57. McCain, C. M. & King, S. R. B. Body size and activity times mediate mammalianresponses to climate change. Global Change Biol. 20, 17601769 (2014).58. Benton, M. J. Exploring macroevolutionusing modern and fossil data. Proc.Biol. Sci. 282, 20150569 doi:10.1098/rspb.2015.0569 (2015).59. Smith, A. G., Smith, D. G. & Funnell, B. M. Atlas of Mesozoic and CenozoicCoastlines Vol. 99 (Cambridge UniversityPress, 1994).60. Gradstein, F. M., Ogg, J. G., Schmitz, M. D. & Ogg, G. M. The Geologic TimeScale 2012. Vol. 2, 1144 (Elsevier, 2012).61. Haq, B. U., Hardenbol, J. & Vail, P. R. Chronology of uctuating sea levels sincethe Triassic.Science 235, 11561167 (1987).62. Butler, R. J., Benson, R. B. J., Carrano, M. T., Mannion, P. D. & Upchurch, P. Sealevel, dinosaur diversity and sampling biases: investigating the common causehypothesis in the terrestrial realm. Proc. R. Soc. Lond. B 278, 11071113 (2011).63. Raup, D. M. Systematists follow the fossils. Paleobiology3, 328329 (1977).64. Scortese, C. R. PALEOMAP Paleogeographic Atlas. 22 (PALEOMAP. ProgressReport #90, Department of Geology, University of Texas at Arlington, 1997).65. Myers, C. E. & Lieberman, B. S. Sharks that pass in the night: usingGeographical Information Systems to investigate competition in the CretaceousWestern Interior Seaway. Proc. Biol. Sci. 278, 681 (2010).NATURECOMMUNICATIONS|DOI:10.1038/ncomms8980 ARTICLENATURECOMMUNICATIONS | 6:7980| DOI: 10.1038/ncomms8980| www.nature.com/naturecommunications 7&2015MacmillanPublishersLimited.Allrightsreserved.66. Myers, C. E., MacKenzie, R. A. & Lieberman, B. S. Greenhouse biogeography:the relationship of geographic range to invasion and extinction in theCretaceous Western Interior Seaway. Paleobiology39, 135148 (2013).67. Foote, M. Origination and extinction components of taxonomic diversity:general problems. Paleobiology 26, 74102 (2000).68. Foote, M. Origination and extinction through the Phanerozoic: a newapproach. J. Geol. 111, 125148 (2003).69. Rode, A. L. & Lieberman, B. S. Integrating evolution and biogeography: a casestudy involving Devonian crustaceans. J. Paleontol. 79, 267276 (2005).AcknowledgementsWe thank Matthew Carrano, Richard Butler and various other data compilers from thePaleobiology Database. This is Paleobiology Database ofcial publication 232. We thankStephen Brusatte, Michael Benton, Paul Wignall, Nick Priest and Richard Butler foroffering valuable critical advice on earlier drafts of this manuscript. A.M.D. is supportedby a Royal Commission for the Exhibition of 1851 Fellowship. M.A.W. thanks theLeverhulme Trust (F/00351/Z) and NERC (NE/K014951/1) for support.AuthorcontributionsA.M.D. designed the project and compiled the data; M.A.W. wrote scripts; A.M.D. andM.A.W. analysedthe data and wrote the manuscript.AdditionalinformationSupplementary Informationaccompanies this paper at http://www.nature.com/naturecommunicationsCompeting nancial interests: The authors declare no competing nancial interests.Reprints and permission informationis available online at http://npg.nature.com/reprintsandpermissions/How to cite this article: Dunhill, A. M. et al. Geographic range did not confer resilienceto extinction in terrestrialvertebrates at the end-Triassiccrisis. Nat. Commun. 6:7980doi: 10.1038/ncomms8980 (2015).ARTICLE NATURECOMMUNICATIONS|DOI:10.1038/ncomms89808 NATURECOMMUNICATIONS | 6:7980| DOI: 10.1038/ncomms8980| www.nature.com/naturecommunications&2015MacmillanPublishersLimited.Allrightsreserved.