Adaptation of the vertebral inner structure to an aquatic life in snakes: Pachyophiid peculiarities in comparison to extant and extinct forms

(1)

HAL Id: hal-02430856

https://hal.sorbonne-universite.fr/hal-02430856

Submitted on 7 Jan 2020

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Adaptation of the vertebral inner structure to an aquatic life in snakes: Pachyophiid peculiarities in

comparison to extant and extinct forms

Alexandra Houssaye, Anthony Herrel, Renaud Boistel, Jean-Claude Rage

To cite this version:

Alexandra Houssaye, Anthony Herrel, Renaud Boistel, Jean-Claude Rage. Adaptation of the vertebral inner structure to an aquatic life in snakes: Pachyophiid peculiarities in comparison to extant and ex- tinct forms. Comptes Rendus Palevol, Elsevier, 2019, 18 (7), pp.783-799. �10.1016/j.crpv.2019.05.004�.

�hal-02430856�

(2)

ContentslistsavailableatScienceDirect

Comptes Rendus Palevol

www . s c ie n c e d i r e c t . c o m

General Palaeontology, Systematics and Evolution (Vertebrate Palaeontology)

Adaptation of the vertebral inner structure to an aquatic life in snakes: Pachyophiid peculiarities in comparison to extant and extinct forms

Adaptation de la structure interne vertébrale à la vie aquatique chez les serpents : particularités des achyophiides en comparaison des formes actuelles et éteintes

Alexandra Houssaye

^a,∗

, Anthony Herrel

^a

, Renaud Boistel

^a,b

, Jean-Claude Rage

^c

aUMR7179CNRS,Département“AdaptationsduVivant”,Muséumnationald’histoirenaturelle(MNHN),57,rueCuvier,CP55,75005 Paris,France

bIPHEP-UMRCNRS6046,UFRSFA,UniversitédePoitiers,40,avenueduRecteur-Pineau,86022Poitiers,France

cUMR7207CNRS,SorbonneUniversité,Département“OriginesetÉvolution”,Muséumnationald’histoirenaturelle(MNHN),57,rue Cuvier,CP38,75005Paris,France

a rt i c l e i n f o

Articlehistory:

Received22March2019

Acceptedafterrevision29May2019 Availableonline15November2019 HandledbyMichelLaurin

Keywords:

Snake Vertebra Microanatomy Aquaticlifestyle Osteosclerosis Pachyophiids

a b s t r a c t

Bonemicroanatomyappearsstronglylinkedwiththeecologyoforganisms.Inamniotes, bone mass increaseis amicroanatomical specializationoften encounteredin aquatic taxaperforminglongdivesatshallowdepths.Althoughpreviousworkhighlightedthe rathergeneralistinnerstructureofthevertebraeinsnakesutilisingdifferenthabitats, microanatomicalspecializationsmaybeexpectedinaquaticsnakesspecialisedforasin- gleenvironment.Thepresentdescriptionofthevertebralmicroanatomyofvariousextinct aquaticsnakesbelongingtotheNigerophiidae,Palaeophiidae,andRusselophiidaeenables towidenthediversityofpatternsofvertebralinnerstructureinaquaticsnakes.Alarge-scale comparativeanalysiswithextantsnakes,includingnumeroussemi-aquaticandaquatic forms,andadditionalextincttaxa,highlightsthat,evenforsnakesspecialisedforasingle environment,vertebralmicroanatomydoesnotcorrelatewellwiththeecology.Thus,it cannotbeusedasaproxyforecologicalinferencesinsnakes.Inaddition,thestudyempha- sizesthestrongdifﬁcultyincharacterizing“osteosclerosis”insnakevertebrae.Finally,it pointsoutanddiscussesthepeculiarityofthemarinehind-limbedsnakes,theonlysnakes showingpachyosteosclerosis.

0/).

Motsclés: Serpent Vertèbre Microanatomie

r é s um é

Lamicroanatomieosseuseapparaîtfortementliéeàl’écologiedesorganismes.Chezles amniotes,l’augmentationdelamasseosseuseestunespécialisationmicroanatomique souventrencontréechezlestaxonsaquatiquesquiréalisentdelonguesplongéesàfaible profondeur.Silestravauxantérieursontmisenévidencelecaractèreplutôtgénéralistede

∗ Correspondingauthor.

E-mailaddresses:[email protected](A.Houssaye),[email protected](A.Herrel),[email protected](R.Boistel).

https://doi.org/10.1016/j.crpv.2019.05.004

creativecommons.org/licenses/by-nc-nd/4.0/).

(3)

Modedevieaquatique Ostéosclérose Pachyophiidés

lastructureinternedesvertèbreschezlesserpentsquisemeuventdansdivershabitats,on peuts’attendreàobserverdesspécialisationsmicroanatomiqueschezlesserpentsaqua- tiquesquisedéplacentdansunseulmilieu.Laprésentedescriptiondelamicroanatomie vertébraledediversserpentsaquatiqueséteintsappartenantauxNigerophiidae,Palaeophi- idaeetRusselophiidaepermetd’élargirladiversitéconnuedespatronsdestructureinterne vertébraledesserpentsaquatiques.Unelargeétudecomparativeavecdesserpentsactuels, incluantdenombreusesformessemi-aquatiquesetaquatiques,ainsiqued’autrestaxons éteints,souligneque,mêmepourlesserpentsappartenantàunseulmilieu,iln’yapasde correspondanceentrelamicroanatomievertébraleetl’écologie.Lacompacitévertébrale nepeutdoncpasêtreutiliséecommeunproxypourréaliserdesinférencesécologiques chezlesserpents.Deplus,l’étudesoulignelagrandedifﬁcultéqu’ilyaàcaractériser l’«ostéosclérose»chezlesvertèbresdeserpents.Enﬁn,ellesouligneetdiscutelapar- ticularitédesserpentsmarins«àpattes»,lesseulsàprésenterunepachyostéosclérose.

4.0/).

1. Introduction

Extantaquaticsnakesarerepresentedwithinseveral taxa,e.g.,Acrochordidae,Homalopsidae,Natricidae,Elapi- dae,which(non-exclusively)adaptedtoagreatvarietyof freshwaterandmarinehabitats(Ineich,2004).Adaptation toanaquaticlifeoccurredindependentlywithinthesetaxa andevenseveraltimeswithinElapidae(Ineich,2004).

Inaddition,ﬁveextinctfamiliesofsnakesadaptedto variousfreshwaterandmarineenvironments:theAnoma- lophiidae, Nigerophiidae, Pachyophiidae, Palaeophiidae, andRussellophiidae.Thesevariousextinctaquaticsnake taxaareknownfromtheCenomaniantotheend ofthe Eocene(thatis,fromabout100to34Myrago).

Variousadaptivefeatures associated withanaquatic lifeareencountered inextantaquaticsnakes, suchasa permeableskintofacilitategaseousexchange,a sublin- gual glandfor salt excretion, a streamlinedhead shape minimisingdrag,apaddle-liketailtoincreasepropulsion (Aubret and Shine,2008; Babonis and Brischoux,2012;

Ineich,2004;Segalletal.,2016,2019).However,mostof theseanatomical features are not preservedin the fos- silrecord.Conversely,onemajoradaptationtoanaquatic lifestyleisobservedinthevertebraeandribsoftheCeno- manianextinct Pachyophiidae:bone massincrease (see Houssaye,2009;RicqlèsanddeBuffrénil,2001).Inthese taxa, bone mass increase consists of pachyosteosclero- sis,i.e.thecombinationofanincreaseinperiostealbone deposits (pachyostosis), which confers to the bones a bloatedaspect, andan increasein inner bonecompact- ness(osteosclerosis;Houssaye,2013).Thisspecializationis thoughttobeassociatedwithapassivecontrolofbuoyancy andbodytrimandit isgenerallyencounteredinpoorly activeswimmersperforminglongdivesinshallowwater environments(Houssaye,2009;RicqlèsanddeBuffrénil, 2001;Taylor,2000).Pachyostosishas,however,notbeen mentionedforanyotherextantorextinctsnaketaxon.A highbonecompactnesshasbeendocumentedinvertebrae ofonlyafewextantaquaticsnakes(Houssayeetal.,2013a).

Thisraisesthequestionof:

•whetherbonemassincrease,observedinmanyamniotes that have secondarily adapted to an aquatic life (see

Houssaye,2009;Houssayeetal.,2016;Ricqlès andde Buffrénil,2001forareview),isindeedabsentinthese otherlineagesofaquaticsnakes;

•howsnakesdifferentlyadaptedtoanaquaticlifestylein termsoftheirvertebralinnerstructure.

Vertebraeandribsarethebonesthemostcommonly affectedbybonemassincreasewhenitoccurs(Houssaye, 2009). Moreover, vertebrae correspond to most of the snakeskeletonandarethefossilsnakebonesthemostfre- quentlydiscovered.Thevertebralmicroanatomy(i.e.the distributionoftheosseoustissueinthebone)ofnumerous extantsnakeshaspreviouslybeeninvestigated(Buffrénil and Rage, 1993; Buffrénil et al., 2008; Houssaye et al., 2010,2013a).Asforextinctsnakes, dataareconversely rare. Theyessentiallypertain topachyophiids (Buffrénil and Rage,1993;Houssaye,2013;Houssayeet al.,2011) andtoalesserextenttoPalaeophiidae(BuffrénilandRage, 1993;Houssayeetal.,2013b).Thecurrentstudyanalyses theinnerstructureofvertebraefromseveralextinctsnakes consideredaquaticorpossiblyaquatic.Coupledwiththe additionofdiverseaquaticsnakesinthecomparativesam- pleofextanttaxa,itwillenable:

•toillustratetheoccurrencesof bone massincrease in aquatic snakesand toobtaina broaderpicture ofthe diversityofvertebralinnerstructurepatternsinaquatic snakes;

•tocorrelate thesepatternswith ecologicalfeatures in extantformsandtopossiblymakeinferencesaboutthe paleoecologyoftheextinctsnakessampled;

•todiscussthedifferenttypesofadaptationofthever- tebraetoanaquaticlifestyleinsnakesthroughouttheir evolutionaryhistory.

2. Materialandmethods

2.1. Material 2.1.1. Fossilmaterial

Anomalophiidae, Nigerophiidae, Palaeophiidae, and Russellophiidaeareonly representedby postcranialele- ments(Rage,1984).Theirsystematicassignmentwasmade

(4)

through vertebral characters that are diagnostic within snakes.PachyophiidaearedeﬁnedasincludingPachyophis, Pachyrhachis,andalltaxamorecloselyrelatedtothesegen- erathantoextantsnakes(Leeetal.,1999).Theirstatusand phylogeneticpositionarestronglydebated(Martilletal., 2015;Palcietal.,2013a,b;Reederetal.,2015).

Concerningthefossilmaterial,taxafromthreeofthe ﬁvefamilies encompassingaquaticformsweresampled (Table1).SomedatarelativetoPachyophiidaearealready available (Buffrénil and Rage, 1993; Houssaye, 2013).

TheAnomalophiidaearerepresentedbyasinglespecies, Anomalophis(Archaeophis)bolcensisAuffenberg,1959,from thelowerEoceneofItaly,whoseremainsarescarce.

Two Nigerophiidae were sampled: the genera Nigerophis and Indophis. Nigerophiidae are known fromtheCenomanianandPalaeoceneofAfricaandAsia.

Moreover,possible nigerophiidswere discoveredin the latest Cretaceousof Indiaand in the Eoceneof Europe (Rage and Werner,1999).Three precloacalvertebrae of NigerophismirusRage,1975a,fromthePalaeoceneoutcrop Krebb deSessao,in Niger, wereanalysed (Fig.1).Their original position along the vertebral column cannot be determinedintheabsenceof comparativematerial.The depositional milieu corresponds to a shallow marine environment(Capetta;pers.com.2007).Wealsoanalysed fourvertebraeofIndophissahniiRage andPrasad,1992, fromtheMaastrichtianofNaskal,AndhraPradesh,India.

VPL/JU Unnumb.3-4 are isolated centra, so that their original position along the vertebral column cannot be determined.As forVPL/JUUnnumb.2, it isconsidereda posterior trunk vertebra and VPL/JU Unnumb.1 a mid- trunkone(JC.R.pers.obs.).Theenvironmentcorresponds toaﬂoodplain/brackishwater(Rage etal.,2004), which isinagreementwiththevertebralcharactersillustrating adaptation to an aquatic life (Rage and Prasad, 1992).

Thesebonesareparticularlysmall,sothatonlyonesection pervertebracouldbemade(exceptforVPL/JUUnnumb.

2;Fig.2).Moreover,correctsectionalplanes(seebelow) weredifﬁculttoobtain.

Within Palaeophiidae, species from the two genera Palaeophis (Maastrichtian–Eocene) and Pterosphenus (Eocene)wereinvestigated.Thesesnakes, whichdisplay various degrees of adaptation to an aquatic life, lived in marine or marginal marine waters (e.g., estuaries, deltas,lagoons,mangroves;ParmleyandDeVore, 2005).

PalaeophiscolossaeusRage,1983aandP.maghrebianusare consideredamongthePalaeophiidae,andthesespeciesare thoughttobelessadaptedtoanaquaticlife(Rage,1983b).

TheP.colossaeussectionsanalysed,whichwerepreviously broadlydescribedbyBuffrénilandRage(1993),wereavail- ableinthecollectionsoftheMNHN.Theyweremadefrom vertebraediscoveredintheLutetianofTamaguilelt,Mali.

Sections fromvertebrae of P.maghrebianus Arambourg, 1952,fromtheYpresianofthePhosphatesofMoroccowere previouslydescribedbyHoussayeetal.(2013b).Fivever- tebraeassignedtoPalaeophistyphaeusOwen,1850(Fig.3), two fromPalaeophis toliapicus Owen,1841, and one of anundeterminedPalaeophisspecies,fromtheYpresianof Prémontré,Aisne,northernFranceandEgem,Pittem,Bel- gium, were analysed. Thesetwo species are considered morehighlyadaptedtoanaquaticlifethanP.maghrebianus

(Rageetal.,2003).Wealsoanalysedtwovertebraefrom PterosphenusschuchertiLucas,1899,fromthelateEocene of Georgia,USA.This species, assumed tohave livedin estuarineorlow-salinityenvironments(Hutchison,1985;

WestgateandWard,1981),isthoughttobeoneofthemost adaptedtoanaquaticlifewithinPalaeophiidae(Rageetal., 2003).

AsforRussellophiidae,wesampledthespeciesRussel- lophistenuisRage,1975b,fromtheearlyEoceneofFrance andIndia.ThetwovertebraecomefromthelateYpresian ofCondé-en-Brie,Aisne,northernFrance.Thistaxonwas thoughttoliveinrivers,lakes,andestuaries(Rage,1983b).

Forcomparativepurposes,sectionsoftwounnumbered vertebrae of the pachyophiid snake Simoliophis roche- bruneifromLesRenardières,inCharente-Maritime,France, describedinBuffrénilandRage(1993),andthreevertebrae ofPalaeophismaghrebianusfromtheYpresianPhosphates ofMorocco,describedinHoussayeetal.(2013b),werealso analysed,butnotdescribed.

2.1.2. Extanttaxa

Our sample includes the various semi-aquatic and aquaticextantsnakesfromHoussayeetal. (2013a)and additionalsemi-aquaticandaquatictaxa(Table2;Fig.4).

Thesetaxa wereadded inorder tobetterrepresentthe diverseecologiesencounteredwithinaquaticsnakes,with deepdivers(e.g.,Hydrophisperonii,Hydrophicelegans)and shallowswimmers(most Aipysurusspecies),the unique pelagicspecies(Hydrophisplaturus),taxaoccupyingdiverse habitats (e.g., Hydrophis stokesii and Hydrophis curtus) andotherswithamorespecializedniche(Emydocephalus annulatus), live-bearers(all homalopsids)and oviparous taxa (Laticauda species),forms living almostexclusively inwater(hydrophiids,acrochordids),andsomeregularly comingontoland(Laticaudaspecies;Heatwole,1999).

Allvertebraeofextanttaxa are mid-precloacalones, where bone mass increase is supposed tobe the most intensewhenitoccurs(seeHoussaye,2013fordatarel- ativetoPachyophiidae).Fossilvertebraeareallprecloacal, but their position along the vertebral column is vari- able(seeabove)asthesamplingchoicewaslimited.The strongestvariationintheintensityofbonemassincrease inpachyophiids,however,essentiallyconcernspachyos- tosis,whereasosteosclerosisismorehomogeneousalong theprecloacalregion(Houssaye,2013).Theimpactofthis potentialbiasforcompactness comparisonsistherefore consideredlimited.

3. Methods

Various methods were used to investigate the vertebral microanatomy of our sample [for the material alreadydigitizedfor Houssayeet al.(2013a),seedetails therein]. If some classical thin sections could be made for a few specimens, virtual ones, through the use of microtomography, were preferred for most specimens, especiallybecauseof therarityofthis materialand the wishtouseanon-destructivetechnique(Tables1and2).

Classicalthinsectionsweremadeinthemid-sagittaland theneutraltransverseplanesusingstandard techniques (seeHoussayeetal.,2008).However,asIndophisvertebrae

(5)

Table1

Listofthematerialofextincttaxaanalysed.Abb.:abbreviation(usedinFig.14);Resol:resolution(in␮m).AllfossilspecimenswerescannedattheESRF, Grenoble,France.ClsandCts:compactness(in%)inlongitudinalandtransversesections,respectively;CL:centrumlength(inmm);USTL:Universitédes SciencesetTechniquesduLanguedoc,Montpellier,France;VPL:VertebratePalaeontologyLaboratory,UniversityofJammu,Jammu,India;MNHN:Muséum nationald’histoirenaturelle,Paris,France;Unnumb.:unnumbered.

Tableau1

Listedumatérieldetaxonséteintsanalysé.Abb.:abréviation(utiliséedanslaFig.14);Resol.:résolution(en␮m).Touslesspécimensfossilesontété scannésàl’ESRF,Grenoble,France.ClsetCts:compacité(en%)encoupelongitudinaleettransversale,respectivement;CL:longueurducentrum(enmm); USTL:UniversitédessciencesettechniquesduLanguedoc,Montpellier,France;VPL:VertebratePalaeontologyLaboratory,UniversityofJammu,Jammu, Inde;MNHN:Muséumnationald’histoirenaturelle,Paris,France;Unnumb.:sansnuméro.

Family Taxon Abb. Age Locality Collectionreference Resol. Cls Cts CL

Nigerophiidae Nigerophismirus Nm Palaeocene KrebbdeSessao,Niger USTLSES105 x 93.9 97.0 3.2 USTLSES106 x 81.4 98.4 3.9 USTLSES107 x 77.5 95.0 5.9

Indophissahnii X Maastrichtian Naskal,India VPL/JUUnnumb.1 x x x x

VPL/JUUnnumb.2 x x x 2.4

Palaeophiidae Palaeophiscolossaeus Pc Lutetian Tamaguilelt,Mali MNHNUnnumb.1 x x 69.8 x

MNHNUnnumb.2 x x x x

MNHNUnnumb.3 x 54.7 x 11.0

MNHNUnnumb.4 x x x x

MNHNUnnumb.A x x x x

MNHNUnnumb.B x x x x

Palaeophismaghrebianus Pm Ypresian PhosphatesofMorocco OCPDEK/GE644 x 92.1 90.4 11.6 OCPDEK/GE645 x 85.1 91.7 8.4 OCPDEK/GE646 x 82.4 88.5 11.5 Palaeophistyphaeus Pt Ypresian Prémontré,France MNHNUnnumb.A 20.2 54.4 60.9 13.8 MNHNUnnumb.B 20.2 73.3 86.4 12.8 MNHNUnnumb.C 5.06 53.7 71.5 5.3 MNHNUnnumb.D 5.06 72.5 89.6 9.9 Ypresian Egem,Belgium MNHNCBL8-13 5.06 52.4 73.3 8.1

Palaeophistoliapicus Pto Ypresian Egem,Belgium MNHNCBL3-6 5.06 70.5 85.7 6.1

MNHNCBL3-6 5.06 63.4 82.0 5.5

Palaeophissp. P Ypresian Egem,Belgium MNHNCBL16 5.06 72.7 74.9 4.3

Pterosphenusschucherti Ps LateEocene Georgia,USA MNHNUnnumb.A x 62.8 82.5 15.5

MNHNUnnumb.B x x 49.1 x

Russellophiidae Russellophistenuis Rt LateYpresian Condé-en-Brie,France MNHNUnnumb.A 5.4 82.6 99.6 3.3 MNHNUnnumb.B 5.4 79.8 98.1 2.8 Pachyophiidae Simoliophisrochebrunei Sr Cenomanian Charente-Maritime,France MNHNUnnumb.A x x 99.8 x

MNHNUnnumb.B x 92.8 x 4.9

Fig.1.Nigerophismirus.KrebbdeSessao,Niger.Palaeocene.A–B.VertebraUSTLSES105.C.VertebraUSTLSES106.D–F.VertebraUSTLSES107;inA,D, leftlateral;B,dorsal;C,F,anterior;E,ventralviews.Thescalebarsequal1mm.

Fig.1. Nigerophismirus.KrebbdeSessao,Niger.Paléocène.A–B.VertèbreUSTLSES105.C.VertèbreUSTLSES106.D–F.VertèbreUSTLSES107;envues A,D,latéralegauche;B,dorsale;C,F,antérieure;E,ventrale.Lesbarresd’échellereprésentent1mm.

Fig.2. Indophissahnii.Naskal,AndhraPradesh,India.Maastrichtian.VPL/JUUnnumb.2inA,anterior;B,leftlateral;C,ventral;andD,dorsalviews.The scalebarequals1mm.

Fig.2. Indophissahnii.Naskal,AndhraPradesh,Inde.Maastrichtien.VPL/JUsansnuméro.2envuesA,antérieure;B,latéralegauche;C,ventrale;etD, dorsale.Labarred’échellereprésente1mm.

(6)

Fig.3.Palaeophistyphaeus.Prémontré(NorthernFrance).Ypresian.MNHNUnnumb.vertebraBinA,anterior;B,dorsal;C,ventral;D,leftlateralviews.

Thescalebarrepresents5mm.

Fig.3. Palaeophistyphaeus.Prémontré(NorddelaFrance).Yprésien.MNHNVertèbrenonnumérotéeBenvuesA,antérieure;B,dorsale;C,ventrale; D,latéralegauche.Labarred’échellereprésente5mm.

wereparticularlythin,onlyonesectionpervertebracould be made for three of them. High-resolution computed tomographywasusedfornumerousextantspecimensat:

•the Steinmann Institut, University of Bonn, Germany (GEphoenix|X-ray v|tome|xs 180 and 240; resolution between 7.5 and 89.9␮m; reconstructions performed usingdatox/ressoftware);

•theMontpellierRioImaging(MRI;MicrotomographRX SkyScan1076;resolution:9.4␮m;reconstructionsper- formedusingNReconsoftware);

•the University of Poitiers, France, using a X8050-16 Viscom model (resolution: between 15.7and 34␮m;

reconstructions performed using Feldkamp algorithm withDigiCTsoftware,version1.15[DigisensSA,France]) atthelaboratoryÉtudes–Recherches–Matériaux(ERM, Poitiers,France;http://www.erm-poitiers.fr/).

Forthefossilones,weresortedtosynchrotronmicroto- mographyontheID19beamline(resolutionofeither7.5or 20.2␮m)attheEuropeanSynchrotronRadiationFacilities (ESRF,Grenoble,France);reconstructionswereperformed usingtheﬁlteredback-projectionalgorithmwiththeESRF PyHSTsoftware.

Threemeasurementsweretakenonthesectionsusing ImageJ(Abramoffetal.,2004):

•globalcompactnessintransversesection(Cts),calculated asthetotalsectionalareaminustheareaoccupiedbycav- itiesandtheneuralcanalmultipliedby100anddivided bythetotalareaminustheareaoccupiedbytheneural canal;

•globalcompactnessofthecentruminlongitudinalsec- tion (Cls), calculated as thetotal areaof thecentrum minustheareaoccupiedbycavitiesmultipliedby100 anddividedbythetotalareaofthecentrum;

•centrumlength(CL),consideredasanindicatorofsize.

Welch’st-tests(Welch,1947)wereperformedoncom- pactnessdatatotestfordifferencespendingonecological categories(semi-aquatic,aquatic,terrestrialandarboreal, fossorial).

4. Description

Thissectiondescribestheinnerstructureofthefossil vertebraesampled.Itthenproposesacomparativeanalysis

basedondatafromextantsnakesandextincttaxaprevi- ouslydescribed.

4.1. Extincttaxa

Noneofthespecimens analyseddisplaysanysignof pachyostosis.

4.1.1. Nigerophismirus

Inlongitudinalsections,therelativesurfaceoccupiedby primaryperiostealbonevariesbetweenvertebrae(Fig.5).

Indeed, primary periostealbone resorptionis restricted totheareasurrounding theneutral point (pointwhere thecentrumgrowthstarts;sensuBuffréniletal.,2008)in USTLSES105andespeciallyUSTLSES106(Fig.5A)sothat remodellingalmostexclusivelyoccurs intheendochon- dralterritory.However,periostealboneresorptionismore intenseinUSTLSES107,whereawidecavityoccupiesthe coreofthecentrum(Fig.5C).Compactnessisgenerallyrel- ativelyhigh(78%<Cls<94%).Cavitiesarerandomlyshaped sothatthereisnotruetrabecularnetworkintheendochon- dralterritory.Inthethreevertebrae,remainsofcalciﬁed cartilageareimportantinthecoreoftrabeculaecloseto theneutralpoint(Fig.5D).

In transverse sections, compactness is very high (95%<Cts<98%).Remodellingisextremelylimited.Lack- inginvertebraeUSTLSES105-106(Fig.5B),itisobserved fromthecoreofthecentrumtothebaseoftheneuralcanal inUSTLSES107.

4.1.2. Indophissahnii

Inlongitudinalsections,variouspatternsareobserved.

Somevertebrae (VPL/JU Unnumb.2-3; Fig.6A) are very compactandshowastronginhibitionofprimaryperiosteal bone resorption. Conversely, another vertebra (VPL/JU Unnumb.4; Fig. 6B) appears highly remodelled in both endochondralandperiostealterritories.

Intransversesections,vertebraeappearverycompact (Fig. 6C). The neural canal is wide and surrounded by a uniquelayeralmost exclusivelyconsisting ofprimary periostealbone,exceptatthebaseoftheneuralcanaland aroundsomesmallcavities.

4.1.3. Palaeophiscolossaeus

Inmostlongitudinalsections,bothendochondraland periostealterritoriesareoccupiedbyaremodelledspon- giosa (Fig. 7A), whereas the primary periosteal bone

(7)

Table2

Listofthematerialofextanttaxaanalysedinthisstudy.Abb.:abbreviation(usedinFig.14);EC:ecologicalcategory;SA:semi-aquatic;EA:essentially aquatic;habitatdataessentiallyfromHeatwole(1999);IneichandLaboute(2002),GibbonsandDorcas(2004),Murphy(2007),PauwelsandVandeWeghe (2008),andDas(2015);␮CT:␮CTused;M:Montpellier;P:Poitiers;B:Bonn(seeMethod);Resol:resolution(in␮m).ClsandCts:compactness(in%)in longitudinalandtransversesections,respectively;CL:centrumlength(inmm).

Tableau2

Listedumatérieldetaxonsactuelsanalysésdanscetteétude.Abb.:abréviation(utiliséedanslaFig.14);CE:catégorieécologique;SA:semi-aquatique; EA:essentiellementaquatique;donnéessurleshabitatsessentiellementissuesdeHeatwole(1999);IneichetLaboute(2002),GibbonsetDorcas(2004), Murphy(2007),PauwelsetVandeWeghe(2008)etDas(2015);␮CT:␮CTutilisé;M:Montpellier;P:Poitiers;B:Bonn(voirMéthode);Resol:résolution (en␮m).ClsetCts:compacité(en%)encoupeslongitudinaleettransversale,respectivement;CL:longueurducentrum(enmm).

Family Taxon Abb. EC Habitat/Ecology Collectionreference ␮CT Resol. Cls Cts CL

Boidae Eunectesmurinus Em SA Freshwater MNHNAC1893197 x x 69.1 73.8 9

MNHNAC1940353 x x 69.1 x 14.3

MNHNSQ-Vert9 x x 58 78.4 15.8

Acrochordidae Acrochordusjavanicus Aj SA Freshwater MNHNSQ-Vert14 x x 66.8 77.5 8.4

AHS0004 M 9.4 48.1 56 4.3

Acrochordusgranulatus Ag EA Mangrovemudﬂats ZRC2.2334 P 15.7 82.0 97.5 2.3

Viperidae Agkistrodonpiscivorus Ap SA Water’sedge MNHN19903854 x x 68 64.3 7.5

Grayiinae Grayiaornata Go EA Freshwater AHS0006 M 9.4 60.8 x 6.3

Natricinae Amphiesmastolatum As SA Freshwater ZFMK18169 B 8.3 58.4 95.5 4.5

Xenochrophispiscator Xp SA Freshwater ZFMK74287 B 9.2 55.8 80.8 3.9

Afronatrixanoscopus Aa SA Freshwater ZFMK65488 B 9.1 66.8 81.1 4.1

Natriciteresfuliginoides Nf SA Freshwater AHS0010 M 9.4 67.1 89.1 2.7

Natrix Nn SA Water’sedge MNHNAC1874535 x x 73 x 5.3

ZFMK64057 B 6.0 65.4 93 3.7

Natrixtessellata Nt SA Freshwater ZFMK24680 B 25.7 70.9 84.1 4.0

Homalopsidae Myrrophischinensis Mc SA Freshwater ZRC2.4805 P 34.0 67.5 86.3 4.3

Hypsiscopusplumbea Hp SA Freshwater ZFMK44891 x x 70 79.9 2.9

Enhydris Ee SA Freshwater ZRC2.5507b P 15.7 87.9 98.5

Subsessorbocourti Sb SA Freshwater MNHN19998361 x x 78.9 87.7 3.8

Erpetontentaculatum Et SA Freshwater AHS0012 B 7.5 93.0 94.8 1.8

AHS0012 B 7.5 80.4 77.9 1.8

AHS0012 B 31.9 81.1 80.5 3.9

AHS0012 B 31.9 79.6 75.2 3.9

AHS0012 B 31.9 89.8 89.0 3.3

Phytolopsispunctata Pp SA Freshwater ZRC2.3554 P 15.7 66.6 87.2 1.4

Cerberusrynchops Cr SA Mangrovemudﬂats MNHN-RA-1998.8583 B 35.3 88.2 96.0 2.4

Homalopsisbuccata Hb SA Freshwater ZRC2.6411 P 15.7 84.0 98.5 2.0

Bitiahydroides Bh EA Mangrovemudﬂats ZRC2.4374 P 20.8 76.5 98.9 2.4

Cantoriaviolacea Cv SA Mangrovemudﬂats ZRC2.3672 P 20.8 65.9 90.5 1.8

Fordonialeucobalia Fl SA Mangrovemudﬂats MNHN-RA-1912.26 B 33.2 77.4 87.6 2.8

Elapidae Laticaudalaticaudata Ll EA Marine ZFMK36425 x x 72.5 87.2 3.1

ZFMK36425 x x 66.2 74.5 2.9

Hydrophismajor Hm EA Marine MNHN19904557 B 44.8 72.1 94.2 4.3

Hydrophisperonii Hpe EA Marine ZRC2.2018 P 17.0 86.1 95.9 3.0

Hydrophisornatus Ho EA Marine MNHN-RA-1994.6997 B 36.0 61.6 87.4 3.9

Hydrophisjerdoni Hj EA Marine ZRC2.2105 P 17.0 77.0 96.1 3.1

Hydrophisgracilis Hg EA Marine ZRC2.2155 P 20.8 96.0 99.6 1.9

Aipysurusduboisii Ad EA Marine MNHN-RA-1990.4519 B 41.0 74.1 93.5 4.6

Aipysuruseydouxii Ae EA Marine MNHN-RA-0.7704 B 40.2 66.9 91.0 3.6

Aipysuruslaevis Al EA Marine MNHN19904506 B 89.9 78.0 87.2 5.0

MNHN19904506 B 60.9 71.4 79.1 6.1

Hydrophiscurtus Hc EA Marine ZRCuncat P 34.0 77.7 92.3 2.8

Hydrophiselegans He EA Marine MNHN-RA-0.1879 B 30.7 81.3 94.0 2.9

Hydrophisstokesii Hs EA Marine ZRC2.2032 P 20.8 74.7 87.3 1.8

Hydrophisschistosus Hsc EA Marine ZRC2.2043 P 20.8 82.7 93.1 2.2

Hydrophisplaturus Hpl EA Marine ZFMK36436 x x 54.9 x 2.3

AHS0014 M 9.4 65.7 82.9 4.2

displaying radially oriented vascular canals (like in all palaeophiids)isrestrictedtotheventraledgeofthecen- trum and the upper part of the cotyle. Intertrabecular spacesareirregularlyshapedandrandomlyoriented.They aremuchsmallertowardtheepiphyses.Globalcompact- nessindicesof57.3%and69.4%werecalculatedforthetwo sectionswithacompletecentrum.Onelongitudinalsection (Fig.7B)displaysarelativelylimitedresorptionofprimary periostealbone,sothatcompactnessishigher.

In transverse sections, compact bone generally consists of two layers surrounding the neural canal and

the periphery of thevertebra (Fig.7C and E).They are connectedbyaspongiosathatoccupiesmostofthesec- tion.Cavitiesarerandomlyshapedanddistributed.Inthe biggest specimens, they are much more numerous and relatively smaller than in the smallest ones, conferring a honeycombstructure(Fig.7D),which isinagreement withthedescriptionofBuffrénilandRage(1993).Aglobal compactness index of 69.8% was calculated (for MNHN Unnumb.1).Inthebiggestsections,remodellingisvery limited at the base of the neural arch causing a sharp interruptionofthestructurein“double-ringsenclosinga

(8)

Fig.4. Consensusphylogenetictreeincludingtheextanttaxasampled,fromFigueroaetal.(2016),withindicationabouttheirecology:green,occasionally aquatic;blue,semi-aquatic;red,essentiallyaquatic.

Fig.4. Arbrephylogénétiquecompositeincluantlestaxonsactuelséchantillonnés,d’aprèsFigueroaetal.(2016),avecindicationdeleurécologie:vert, occasionnellementaquatique;bleu,semi-aquatique;rouge,essentiellementaquatique.

spongiosa”(Fig.7D).Furthermore,onesection(Fig.7E)dis- playsaparticularlythicklayerofprimaryperiostealbone surroundingitsperiphery.Thissectionappearsthuschar- acterizedbyarelativeinhibitionofprimaryperiostealbone resorption.

The differences highlighted in both longitudinal and transversesectionssuggestsigniﬁcantintraspeciﬁcand/or intracolumnarvariabilityinthistaxon.

4.1.4. Palaeophistyphaeus

Allvertebraedisplayawidecavityinthecoreofthecen- trum(Fig.8).However,thethicknessoftheventrallayer ofcompactcortexandthetightnessofthespongiosavary betweenthevertebraeinlongitudinalsection,fromathick compactcortex(Fig.8A)toathinnercortexsurrounding aloosespongiosa(Fig.8C).Intransversesection,whereas somevertebraeshowa compactneuralarchandneural spine(Fig.8B),othersshowahollowneuralspinesome- timesassociatedwithlargecavitiesintheventralpartof theneuralarch(Fig.8D).Asaresult,somevertebraeare much morecompact(MNHNUnnumb.A-B)than others (MNHNUnnumb.C-D).

4.1.5. Palaeophistoliapicus

ThestructureobservedinP.toliapicus(Fig.9)isvery similartothatpreviouslydescribedinP.typhaeus.

4.1.6. Palaeophissp.

ThisisalsothecaseforthevertebraofPalaeophissp.

(Fig.10).

4.1.7. Pterosphenusschucherti

Thetwolongitudinalsectionsdisplaysimilarfeatures.

Like in the other palaeophiids, a wide cavity occupies the core of the centrum, whereas the rest of the sec- tionconsistsofaspongiosa (Fig.11).Thelatteris loose in the periosteal territory,with wide randomly shaped intertrabecular spaces, but much tighter in the endochondralterritory,withrelativelysmallnumerouscavities (Fig.11A).Intransversesection,MNHNUnnumb.Adisplays arelativelyfeeblyremodelledstructure.Wideresorption cavitiesarerestrictedtotheneuralspineandthecoreof thecentrum,whereastherestofthesectionisrelatively compact,withrathersmallandscarcecavities(Fig.11B).

Conversely, resorptionis much more intense in MNHN Unnumb.B,whichdisplaysthestructurein“double-rings connectedbyveryfewtrabeculae”characteristicofextant squamates(Houssayeetal.,2010;Fig.11C).Compactness isthusmuchlowerinthelatter(Cts=49.1%versus82.5%).

4.1.8. Russellophistenuis

Inlongitudinalsections,vertebraearehighlycompact (meanCls=81.2%). Periosteal bone remodelling appears veryinhibited.Thereisnotruetrabecularnetworkinthe endochondralterritory,but afewwidesub-sagittalcav- itiesoccur(Fig.12A–B).Vascularization appears lacking inprimaryperiostealbone.Intransversesections,cavities arealmostlacking.Compactnessisthusveryhigh(mean Cts=98.9%).Theneuralcanalisratherwideinthisspecies, likeinIndophis.

(9)

Fig.5.Nigerophismirus.KrebbdeSessao,Niger.Palaeocene.A,B.VertebraUSTLSES106;A,longitudinalsection(LS)ofthecentruminpolarizedlight (PL);arrowsindicatethelimitbetweenperiostealandendochondralterritories;B,halftransversesectionshowingtheabsenceofremodelling.Scalebars:

500␮m.C,D.vertebraUSTLSES107;C,longitudinalandhalftransversesectionsinnaturallight(NL);scalebars:1mm;D,remainsofcalciﬁedcartilage (indicatedbyarrows)insidethetrabeculaeofendochondralorigininLS;scalebar:500␮m.

Fig.5. Nigerophismirus.KrebbdeSessao,Niger.Paléocène.A,B.VertèbreUSTLSES106;A,coupelongitudinale(LS)ducentrumenlumièrepolarisée (PL);lesﬂèchesindiquentlalimiteentrelesterritoirespériostiquesetendochondraux;B,demi-coupetransversaleillustrantl’absencederemaniement.

Barresd’échelle:500␮m.C,D.VertèbreUSTLSES107;C,coupeslongitudinaleetdemi-transversaleenlumièrenaturelle(NL);échelle:1mm;D,restes decartilagecalciﬁé(pointépardesﬂèches)àl’intérieurdestravéesd’origineendochondraleenLS;échelle:500␮m.

Fig.6.Indophissahnii.Naskal,AndhraPradesh,India.Maastrichtian.A.VPL/JUUnnumb.3.B.VPL/JUUnnumb.4.C.VPL/JUUnnumb.1;A–B,longitudinal sectionsinNL;C,transversesectioninPL.Thescalebarsequal300␮m.

Fig.6. Indophissahnii.Naskal,AndhraPradesh,Inde.Maastrichtien.A.VPL/JUsansnuméro.3.B.VPL/JUsansnuméro.4.C.VPL/JUsansnuméro.1;A–B, coupeslongitudinalesenNL;C,coupetransversaleenPL.Barresd’échelle:300␮m.

(10)

Fig.7.Palaeophiscolossaeus.Tamaguilelt,Mali.Lutetian.A–B.LongitudinalsectionsofA,MNHNUnnumb.3andB,MNHNUnnumb.AinNL.C–E.Vertebral transversesections.A,MNHNUnnumb.4;B,MNHNUnnumb.1;C,MNHNUnnumb.B.ThescalebarsequalA,C,1mm,B,D,2mm,E,1.5mm.

Fig.7. Palaeophiscolossaeus.Tamaguilelt,Mali.Lutétien.A–B.CoupeslongitudinalesdeA,MNHNUnnumb.3etB,MNHNUnnumb.AenNL.C–E.Coupes transversales.A,MNHNNsansnuméro.4;B,MNHNsansnuméro.1;C,MNHNsansnuméro.B.Lesbarresd’échellereprésententA,C,1mm,B,D,2mm,E, 1,5mm.

4.2. Comparisonswithextanttaxa

Manyextantaquaticsnakesdisplayahighinnercom- pactness. Average compactness values for the extant semi-aquaticandaquaticformsare72.9%(Cls)and86.9%

(Cts).Therearenostrikingdifferencesbetweenoccasion- ally aquatic and semi-aquatic species on the one hand (mean Cls=72.1%;mean Cts=84.1%)and predominantly aquaticones(meanCls=73.9%;meanCts=90.7%),though thedifferenceintransversesection,madeattheneutral point,indicatesahigherthicknessofthecompactcorti- cal bone in thelatter. Calculations based ondata from

Houssayeetal.(2013a)indicatelowervaluesforterres- trial(generalistandarboreal)taxa(meanCls=71.3%;mean Cts=76.5%),whereasfossorialspeciesshowhighcompact- nessvalues(meanCls=74.9%;meanCts=86.0%),closeto those observed in aquatic snakes (Fig.13).There is no signiﬁcantdifferenceinthecompactnessinlongitudinal sectionforthesefourecologicalcategories(semi-aquatic, aquatic,terrestrial and arboreal, fossorial;P=0.60 for a Welch’st-test),but itissigniﬁcantintransversesection (P<0.001).

Comparisons ofthecompactness valuesobtainedfor theextinctandextanttaxafromoursampleandSimoliophis

(11)

Fig.8.Palaeophistyphaeus.Prémontré(NorthernFrance).Ypresian.A,B.MNHNUnnumb.B.C–D.MNHNUnnumb.C;A,C,longitudinalvirtualsectionof thecentrum;B,D,transversevirtualsection.Thescalebarsequal1mm.

Fig.8. Palaeophistyphaeus.Prémontré(NorddelaFrance).Yprésien.A,B.MNHNsansnuméro.B.C–D.MNHNsansnuméro.C;A,C,coupevirtuelle longitudinaleducentrum;B,D,coupetransversalevirtuelle.Lesbarresd’échellereprésentent1mm.

andPalaeophismaghrebianussections(seeTable1)reveal that all fossil specimens sampled exhibit compactness indicesintherangeofthoseobservedintheextantaquatic snakes(Fig.14).Moreover,nocleardistinctionisobserved between semi-aquatic and almost exclusively aquatic forms(Fig.13).Palaeophiidsdisplayawiderangeofcom- pactnessvalues with someparticularly light forms and thePalaeophismaghrebianusspecimensexhibitingrather high compactness values. Nigerophis, Russelophis and Simoliophisexhibithighcompactnessvalues,buttheyare notdistinctfromsomeextantsnakes(e.g.,Enhydris,Cer- berusrynchops,Homalopsisbuccata).Whatisofparticular interestisthatthesesnakesdisplayvariousecologies,from mangrovemudﬂats forAcrochordusgranulatus,Cerberus rynchops,andBitiahydroides,toapurelymarinelifestyle,

forHydrophis peronii,H.gracilis,H.elegans,H.schistosus, H.stokesii,andH.curtus,throughfreshwaterenvironments forEnhydris,Erpetontenteculatum,andHomalopsisbuccata.

5. Discussion

5.1. Histologicalandmicroanatomicalfeaturesofthe extincttaxa

Nigerophis.Vertebraearecharacterizedbytheinhibition ofbothperiostealboneandcalciﬁedcartilageresorption, whichevokesosteosclerosis.WhereasvertebraeUSTLSES 105-106areconsideredmid-trunkvertebrae,USTLSES107 couldcorrespondtoamoreanteriortrunkvertebra(JCR;

pers.obs.)Thevariationincompactnessobservedbetween

(12)

Fig.9.Palaeophistoliapicus.Egem(Belgium).Ypresian.MNHNCBL3-6.A,C.Longitudinalvirtualsectionofthecentrum.B,D.Transversevirtualsection.

Thescalebarsequal1mm.

Fig.9. Palaeophistoliapicus.Egem(Belgique).Yprésien.MNHNCBL3-6.A,C.Coupelongitudinalevirtuelleducentrum.B,D.Coupetransversalevirtuelle.

Lesbarresd’échellereprésentent1mm.

thesevertebraeisconsistentwithamaximalosteosclerosis intensityinthemid-trunkregion.BuffrénilandRage(1993) described the inner organization of specimens of both NigerophismirusandPalaeophiscolossaeusasahoneycomb weavestructurecharacterizedbyanintenseremodelling activity.AsforN.mirus,thisresultis notin accordance withthesectionsanalysedhere.Noillustrationofthese sectionsisavailableintheliterature;however,two lon- gitudinalsectionsassigned tothistaxonwereloanedby V.deBuffrénil.Themicroanatomyisrelativelycomparable towhat isobservedin vertebraUSTLSES107. Unfortu- nately,nocastorillustrationoftheoriginalvertebraeis

available,sothatnoinferenceabouttheoriginalposition ofthesevertebraealongthevertebralcolumncanbemade.

Indophis. Osteosclerosis has been observed in most sectionsanalysed, although VPL/JU Unnumb. 4 displays a pattern similar to that of most extant squamates.

Osteosclerosis in this taxon is probably restricted to a peculiar region of thevertebral column that cannot be determinedbasedonthissample.

Palaeophiidae.NoneofthePalaeophiscolossaeusverte- braedisplaysosteosclerosis.Theirinnerstructureisrather spongious.Thevariouspatternsobservedshowanimpor- tantintraspeciﬁcvariationinvertebralmicroanatomy.Like

(13)

Fig.10.Palaeophissp.Egem(Belgium).Ypresian.MNHNCBL16.A.Longitudinalvirtualsectionofthecentrum.B.Transversevirtualsection.Thescalebars equal1mm.

Fig.10. Palaeophissp.Egem(Belgique).Yprésien.MNHNCBL16.A.Coupelongitudinalevirtuelleducentrum.B.Coupetransversalevirtuelle.Lesbarres d’échellereprésentent1mm.

Fig.11. Pterosphenusschucherti.Georgia,USA.LateEocene.MNHNUnnumb.A.LongitudinalsectioninNL.Thescalebarequals3mm.B–C.Halftransverse sectionsinNL.B,MNHNUnnumb.A;C,MNHNUnnumb.B.Thescalebarsequal2mm.

Fig.11. Pterosphenusschucherti.Géorgie,États-Unis.Éocènesupérieur.MNHNsansnuméro.A.CoupelongitudinaleenNL.Labarred’échellereprésente 3mm.B–C.Demi-coupestransversalesenNL.B,MNHNsansnuméro.A;C,MNHNsansnuméro.B.Lesbarresd’échellereprésentent2mm.

forP.colossaeus,novertebraofPalaeophistyphaeusdisplays osteosclerosis,thoughsomevertebraeshowcompactneu- ralarchesandspines,whereasothersdonot,reﬂectinga ratherhighintraspeciﬁcvariability.Thesevertebrae,however,alldisplaya largecentralcavity inthecoreof the centrum.TheP.toliapicusandPalaeophissp.innerstruc- tureappears very similartothat of P.typhaeus. Thisis also thecase for Pterosphenus. The endochondral terri- toryinthePalaeophisvertebraeconsistsinaratherloose spongiosa,which becomes tighterin some(butnot all)

ofthelargestvertebrae.WithinPalaeophiidae,important intraspecificvariationinmicroanatomicalorganizationis observed,whichprobablyreflectsimportantintracolum- narand sizevariation inthevertebralmicroanatomyof thesetaxa.Furtheranalysesonvertebraeofvarioussizes illustrating diverse clearly defined positions along the vertebralcolumnwouldberequiredtoexplainthisvari- ability.

MostPalaeophisvertebrae,exceptthoseofP.colossaeus, show an open cavity in the core of the centrum, even