A revision of the trionychid turtle Procyclanorbis sardus Portis, 1901
from the late Miocene of Sardinia (Italy)
Une révision de la tortue trionychidé Procyclanorbis sardus Portis, 1901 du
Miocène supérieur de Sardaigne (Italie)
Georgios L. Georgalis
a,b,∗, Daniel Zoboli
c, Gian Luigi Pillola
c, Massimo Delfino
b,daDepartment of Geosciences, University of Fribourg, 6, chemin du Musée, 1700 Fribourg, Switzerland
bDipartimento di Scienze della Terra, Università di Torino, 35, via Valperga Caluso, 10125 Torino, Italy
cDipartimento di Scienze Chimiche e Geologiche, Università di Cagliari, 51, via Trentino, 09127 Cagliari, Italy
dInstitut Català de Paleontologia Miquel Crusafont, Universitat Autònoma de Barcelona, Edifici ICTA -ICP, Carrer de les Columnes s/n, Campus de la UAB,
08193 Cerdanyola del Vallès, Barcelona, Spain
Keywords: Procyclanorbis Trionychidae Sardinia Miocene Insularity
a b s t r a c t
Procyclanorbis sardus Portis, 1901 is the first fossil trionychid turtle described from Sardinia. This late Miocene taxon was originally considered to have affinities with the African and southern Asian cyclanor-bines. We here redescribe in detail the holotype specimen of this species, which has suffered severe degradation since its original publication. A comparison between the original state of the fossil and its current state of preservation is provided. On the basis of its anatomy, affinities of Procyclanorbis sardus with cyclanorbines are discarded and this taxon is demonstrated to be an indeterminate pan-trionychine. The distribution of fossil trionychids in the Mediterranean Islands is also discussed.
Mots clés : Procyclanorbis Trionychidae Sardaigne Miocène Insularité
r é s u m é
Procyclanorbis sardus Portis, 1901 est la première tortue trionychidé fossile décrite en Sardaigne. Ce taxon du Miocène supérieur a été initialement considéré comme ayant des affinités avec les cyclanorbines d’Afrique et d’Asie du Sud. Nous décrivons ici en détail et figurons l’holotype de cette espèce, qui a subi une importante dégradation depuis sa publication originale. Une comparaison entre l’état originel du fossile et son état de conservation actuel est fournie. Sur la base de son anatomie, les affinités de Procyclanorbis sardus avec les cyclanorbinés sont rejetées car ce taxon s’avère être un pantrionychiné indéterminé. La distribution des trionychidés fossiles dans les îles de la Méditerranée est également discutée.
1. Introduction
Procyclanorbis sardus Portis, 1901 is the first named fos-sil soft-shelled turtle from the island of Sardinia, Italy (Portis, 1901a). This taxon has had a problematic taxonomic history, being
∗ Corresponding author. Department of Geosciences, University of Fribourg, Chemin du Musée 6, 1700 Fribourg, Switzerland.
E-mail address:georgios.georgalis@unifr.ch(G.L. Georgalis).
originally described as a European member of pan-cyclanorbines (Portis, 1901a), later considered as a member of Amyda (Hummel, 1929; Comaschi Caria, 1959) and subsequently treated as an indeterminate pan-trionychine (Georgalis and Joyce, 2017). We restudied in detail the holotype carapace that still exists, severely damaged, in the collections of the Museo Sardo di Geologia e Pale-ontologia “Domenico Lovisato”, Cagliari, Italy (MDLCA), and we herein redescribe and attempt to revise this historical taxon reeval-uating its taxonomic status for the first time on the basis of first hand observation and with a modern approach.
http://doc.rero.ch
Published in "Annales de Paléontologie 103(2): 127–134, 2017"
which should be cited to refer to this work.
InstitutionalAbbreviations:MDLCA:MuseoSardodiGeologiae Paleontologia“DomenicoLovisato”,Cagliari,Italy.
2. Materialandmethods
TheholotypecarapaceofProcyclanorbissardusiscurated,along with its natural mold, at MDLCA under thecollection number MDLCA 14007.Otherspecimens (another carapace,a skull,and additionalplastralelements)thatwerealsooriginallyreferredto P.sardusbyPortis(1901a)couldnotbelocatedandweagreewith Kotsakis(1985)thattheywereprobablydestroyedduringWorld War II.The partialcarapace of atrionychid, alsofromthetype area,thatwassubsequentlyreferredtoP.sardusbyComaschiCaria (1959)isalsohousedinthecollectionsofMDLCAunderthenumber MDLCA14008.
3. Geologicalsetting
The CalcaridiCagliari Formation (Gandolfiand Porcu, 1967; Cherchi,1974)carbonatesuccessionisrepresentedbythreemain lithofacies which are,from thebottomto top:“PietraCantone“, “Tramezzario”and“PietraForte”.Notethatthenames,informally adoptedin themostrecent officialgeologicalmap (Barca etal., 2005),arederivedfromthenamesusedbyquarrymen.
The“PietraCantone”iscomposedofastratifiedyellow marly-arenaceous limestone with common intense bioturbation. The intermediatelithofacies(“Tramezzario“)isrepresentedbywhitish calcarenites,islocallymarlyandbearsabundantbioclastic com-ponents, and shows widespreadphenomenaof synsedimentary breccias, slumpings, erosionalsurfacesand faulting. The“Pietra Forte” representsmainlythetopof thesuccessionandconsists ofmassivecoarsebioclastic(mainlyrodalgal-mollusc) biostroma-llimestones[forfurtherdetailsonlithologyandfaciesdistribution seeBarcaetal.(2005)].
The Calcari di Cagliari Formation is considered to be late Mioceneinage(Kotsakis,1985;ZoboliandPillola,2016).The Tor-tonian/Messinianboundaryistentativelyplacedwithintheupper partofthe“PietraCantone”(CherchiA.,personalcommunication, 2016).
Theholotypecarapacewithitsinternalmold,asalsothereferred skullandplastralelementsdescribedbyPortis(1901a)andnow lost,asalsothereferredcarapaceimprintdescribedbyComaschi Caria(1959),wereallfoundintheIsMirrionisareainCagliari.The associatedreptilefaunafromIsMirrionisconsistsonlyofthetype materialofthecrocodylianTomistomacalaritanumCapellini,1890. TheentireareaofIsMirrionisandtheadjacentTuvixedduand TuvumannuhillsandSant’Avendrace,currentlyinhabitedbutwith severaloutcropsandsectionsstillcropping-out,wasintensively quarriedforbuildingmaterialssincehistoricaltimes.Thestudied materialoriginates,inmostlikelihood,fromthelower “Tramez-zario”facies;therefore,wetentativelyassignaMessinianageto theseremains.
4. Historicalbackground
Portis (1901a) originally establishedthe new species Procy-clanorbissardusonthebasisofanincompletecarapaceanditsmold fromthelateMioceneofIsMirrionis,CagliariandSardinia.The sameauthoradditionallyreferredtothesamespeciesarather com-pletebutcrushedskullandtwoplastralfragmentsfromthesame locality,as alsoanother,partialcarapacefromthelate Miocene of Sassari, nearNulvi,northernSardinia(Portis,1901a).On the basisofallthismaterial,andmostlythemorphologyofthenuchal andthecostals,Portis(1901a)consideredProcyclanorbissardusto bethefirstEuropeanmemberofPan-Cyclanorbinae,acladethat
hasextantrepresentativesonlyinAfricaandsouthernAsia,and, at that time, a poor fossil record confined tofew finds in the Indian subcontinent(Lydekker,1885, 1889).Portis (1901a)also envisagedsimilaritiesof hisnewSardinianpan-trionychidwith certainCentralEuropeanfinds,morespecificallywithTrionyx ger-gensiReinachvon,1900fromtheearlyMioceneofGermanyand TrionyxpreschenensisLaube,1900fromtheearlyMioceneofthe CzechRepublic.Hefurthermoreconsideredthisresemblanceas adequateenoughtosuggestcongenericaffinitiesbetweenthe Sar-dinian,GermanandCzechspecimens,thereforerecombiningboth T.gergensi and T. preschenensisinto hisnewgenus Procyclanor-bis,andtherebytreatingthemasthenorthernmostoccurrences ofpan-cyclanorbinesknowntothatdate(Portis,1901a).
Since then,only few mentions of Procyclanorbissardus have occurredinthechelonianliterature.Furthermore,besidessporadic simplementionsofjustthename(e.g.Bergounioux,1954;Kotsakis andPalombo,1979;ComaschiCaria,1986;Kotsakis,1989;Karl, 1999),onlyfewauthorshavedealtwiththetaxonomicaffinitiesof theSardiniantaxon.Fucini(1912)wasthefirsttoexpressdoubts onthevalidityofthegenusProcyclanorbisandstatedthatP.sardus couldonlybedifferentiatedfromTrionyxpliocenicusFucini,1912 fromthePlioceneofTuscany,Italy,onthebasisofminor morpho-logicalcharacters.Fewyears later,Hummel(1929,1932)defied Portis’s(1901a)originalidentificationoftheSardinianmaterialas acyclanorbineandheratherincludeditintoAmyda,recombining itasTrionyx(Amyda)sardus.Such subgenericassignmentwasa commonpracticeformostEuropeanfossiltrionychidsaccording toHummel(1929),whoerroneouslyalsoreferred Trionyx triun-guis to Amyda. Bergounioux (1935) mentioned the presence of theotherwise CzechtaxonTrionyxpontanusLaube,1895,inthe MioceneofSardiniawithoutprovidinganyotherinformation,but itisnowbelievedthatthisisratheranerror,andthatthisauthor intended to mean instead Trionyx sardus (Georgalis and Joyce, 2017). In the same paper, Bergounioux (1935) mentioned that P.sarduswasalsoknownfromSwitzerland,againmostprobably anerror.Comaschi Caria(1959) later describednew trionychid remainsfromtheMioceneofSardinia.Thenewfindsoriginated fromtheMiocenelocalitiesofIsMirrionis(typelocalityofP.sardus) andSant’Avendrace,withtheauthorassigningthemtoP.sardus andtheOligoceneFrenchtaxonTrionyxburdigalensisBergounioux, 1935,respectively,buttreatingbothspeciesasmembersofAmyda (ComaschiCaria,1959).Fewyearslater,inhiscompendium,Kuhn (1964) continuedto treat P.sardus as a pan-trionychineand a memberofTrionyx.Broin(1977)madeabriefmentionofP.sardus statingthatthereferredskullthatwasoriginallydescribedbyPortis (1901a)belongedinfacttoacheloniidmarineturtle.Inhisreview oftheItaliantrionychids,Kotsakis(1985)followedtheopinionof Broin(1977)thatthereferredskulldoesnotbelongtotrionychids, furthermentioningthatthisspecimenwasprobablylost(destroyed duringtheWorldWarII),andhetentativelytreatedallMiocene findsfromSardiniaaspertainingtoonespecies,P.sardus,which heconsideredas a memberof Trionyx. Delfino(2002) followed Kotsakis(1985)andreportedP.sardusasatentativevalidspeciesof Trionyx.Chesi(2009)alsoconsideredP.sardusasamemberof Tri-onyxbutnotedthatthevalidityofthistaxonshouldbetestedusing amodernsystematicapproach.HefurtherdescribednewSardinian findsfromtheearlyMiocenelocalityofNoragugume, whichhe treatedasanindeterminatepan-trionychid(Chesi,2009).Intheir reviewoftheMiocenereptileshousedattheMDLCA,Zoboliand Pillola(2016)mentionedP.sardusandprovidedanewfigureof theactualpreservationstateoftheholotypespecimen.They addi-tionallyshowedthatthecarapacereferredtoT.burdigalensisby ComaschiCaria (1959)is infact acheloniid, andthis specimen isnotlocatedinamuseum,butisawalledpartofafountainin Sant’Avendrace(Cagliari).IntheiroverviewofallOldWorldfossil pan-trionychids,GeorgalisandJoyce(2017)brieflydiscussedthe
statusofP.sardusonthebasisofitspublisheddescriptions.They consideredthatthereferredskullandplastralelementsbelongin facttocheloniids,whereas theholotypecarapaceandtheother tworeferredcarapacesofPortis(1901a)andComaschiCaria(1959) representindeterminatepan-trionychines.Assuch,Procyclanorbis sarduswasconsideredtobeanomendubium(GeorgalisandJoyce, 2017).
5. Systematicpaleontology
Class:ReptiliaLaurenti,1768 Order:TestudinesBatsch,1788 Family:TrionychidaeGray,1825
Sub-Family:Pan-TrionychinaeGeorgalisandJoyce,2017
Pan-Trionychinaeindet.(Figs.1–3) Synonymy:
1901aProcyclanorbissardusPortis:Plate1Plate1. 1912ProcyclanorbissardusPortis:Fucini,1912,p.3. 1929TrionyxsardusPortis:Hummel,1929,p.25.
1954ProcyclanorbissardusPortis:Bergounioux,1954,p.191. 1959AmydasardusPortis:ComaschiCaria,1959,p.38. 1977ProcyclanorbissardusPortis:Broin1977,p.191.
1979ProcyclanorbissardusPortis:KotsakisandPalombo,1979, p.624.
1983TrionyxsardusPortis:EsuandKotsakis,1983,p.198. 1986AmydasardaPortis:ComaschiCaria,1986,p.29.
Descriptionoftheholotype:
Portis(1901a)describedboththecarapaceanditsinternalmold, butfiguredonlytheformerspecimen.Judgingfromthepublished figure(Fig.1)andthecurrentpreservationstateofthisspecimen (Fig.2),itseemsthatithassufferedalotofdamagesinceits orig-inaldescription.Indeed,thecarapaceismuchbetterpreservedin Plate 1.1of Portis(1901a),while currentlytheposterior halfof thespecimenisalmosttotallymissing.Suchdamagewas proba-blycausedduringtheWorldWarII,althoughPortis(1901a)already mentionedthatthewholeturtlematerialfromIsMirrionishad suf-fereddamageduringthetransportfromCagliaritoTurin,wherethe authorwasbasedatthattime.
Judgingfromthepublishedfigureandtheoriginaldescription, where thespecimen appearsmore complete (Fig. 1a),it seems thattheholotypepertainstoamedium-sizedtrionychid,witha carapacelengthofabout45cm.Themarginsofthecarapace, how-ever,areuniversallynotpreserved.Especially,thelattermargin isseverelydeformed,renderingthesizeoflastcostalsambiguous. Thereisnopreneural.Thenuchalisratherenlargedandsitsanterior tothediscformedbythecostals.Therearesevenneurals.Thefirst twoneuralsarelargeandelongated,especiallyneuralIwhichis alsorelativelywide.Itisnotpossibletodeterminewhetherthere isareversalintheneuralorientation,asisthetypicalcondition
Fig.1. TheoriginalplateofPortis(1901a)withtheholotypecarapaceofProcyclanorbissardusandthereferredplastralandcranialmaterial.A.Theholotypecarapace.B.The
nowlostpartiallefthyo-hypoplastron(nowidentifiedasaprobablecheloniid).C.Thenowlostskull(nowidentifiedasaprobablecheloniid).
PlancheoriginaledePortis(1901a)avecl’holotypedeProcyclanorbissardus(Carapace)etlematérielréféré(plastronetcrâne).A.Carapace,holotype.B.Hyo-hypoplastrongauche
aujourd’huiperdu(maintenantidentifiécommeunchéloniidéprobable).C.Crâneaujourd’huiperdu(maintenantidentifiécommeunchéloniidéprobable).
Fig.2. Thecurrentstateofpreservationoftheholotypepartialcarapace(MDLCA14007)ofProcyclanorbissardus(A,C)anditsimprint(B,D).Abbreviations:Nu–nuchal,
N1–N4–neuralsI–IV,C1–C5–costalsI–V.
Étatdeconservationactueldel’holotype(carapacepartielle)(MDLCA14007)deProcyclanorbissardus(A,C)etsonempreinte(B,D).Abréviations:Nu-nucale,N1–N4-neurales
I–IV,C1–C5-costalesI–V.
forpan-trionychines(Meylan,1987).Thelastneuralmeetsatthe midlinecostalsVIandtheanteriormostpartofcostalsVII.There areeightpairsofcostals.Costalsarerelativelylarge,withcostalsII beingdistallyexpanded,whereascostalsVIIIseemtobereduced andshortandfullymeetatthemidlineofthecarapace.However, thereducedsizeofcostalsVIIIshouldbetakenwithcaution,due tothebreakageattheposteriormarginofthecarapace.The sculp-turingpatternconsistsofanetworkofridgesatthelateralmargins ofthecarapace,whereasitslightlyfadestowardsthecenterofthe shell,consistingmostlyofsmalltuberclesattheneuralregion.The sculpturinginthenuchalregionhasbeentotallyfadedoutanditis impossibletostatethenatureofthatpatternatthisregionofthe shell.
Inthecurrentandseverelydamagedform,onlytheanteriorpart ofthecarapaceispreserved(Fig.2).Assuch,onlythenuchal,the firstfourneurals,thefirstfourrightandfirstthreeleftcostalscan beobserved,whereasinitsnaturalmold,remnantsofthenuchal andthefirstfiveleftandfourrightcostalsarepreserved(Fig.2). Ofcourse,thevisiblepreservedfeaturesintheimprintshouldnot betakenintofullconsideration,astheydonotreliablyreflectthe external arrangement ofthebones (Georgalisand Joyce, 2017). Indeedthesize,shapeandinclinationofallpreservedcostalsare radicallydifferentfromthoseobservedintheactualcarapace.The
sculpturingpatterniscurrentlynotwellpreserved,anditappears thatithasseverelyfadedout(Fig.3).
6. Discussion
6.1. TaxonomicidentificationandstatusofProcyclanorbis sardus
Avast numberoftrionychidtaxahavebeennamedfromthe EuropeanMiocene,in particular fromthecentral and southern partsofthecontinent(GeorgalisandJoyce,2017).Thevalidityofthe majorityofthesespecieshasbeenrecentlyrejectedbyGeorgalis andJoyce(2017),whodemonstratedthepresenceoftwo differ-entpan-trionychidlineagesintheEuropeanMiocene,belonging totheextantpan-trionychinegeneraRafetusandTrionyx. Accord-ingtotheseauthors,onlydifferencesinthenumberandextent ofsculpturingcallositiesontheplastronisareliablecharacterfor discriminatingfossilsofthesetwolineagesbasedonshell mate-rial,andminorcarapacialcharactersamongRafetusandTrionyx, suchasdifferencesinthesizeofthelastcostals,shouldbe bet-terconsideredtentativeandvariable.Furthermore,thepresenceof pan-cyclanorbinesintheEuropeanfossilrecordwasrecently dis-carded,astheEuropeanpurportedmemberofthisclade,Trionyx
Fig.3. AcloseupofthesculpturingpatternofMDLCA14007(holotypeof
Procy-clanorbissardus).
Vuerapprochéesurl’ornementationdelacarapaceMDLCA14007(holotypede
Procy-clanorbissardus).
melitensisLydekker,1891,fromtheMioceneofMalta,wasshown tobeinfactacheloniid(Georgalisand Joyce,2017;seebelow). Therefore,onlymaterialcontainingplastralelementscouldbe ade-quatelyassignedtoeitherRafetusorTrionyx.Theidentificationof thenowlostpartialhyo-hypoplastronreferredtoP.sardusbyPortis (1901a)asaprobablecheloniid(seebelow),leavestheSardinian formasacarapacebasedonlytaxon.Forthisreason,inaddition withtheincompletenatureoftheholotype,P.sardushastobe consideredtobea nomendubium,pertaining anindeterminate pan-trionychine.
Portis(1901a)envisagedhisnewspeciesProcyclanorbissardus aspertainingtocyclanorbines.Henotedstrongresemblancewith theextantAfricancyclanorbinegeneraCyclanorbisandCycloderma, whereasamongextincttaxa,Procyclanorbissarduswasmost simi-larwithTrionyxgergensiandT.preschenensis,forwhichheformally suggestedcongenericaffinitieswithhisnewSardinianform(Portis, 1901a).However,allsuchaffinitieswereproposedonthebasis ofhighlyvariablecharacters,suchasthesculpturingpatternand theshape and size ofcostals andneurals (Meylan,1987;Vitek andJoyce,2015).Furthermore,thetypecarapaceofP.sarduscan bereadilyexcludedfromPan-Cyclanorbinaebytheabsenceofa preneuralandthelackofsplitcostiformprocessesonthenuchal (Meylan,1987).Additionally,thecarapacialsculpturingofP.sardus isnotsoprominentasthatofextantcyclanorbines.Furthermore, thesuggestionofHummel(1929,1932)andComaschiCaria(1959) thattheSardiniantaxonbelongstoAmyda,isalsoattributedto highlyhomoplasticandvariablecharactersandthelattergenus should be confined only to Asian forms. In particular, judging fromthecarapacemorphology,itseemsmostprobablethatP. sar-dusbelongstothesamelineagewithtrionychines,althoughthe absenceofplastralmaterialpreventsanydefiniteconclusion.Exact
affinitieswiththethreevalidpan-trionychidtaxafromthe Neo-geneofEurope,Rafetusbohemicus(Liebus,1930),fromtheMiocene of theCzechRepublic, Trionyxvindobonensis Peters, 1855,from theMioceneofcentralandwesternEurope,andTrionyx plioceni-cusFucini,1912,fromthePlioceneofItaly,cannotbemadewith certaintyduetotheabsenceofplastralmaterialfortheSardinian taxon.Aswasstatedabove,theholotypecarapaceofP.sardusseems tobearrathersmallcostalsVIII,acommonfeatureoftheRafetus lineage. However,thedamagingof theposteriormarginsofthe carapacehinderstheexactshapeandsizeoftheseelements,and wearethereforereluctanttomakeanygenericassignmentofthe Sardinianform.Assuch,thefactthatthereisnoreliableplastral material,inadditionwiththeincompletenatureoftheholotype and its unfortunatesubsequentseveredamaging, prompt usto considerProcyclanorbissardustobeanomendubium.
RegardingtheskullthatwasoriginallyreferredtoProcyclanorbis sardusbyPortis(1901a),thisspecimenisnowlost,butitwasatleast figured.Broin(1977)andsubsequentlyEsuandKotsakis(1983)and Kotsakis(1985)consideredthatthisspecimendoesnotbelongto pan-trionychids,butinsteadithascheloniidaffinities.Indeed, judg-ingfromthepublishedimageofthisspecimen(Fig.1c),itseems thattheskulldidnotbelongtopan-trionychids:itsbasicranium appearstobeextremelyslender(andnotbroadasinmost pan-trionychids)andpossiblyalso,prepalatineforaminaarepresent (whicharetotallyabsentinpan-trionychids).
TheplastralfragmentsthatwereoriginallyreferredbyPortis (1901a)toProcyclanorbissarduscorrespondtoapartialleft hyo-hypoplastron.Thismaterialisalsolost,probablyduringtheWorld War II. In any case, judging from the published figure of the originalpublication(Fig.1b),andtheshapeandthesizeofthe hyo-hypoplastron,itseemsthattheseelementsalsopertaintoamarine turtle. Indeed, even in the original description, Portis (1901a) admittedthatatfirstglancethisplastralmaterialseemedto per-taintoacheloniid,butafterhissubsequentstudyhedenotedstrong resemblancewiththeplastronoftheextantAfricancyclanorbines CyclanorbisandCycloderma.Pan-cyclanorbinehyo-hypoplastraare characterizedbytheirfusionsoonafterhatchling(Meylan,1987). Portis(1901a)alsostatedthepresenceofsculpturingonthe hyo-hypoplastron,thoughthischaractercouldnotbeevaluatedinthe accompanying imageof thespecimen.Onthebasis of theonly existingandpoorqualityfigure,theplastralelementsreferredto P.sardusseemtohavestrongerresemblancetocheloniidsrather thanthatofanypan-trionychid.
As for the putative, now lost, carapace from Sassari, Portis (1901a)onlybrieflydescribedthisspecimen,withoutfiguringit, stating that this specimen was smaller than the holotype and apparentlypertained toa youngerindividual.Fortunately, how-ever,thepartialcarapaceimprintfromthetypelocalitydescribed byComaschiCaria(1959)asreferabletoP.sardusisstilllocatedin thecollectionsofMDLCAundertheaccessionnumber14008. How-ever,thisspecimenalsohassufferedseveredegradationsinceits originaldescription(Fig.4).Althoughitwasinitiallyanalmost com-pleteimprintofacarapace,missingonlyitsupperrightandlower marginsoftheshell,initscurrentstateofpreservation,largeparts ofthecarapaceimprintaremissingandtheedgesofmostcostals havefaded.Weconsiderthisspecimenaswelltobean indeter-minatepan-trionychine,onthebasisoftheabsenceofperipherals andshellscutes.
6.2. TrionychidsfromtheMediterraneanIslands
As evidenced by the fossil record, soft-shelled turtles vari-ouslyoccurredintheMediterraneanislands(GeorgalisandJoyce, 2017).Thiscladehasnoextantrepresentativesinthe Mediter-raneanislands,althoughlivingindividualsofTrionyxtriunguishave beenrepeatedlyreportedfromtheDodecaneseIslandsinGreece,
Fig.4.ThecurrentstateofpreservationofthetrionychidreferredtoProcyclanorbis
sardusbyComaschiCaria(1959)(MDLCA14008).
Étatde conservationactueldu trionychidérapportéà Procyclanorbissarduspar
ComaschiCaria(1959)(MDLCA14008).
specificallyKos,Symi,Leros,KalymnosandRhodes,some kilome-tersawayfromthesouthwesterncoastofAsiaMinor(Corsini-Foka andMasseti,2008).However,thesesightingsoflivingindividualsof T.triunguisshouldbebetterconsideredasrandomcasesofmarine dispersalsacrossnarrowstraightsoftheAegeanSea,asthisspecies hasbeenwelldocumentedtoswim atcertainmarinedistances fromthecoast(Taskavak etal.,1999).Nevertheless,mostfossil findsfromthisregionareratherfragmentary,hinderingtheexact taxonomicaffinitiesoftheMediterraneanIslandspan-trionychids. SuchremainshavebeenfoundintheEoceneoftheBalearicIslands (Mallorca; JiménezFuenteset al.,1990)and Sardinia(Kotsakis, 1985),andtheMioceneofCyprus(Reed,1932;Hadjisterkotisetal., 2000),Crete(Georgalisetal.,2016),Sicily(DeGregorio,1883),and ofcourseSardinia.Amongthese,onlytheMioceneSardinianand Sicilianmaterialhasbeenidentifiedatthespecieslevel,withthe two supposedly endemictaxa Procyclanorbis sardusandTrionyx ragusensisDeGregorio,1883,respectively.Thelatteroccurrence is rather problematic, as theonly known specimenis lostand hasnever beenfigured,and therefore,T. ragusensisshould bet-ter beconsideredanomendubium(Georgalisand Joyce,2017). Curiously,pan-trionychidsaretotallyabsentfromthewell-known MiocenefaunasoftheAegeanIslands(GeorgalisandKear,2013), withtheexceptionofCrete,fromwheretheywereonlyrecently described(Georgalisetal.,2016).Additionally,theirtotalabsence fromCorsicaseemsbizarre,aspan-trionychidsareabundantinthe latePaleogeneandNeogeneofsouthernFranceandnorthwestern and centralwesternItaly(e.g.Portis,1879,1883;Ristori, 1895; Bergounioux,1933).ThecaseofMaltaisintriguing,asfromthat island,Lydekker(1891)establishedTrionyxmelitensis,apurported trionychidthatwaseitherassignedtocyclanorbines(Lapparentde BroinandVanDijk,1999),Trionyxsensulato(Kotsakis,1985),orthe AsianNilssonialineage(Hummel,1929).However,Georgalisand
Joyce(2017)recentlyshowedthattheholotypeandonlyknown specimenofT.melitensispertainstoacheloniid,andmore specifi-callytoTrachyaspisoraTrachyaspis-likegenus,amarineturtlethat ischaracterizedbya distinctivesculpturing pattern.Other pur-portedoccurrencesoffossiltrionychidsfromMalta(Gulia,1843; Cooke,1890)mostprobablypertaintothesame individual,the holotypeof T.melitensis (Zammit-Maempel,1979).Accordingly, theholotypeofP.sardusremainsthemostcompletefossil pan-trionychidspecimenfromtheMediterraneanIslands.
Of course the paleogeographyof the Mediterranean Islands wastotallydifferentduringthePaleogeneandtheNeogene,with certainislandseitherconnected withtheEuropeanand African mainlandoremergingonlymorerecently(EsuandKotsakis,1983; Rögl,1999).Thisfactinevitablyhindersourunderstandingofthe MediterraneanIslandstrionychids,anditcannotbeknownwith certaintyiftheyrepresentindeedinsularformsoraresimply repre-sentativesofcontinentaltaxa.Thescarcenessoffossiltrionychids fromNorthAfrica(Georgalisand Joyce,2017)alsohampersthis situation,althoughfew complete findsclearly denotethe pres-enceoftheTrionyxtriunguislineagealreadyinthatregion(Wood, 1987).InthecaseofSardiniaandProcyclanorbissardus,duringthe Tortonian–MessiniantheSardo-CorsicanMassifandtheTuscany areaformedanarchipelagoofislands,isolatedfromcontinental Europe(Casanovas-Vilaretal.,2011).Thisinsularpaleobioprovince comprisedahighlyuniqueislandvertebratefauna,asitistestified byfossils recoveredfromtheFiumeSantolocality (Portotorres, northwesternSardinia), whichinclude theprimateOreopithecus andseveralpeculiarbovidsand rodents(Abbazziet al.,2008b). Consequently,wesuggestthatP.sarduswasaninsulartaxon. Sub-sequently,theSardinia-CorsicaareawasisolatedfromTuscanyby themid–lateMessinianduetotheopeningoftheTyrrhenianSea (Palombo,2009).
7. Conclusions
The holotype shell of Procyclanorbis sardus Portis, 1901a, is described herein and the taxonomic status of this species is reevaluated.Newanddetailedfiguresoftheholotype(including interpretativedrawings),whichhassufferedseveredamagingsince itsoriginaldescription,areprovided.AffinitiesofP.sarduswith cyclanorbinesarediscardedonthebasisofitsshellanatomy,and theSardiniantaxonclearlybelongs totrionychines.However,a referraltoeitherRafetusorTrionyx,thetrionychinelineagesthatare presentintheMioceneofEurope,iscurrentlynotpossible.Theskull andthehyo-hypoplastronthatwereoriginallyreferredtoP.sardus byPortis(1901a),belonginfacttocheloniidturtles.Procyclanorbis sardusisconsideredtobeanindeterminatepan-trionychineand thenameisconsideredanomendubium.However,evenifitdoes notbeardistinctivediagnosticfeatures,theholotypespecimenof P.sardusrepresentsthebest-preservedtrionychidfossilfromthe MediterraneanIslands.
Despiteaconspicuousfossilrecord(Delfino,2002;Chesietal., 2007;Chesi,2009),theonlyvalidturtlespeciesfromSardiniais thereforeTestudopecoriniiDelfino,2008thatwasdescribedonthe basisofacompleteshellfromtheEarlyPleistoceneoftheD4local faunaofCapoMannu(Abbazzietal.,2008a).Thestatusofnomen dubiumforthetrionychidturtleProcyclanorbissardusPortis,1901a followsthatofPalaeopythonsardusPortis,1901,whoseholotype wasoriginallyreferred toa pythonidsnake(Portis,1901b), but thatisinfactanindeterminateacanthomorphanfish(Delfinoetal., 2014). Similarly, we planned the revision of Tomistoma calari-tanumCapellini,1890,becauseisnotclearifthisspecies,originally described in two papers publishedin thesame year(Capellini, 1890a,b)isvalidornot(seeKotsakisetal.,2004,andPirasetal., 2007,andliteraturetherein)andthetypewasseverelydamaged
duringtheWorldWarII,andthereforefewmorphological char-actersareleftforitsrevision.Therevisionoftypematerialsand theretrievalofnewremainsfromthetypeorneighboring local-ities[notalwayspossiblebutveryuseful;seeZobolietal.(2016) forarecentexampleconcerningaSardinianmonkey]ismandatory toreassessthevalidityoftaxathatwereerectedinthelate nine-teenthorearlytwentiethcenturybyenthusiasticpaleontologists thatknewverywelltheliterature,buthadlittledirectfamiliarity withthemorphologyandvariationofextantandextinctreptiles (Delfinoetal.,2014).
Disclosureofinterest
Theauthorsdeclarethattheyhavenocompetinginterest.
Acknowledgements
WewouldliketothankWalterJoyce(UniversityofFribourg) forusefulcommentsthatimprovedthequalityofthemanuscript. WealsothankoureditorRomainVullo(CNRSGéosciencesRennes) andourreviewersJulienClaude(Institutdessciencesdel’évolution deMontpellier)andAdánPérez-García(GrupodeBiología Evolu-tiva,FacultaddeCiencias,UNED,Madrid)fortheirvaluablehelp duringtheEditorialandReviewingprocess.MDwassupportedby FondidiAteneodell’UniversitàdiTorino(2015–2016),Generalitat deCatalunya(2014SGR416GRCandCERCAProgram)and Span-ishMinisteriodeEconomíayCompetitividad(CGL2016-76431-P). DZacknowledgestheSardiniaRegionalGovernmentforfinancial supportofPhDscholarship(P.O.R.SardegnaF.S.E.Operational Pro-grammeoftheAutonomousRegionofSardinia,EuropeanSocial Fund2007-2013-AxisIVHumanResources,Objectivel.3,Lineof Activity l.3.1.).GLPwassupportedby“ContributoAteneoperla Ricerca”oftheCagliariUniversity.
References
Abbazzi,L.,Carboni,S.,Delfino,M.,Gallai,G.,Lecca,L.,Rook,L.,2008a.Fossil
ver-tebrates(MammaliaandReptilia)fromCapoMannu(LatePliocene,Western Sardinia,Italy)withdescriptionofanewTestudo(CheloniiTestudinidae)species. RivistaItalianadiPaleontologiaeStratigrafia114,119–132.
Abbazzi,A.,Delfino,M.,Gallai,G.,Trebini,L.,Rook,L.,2008b.Newdataonthe
ver-tebrateassemblageofFiumeSanto(North-WestSardiniaItaly)andoverview ontheLateMioceneTusco-Sardinianpalaeobioprovince.Palaeontology52, 251–425.
Barca,S.,Melis,E.,Annino,E.,Cincotti,F.,Ulzega,A.,Orrù,P.,Pintus,C.,2005.
NoteIllustrativedellaCartaGeologicad’Italiaallascala1:50.000.In:Foglio557 “Cagliari”.ServizioGeologicod’Italia,RegioneAutonomadellaSardegna,S.EL.CA, Firenze.
Bergounioux,F.M.,1933.Monographiepaléontologiquedelafaunedevertébrés
dessablesdeMontpellier2.ChéloniensfossilesdessablesPliocènesastiensde Montpellier.TravauxduLaboratoiredeGéologiedelaFacultédessciencesde Lyon23,1–32.
Bergounioux,F.M.,1935.Contributionàl’étudepaléontologiquedeschéloniens:
chéloniensfossilesdubassind’Aquitaine.MémoiresdelaSociétéGéologiquede France25,1–215.
Bergounioux,F.M.,1954.LeschéloniensfossilesdesterrainstertiairesdelaVénétie.
MemoriedegliIstitutidigeologiaemineralogiadell’UniversitàdiPadova18, 1–115.
Broin,F.de,1977.Contributionàl’étudedesChéloniens:Chélonienscontinentaux
duCrétacéetduTertiairedeFrance.MémoiresduMuséumNationald’Histoire Naturelle,SérieC,38.,pp.1–366.
Capellini,G.,1890a.Sulcoccodrillianogarialoide(Tomistomacalaritanus)scoperto
nellacollinadiCagliarinel1868.RendicontidellaRealeAccademiadeiLincei, serie46,149–151.
Capellini,G.,1890b.Sulcoccodrillianogarialoide(Tomistomacalaritanus)scoperto
nellacollinadiCagliarinelMDCCCLXVIII.AttidellaRealeAccademiaLincei, MemoriedellaClassediScienzeFisiche.MatematicheeNaturali,serie46, 507–533.
Casanovas-Vilar,I.,Dam,J.A.,van,Trebini,L.,Rook,L.,2011.Therodentsfromthe
LateMioceneOreopithecus-bearingsiteofFiumeSanto(Sardinia,Italy).Geobios 44,173–187.
Cherchi, A.,1974.AppuntibiostratigraficisulMiocene dellaSardegna(Italia),
ActesVCongrèsduNéogeneMéditerranéeLyon1971.MémoireduBureaude RecherchesGéologiquesetMinières78,pp.433–445.
Chesi,F.,(PhDdissertation)2009.Ilregistrofossileitalianodeicheloni.Università
diFirenze,Florence(Unpublished).
Chesi,F.,Delfino,M.,Abbazzi,L.,Carboni,S.,Lecca,L.,Rook,L.,2007.Newfossil
ver-tebrateremainsfromSanGiovannidiSinis(LatePleistocene,Sardinia):thelast Mauremys(ReptiliaTestudines)intheCentralMediterranean.RivistaItalianadi PaleontologiaeStratigrafia113,287–297.
ComaschiCaria,J.,1959.NuovirestidicheloninelMiocenedellaSardegna.Bollettino
dellaSocietàGeologicaItaliana78,37–44.
ComaschiCaria,I.,1986.AnimaliepiantefossilidellaSardegna.(Terzaedizione
ampliataeconunguidabreveallostudiodellaPaleontologia).Edizionidella Torre,Cagliari.
Cooke,J.H.,1890.NotesonthefossilreptilesofMalta.IlNaturalistaMaltese1,
18–20.
Corsini-Foka,M.,Masseti,M.,2008.OntheoldestrecordoftheNilesoft-shelled
turtle,Trionyxtriunguis(Forskal,1775),intheEasternAegeanislands(Greece). ZoologyintheMiddleEast43,108–110.
DeGregorio,A.,1883.Nuovifossiliterziari(vertebratieinvertebrati).Palermo.
Delfino,M.,(PhDdissertation)2002.ErpetofauneitalianedelNeogeneedel
Qua-ternario.UniversityofModenaandReggioEmilia,ModenaandReggioEmilia (Unpublished).
Delfino,M.,Zoboli,D.,Carnevale,G.,Pillola,G.L.,2014.Therediscovered
holo-typeofPalaeopythonsardusPortis,1901fromtheMioceneofSardiniabelongs to a fish, not asnake. Bollettinodella Società PaleontologicaItaliana 53, 89–92.
Esu,D.,Kotsakis,T.,1983.LesvertébrésetlesmollusquescontinentauxduTertiaire
delaSardaigne:paléobiogéographieetbiostratigraphie.GeologiaRomana22, 177–206.
Fucini,A.,1912.TrionyxpliocenicusLaw.PaleontographiaItalica18,1–28.
Gandolfi, R., Porcu, A., 1967. Contributo alla conoscenza delle microfacies
miocenichedellecollinediCagliari(Sardegna).RivistaItalianadiPaleontologia eStratigrafia73,313–348.
Georgalis,G.L.,Joyce,W.G.,2017.AreviewofthefossilrecordofOldWorldturtlesof
thecladePan-Trionychidae.BulletinofthePeabodyMuseumofNaturalHistory 58,115–208.
Georgalis,G.L.,Kear,B.P.,2013.ThefossilturtlesofGreece:anoverviewoftaxonomy
anddistribution.Geobios46,299–311.
Georgalis,G.L.,Villa,A.,Vlachos,E.,Delfino,M.,2016.Fossilamphibiansandreptiles
fromPlakias,Crete:aglimpseintotheearliestlateMioceneherpetofaunasof southeasternEurope.Geobios49,433–444.
Gray,J.E.,1825.Asynopsisofthegeneraofreptilesandamphibia,withadescription
ofsomenewspecies.AnnalsofPhilosophy10,193–217.
Gulia,G.,1843.GuscioditestugginefossiletrovatoaMalta.RepertoriodiConoscenze
UtileZoologiaFossile27,77–78.
Hadjisterkotis,E.,Kassapis,I.,Klerides,N.,2000.Paleontologicalevidenceforthe
presenceofTrionyxsp.onCyprus,withnewinformationonlivingandalien speciesofterrapins.Newsletter,HellenicZoologicalSociety33,8–10.
Hummel, K., 1929. Die fossilen Weichschildkröten (Trionychia): eine
morphologisch-systematischeundstammesgeschichtlicheStudie.Geologische undPaläontologischeAbhandlungen16,359–487.
Hummel,K.,1932.Pars52:Trionychiafossilia.In:Quenstedt,W.(Ed.),Fossilium
Catalogus.I:Animalia.W.Junk,Berlin.
JiménezFuentes,E.,RamosGuerrero,E.,MartínDeJesús,S.,PérezRamos,E.,Mulas
Alonso,E.,1990.QueloniosdelEocenomediodeMallorca.Paleontologiai
Evolu-ció23,153–156.
Karl,H.V.,1999.DieZoogeographiederkänozoischenWeichschildkröteTrionyx
triunguisForskål,1775(Testudines:Trionychidae).Joannea–Geologieund Paläontologie1,27–60.
Kotsakis,T.,1985.LesTrionychidae(TestudinataReptilia)fossilesdel’Italie.
Bollet-tinodellaSocietàPaleontologicaItaliana24,161–168.
Kotsakis,T.,1989.LateTurolianamphibiansandreptilesfromBrisighella(northern
Italy):preliminaryreport.BollettinodellaSocietaPalaeontologicaItaliana28, 277–280.
Kotsakis,A.,Delfino,M.,Piras,P.,2004.ItalianCenozoiccrocodilians:taxa,
tim-ingandpalaeobiogeographicimplications.Palaeogeography,Palaeoclimatology, Palaeoecology210,67–87.
Kotsakis,T.,Palombo,M.R.,1979.Vertebraticontinentaliepaleogeografiadella
SardegnaduranteilNeogene.AnnalesGéologiquesdesPaysHelléniques.Tome HorsSérie2,621–630.
Kuhn,O.,1964.Part107,Testudines.In:Kuhn,O.(Ed.),FossiliumCatalogus,Volume
I:Animalia.W.Junk,TheHague.
LapparentdeBroin,F.,VanDijk,P.P.,1999.CheloniafromtheLateMioceneBaynunah
Formation,EmirateofAbuDhabi,UnitedArabEmirates:palaeogeographical implications.In:Whybrow,P.J.,Hill,A.(Eds.),FossilVertebratesofArabia.Yale UniversityPressYale,pp.136–162.
Laube,G.C.,1895.VorläufigerBerichtüberSchildkrötenresteausderböhmischen
Braunkohlenformation.VerhandlungenderGeologischenBundesanstalt1895, 391–392.
Laube,G.C.,1900.NeueSchildkrötenundFischeausderböhmischen
Braunkohlen-formation.Lotos2,37–56.
Liebus,A.,1930.NeueSchildkrötenausdenTertiärenSüsswassertonenvonPreschen
beiBilininBöhmen.RozpravyStátníhogeologickéhoústavu ˇCeskoslovenské Republiky4,1–57.
Lydekker,R.,1885.IndianTertiaryandpost-TertiaryVertebrata.Siwalikand
Ner-badaChelonia.MemoirsoftheGeologicalSurveyofIndia,21.,pp.155–208.
Lydekker,R.,1889.NotesonSiwalikandNarbadaChelonia.RecordsoftheGeological
SurveyofIndia22,56–58.
Lydekker,R.,1891.OnanewspeciesofTrionyxfromtheMioceneofMaltaanda chelonianscapulafromtheLondonClay.TheQuarterlyJournaloftheGeological SocietyofLondon47,37–40.
Meylan,P.A.,1987.Thephylogeneticrelationshipsofsoft-shelledturtles(Family
Trionychidae).BulletinoftheAmericanMuseumofNaturalHistory186,1–101.
Palombo,M.R.,2009.Biochronology,paleobiogeographyandfaunalturnoverin
westernMediterraneanCenozoicmammals.IntegrativeZoology4,367–386.
Peters,K.F.,1855.SchildkrötenresteausdenösterreichischenTertiärablagerungen.
DenkschriftenderKaiserlichenAkademiederWissenschaften, mathematisch-naturwissenschaftlicheClasse9,1–22.
Piras,P.,Delfino,M.,DelFavero,L.,Kotsakis,T.,2007.Phylogeneticpositionofthe
crocodylianMegadontosuchusarduini(deZigno,1880)andtomistomine palaeo-biogeography.ActaPalaeontologicaPolonica52,315–328.
Portis,A.,1879.DialcunifossiliterziariidelPiemonteedellaLiguriaappartenenti
all’ordinedeichelonii.MemoriedellaRealeAccademiadelleScienzediTorino, serie2,2.,pp.113–134.
Portis,A.,1883.NuovicheloniifossilidelPiemonte.MemoriedellaRealeAccademia
delleScienzediTorino,serie2,35.,pp.369–378.
Portis,A., 1901a.Il ProcyclanorbissardusPort.,nuovotrionychidefossile della
Sardegna.BollettinodellaSocietàGeologicaItaliana20,51–79.
Portis,A.,1901b.IlPalaeopythonsardusPort.NuovopitonidedelMiocenemedio
dellaSardegna.BollettinodellaSocietàGeologicaItaliana20,247–253.
Reed,F.R.C.,1932. NewMiocenefaunasfromCyprus.GeologicalMagazine69,
511–517.
Reinach von, A., 1900. Schildkrötenreste im Mainzer Tertiärbecken und in
benachbartenungefährgleichaltrigen.AbhandlungenderSenckenbergischen NaturforschendenGesellschaft28,1–135.
Ristori,G.,1895.ChelonianifossilidiMontebambolieCasteani.Memoria
paleonto-logicadelprof.GiuseppeRistori.ConappendicesuiChelonianifossilidelCasino (Siena),21.PubblicazionidelRealeIstitutodiStudisuperioriinFirenze,lezione discienzefisicheenaturale,pp.1–104.
Rögl, F., 1999. Mediterranean and Paratethys. Facts and hypotheses of an
OligocenetoMiocenepaleogeography(shortoverview).GeologicaCarpathica 50,330–349.
Taskavak,E.,Reimann,M.J.,Polder,W.N.,1999.FirstrecordoftheNilesoft-shelled
Turtle,Trionyxtriunguis,fromKosIsland,Greece,withcommentsonits occur-renceintheeasternMediterranean.ChelonianConservationandBiology3, 510–512.
Vitek,N.S.,Joyce,W.G.,2015.AreviewofthefossilrecordofNewWorldturtlesof
thecladePan-Trionychidae.BulletinofthePeabodyMuseumofNaturalHistory 56,185–244.
Wood,R.C.,1987.FossilturtlesfromtheSahabiFormation.In:Boaz,T.,El-Amauti,
A.,Gaziry,A.W.,deHeinzelin,J.,DechantBoaz,D.(Eds.),NeogenePalaeontology andGeologyofSahabi.AlanR.Liss,NewYork,pp.107–112.
Zammit-Maempel,G.,1979.TheIndo-PacificaffinityofsomeMalteseTertiaryfossils.
TheCentralMediterraneanNaturalist1,1–12.
Zoboli,D.,Pillola,G.L.,2016.IrettilimioceniciconservatinelMuseoSardodi
Geolo-giaePaleontologiaDomenicoLovisato(Cagliari,Italia).MuseologiaScientifica nuovaserie10,81–87.
Zoboli,D.,Pillola,G.L.,Rook,L.,2016.NewremainsofMacacamajoriAzzaroli,1946
(Primates,Cercopithecidae)fromIsOreris(Fluminimaggiore,southwestern Sar-dinia).BollettinodellaSocietàPaleontologicaItaliana55,227–230.
