Anatomy of Juravenator starki (Theropoda: Coelurosauria) from the ...

1 Introduction

Small non-avian theropods of Late Jurassic age arerare worldwide (WEISHAMPEL et al 2004) Only ahandful of them exist in Europe and until recentlythese dinosaurs were best represented by two incom-plete skeletons of Compsognathus longipes (WAGNER

1861) Despite the historical significance of Comp -sognathus the discovery of which played a key role in

the evolutionary debate of the 19th century (DESMOND

1982 OSTROM 1978 CHIAPPE 2007) the rather un -specialized morphology of ldquocompsognathidsrdquo (Comp-sognathus and a series of alleged Late Jurassic-EarlyCretaceous relatives ndash CURRIE amp CHEN 2001 HOLTZ

et al 2004 HWANG et al 2004 NAISH et al 2004 JI

et al 2007b GISHLICK amp GAUTHIER 2007) has com-plicated interpretations of this grouprsquos monophyly Inthis paper ldquocompsognathidsrdquo are used without any

Anatomy of Juravenator starki (Theropoda Coelurosauria)from the Late Jurassic of Germany

Luis M Chiappe Los Angeles and Ursula B Goumlhlich Vienna

With 26 figures

CHIAPPE L M amp GOumlHLICH U B (2011) Anatomy of Juravenator starki (Theropoda Coelurosauria)from the Late Jurassic of Germany ndash N Jb Geol Palaumlont Abh 258 257ndash296 Stuttgart

Abstract We provide a detailed study of the morphology of the holotype of Juravenator starki fromthe Late Jurassic of the Solnhofen area of southern Germany The incompletely ossified surface ofmultiple bones and lack of several skeletal fusions indicate that Juravenator starki is based on animmature specimen Nonetheless numerous unique morphologies and bone proportions distinguishthis taxon from Compsognathus longipes the only previously named non-avian theropod dinosaurfrom the Late Jurassic of the Solnhofen Archipelago Yet its skeletal anatomy is most similar to thatof Compsognathus and other theropods that have often been regarded as closely related to the latter ndashsometimes within a monophyletic Compsognathidae Juravenator is characterized by having a smallsize (~ 075-meter-long in the holotype) with few maxillary teeth lack of a premaxillary-maxillarydiastema an antorbital fenestra subequal in length to orbit an elongate scapula that is narrowest at itsneck a proportionally short humerus and high and abruptly tapered manual claws and bow-like zygapophysial articulations in the mid-caudal vertebrae Portions of the epidermis preservedmainly along the tail provide the only glimpse of the morphology of the skin of basal coelurosaursand structures newly revealed under UV light hint at the possibility of filamentous integumentarystructures ndash akin to those interpreted as proto-feathers in other basal coelurosaurs ndash also covering thebody of this dinosaur The discovery of Juravenator has provided evidence of morphologies ndash fromdetails of the skull to the epidermis ndash that are poorly known in other theropods interpreted as ator near the base of Coelurosauria and thus contributes significantly to our understanding of theevolutionary history of this clade The exquisitely preserved holotipic skeleton adds significantly tothe meager record of small-bodied Late Jurassic theropods

Key words Juravenator Theropoda anatomy taphonomy preservation Jurassic Solnhofen Lime-stones

copy 2010 E Schweizerbartrsquosche Verlagsbuchhandlung Stuttgart Germany wwwschweizerbartde

DOI 1011270077-774920100125 0077-774920100125 $ 1000

N Jb Geol Palaumlont Abh 2583 257ndash296 ArticleStuttgart December 2010

dm
Texte surligneacute
dm
Texte surligneacute
dm
Texte surligneacute
dm
Texte surligneacute
dm
Texte surligneacute

258 L M Chiappe and U B Goumlhlich

implication of the putative monophyletic nature ofa group that includes Compsognathus longipes(WAGNER 1861 OSTROM 1978 PEYER 2006) Huaxia-gnathus orientalis (HWANG et al 2004) Sinosauro -pteryx prima (JI amp JI 1996 CHEN et al 1998 CURRIE

amp CHEN 2001) Scipionyx samniticus (DAL SASSO ampSIGNORE 1998) and Juravenator starki (GOumlHLICH amp

CHIAPPE 2006) among other non-avian theropodsusually regarded at or near the base of the coeluro -saurian clade The morphological diversity of thesetheropods has remained as a poorly understood chap-ter in the evolutionary history of Late Jurassic-EarlyCretaceous theropod dinosaurs (CURRIE amp CHEN

2001 HOLTZ et al 2004 PEYER 2006) However

Fig 1 Geographic location of Juravenator starki (JME Sch 200) Compsognathus longipes and selected specimensof Archaeopteryx lithographica (reconstructions not to scale) Barred and mottled patterns indicate the approximatedepositional areas during the Late Jurassic

these generally unspecialized theropods provide keyevidence for understanding the morphological trans-formations that occurred near the base of Coeluro -sauria an extremely diverse clade of dinosaurs thatincludes all living birds

In 1998 ending a decade-long collecting program(1989-1998) by the Jura-Museum Eichstaumltt underthe direction of GUumlNTER VIOHL volunteers KLAUS-DIETER and HANS-JOACHIM WEISS (Kelkheim-Fisch-bach) unearthed a small Late Jurassic theropod fromthe FRANZ STARK Quarry of Schamhaupten (districtof Eichstaumltt Southern Franconian Alb Bavaria Ger-many) (Fig 1) The WEISS brothers first split a slabcontaining portions of the skull and cervicals whichled them to quarry the adjacent slabs The specimen(JME Sch 200) was brought to the Jura-Museum Eich-staumltt for a multi-year preparation by PINO VOumlLKL who

revealed a nearly complete skeleton with traces of softtissues (for more details on the history of discoveryand preparation see TISCHLINGER et al [2006] andGOumlHLICH et al [2006]) The specimen was prelimi -narily described and named Juravenator starki byGOumlHLICH amp CHIAPPE (2006) Based on several synapo -morphies including the presence of a round orbitsome maxillary and dentary teeth without anteriorserrations and axial epipophyses not extendingbeyond the posterior rim of the postzygapophysesGOumlHLICH amp CHIAPPE (2006) identified Juravenator asa coelurosaurian and clustered this taxon with otherbasal coelurosaurians within Compsognathidae Asmentioned above however the monophyly of ldquocomp-sognathidsrdquo and their relationships to other coeluro-saurians are far from settledWhile a number of phylo -genetic studies (eg HWANG et al 2004 SENTER 2007

Anatomy of Juravenator starki from the Late Jurassic of Germany 259

Fig 2 Stratigraphic position of STARK Quarry (Schamhaupten Member) and Juravenator starki (JME Sch 200)Stratigraphic section and table modified after VIOHL amp ZAPP (2006 2007)


XU et al 2009 CHOINIERE et al 2010) have recoveredthem as a monophyletic clade taxa usually ascribedas ldquocompsognathidsrdquo are distributed across the basalportion of the coelurosaurian tree in other studies(eg BUTLER amp UPCHURCH 2007) Furthermorecladistic analyses in which compsognathids are re -covered as monophyletic usually limit the taxonomicsampling of these theropods to better known taxasuch as CompsognathusSinosauropteryx and Huaxia -gnathus but fail to include a variety of other taxa thathave also been described as members of the samegroup (eg Mirischia asymmetrica Sinocalliopteryxgigas) (MARTILL et al 2000 NAISH et al 2004 JI et al2007a)

The phylogenetic relationships of Juravenator to -gether with that of many other basal coelurosaurians(and the monophyly of ldquocompsognathidsrdquo) needs tobe studied in light of more comprehensive cladisticanalyses This notewithstanding the discovery ofJuravenator resulted in an exceptionally well-pre -served specimen ndash possibly the most complete non-avian theropod from Europe ndash and a startling newexample of a small-bodied Late Jurassic carnivorousdinosaur Juravenator thus provides important infor-mation for better understanding the role played bybasal coelurosaurs in the evolutionary history ofthero pods The present paper provides a detaileddescription of both the skeletal morphology and in -tegumentary anatomy of this dinosaur

Inst i tut ional abbreviat ions ndash AMNH AmericanMuseum of Natural History (New York) BSPG BayerischeStaatssammlung fuumlr Palaumlontologie und Geologie (Munich)CAGS Chinese Academy of Geological Sciences (Beijing)GMV Geological Museum of China (Beijing) JME Jura-Museum Eichstaumltt (Eichstaumltt) SMNK Staatliches Museumfuumlr Naturkunde Karlsruhe (Karlsruhe)

2 Geological setting

During the Late Jurassic southern Germany waslargely submerged by a shelf sea wedged between twolarge islands to the north and the deeper Tethys Oceanto the south (BARTHEL et al 1990) A series of lime-stone-filled basins developed within this shallow seaone of them was the small (~ 35 km2) SchamhauptenBasin (Southern Franconian Alb) in which limestoneswere deposited amid a large complex of dolomiticreefs (RENESTO amp VIOHL 1997 ZEISS 2001 VIOHL ampZAPP 2006 2007) (Figs 1-2)

The skeleton of Juravenator (JME Sch 200) is con-tained in the strongly silicified laminated limestone

that outcrops at the FRANZ STARK Quarry of Scham-haupten VIOHL amp ZAPP (2006 2007) identified thissilicified plattenkalk as belonging to the Late Kim -meridgian Schamhaupten Subformation which is con -sidered to be part of the Painten Formation Based onrecent biostratigraphic studies (ZEISS 2001 VIOHL ampZAPP 2006 2007 SCHWEIGERT 2007) the top lime -stones at Stark Quarry are dated as uppermost Kim-meridgian (Beckeri Zone Ulmense Subzone) (Fig 2)thus slightly older than the neighboring SolnhofenLithographic Limestones (Solnhofener Plattenkalk) ndashthe celebrated deposits that have yielded both Comp-sognathus and Archaeopteryx (WELLNHOFER 2008)

The Schamhaupten plattenkalk (which is differentthan the Solnhofen plattenkalk) comprises two maintypes of beds bindstones and detrital carbonatesMost articulated fossils are contained within bind -stones while detrital carbonates most typically yieldshells and isolated skeletal elements (VIOHL amp ZAPP

2006) ndash Juravenator is embedded in a 35-cm-thickbindstone belonging to section layer E3 (see VIOHL ampZAPP 2006 fig 5 2007 fig 4) Containing more than200 taxa the Schamhaupten plattenkalk has preserveda great variety of fossil plants nano- and microfossilsinvertebrates and vertebrates (VIOHL 1999 VIOHL ampZAPP 2006) Fish are the most abundant vertebratesbut a variety of reptiles ndash among these the turtlesSolnhofia and Eurysternum the sphenodontid Lepto-saurus and Juravenator (RENESTO amp VIOHL 1997VIOHL 1999 2006 2007) ndash are also well-representedThe fact that most Schamhaupten fossils are of marineorigin suggests good connections to the Tethys Oceanbut the remains of terrestrial taxa also indicate theexistence of nearby islands

The exceptional preservation illustrated by therichness of articulated fossils ndash fish reptiles crusta -ceans sea urchins with in situ spines and others ndashdesignates the Schamhaupten plattenkalk as a Kon -servat Lagerstaumltte although one in which organisms indifferent stages of decay are also preserved Thisextraordinary preservation has been recently ex -plained by the taphonomic model of VIOHL amp ZAPP

(2006 2007) These authors suggest that a stratifiedsalinity developed within the Schamhaupten Basinits bottom zone was hypersaline anaerobic or dys -aerobic and thus lethal while its upper water columnwas suitable for marine life The hostile nature of thebasinrsquos bottom suggests the absence of benthonicscavengers althoughVIOHL ampZAPP (2006 2007) inter -preted the proliferation of microbial mats as respon -sible for the generation of bindstones by trapping sedi-


Fig 3 Photograph under normal light of the skeleton of Juravenator starki (JME Sch 200) Dashed line on inset highlightsthe boundaries of the two blocks in which the specimen was collected


ment and for the extraordinary preservation of Scham-haupten by rapidly sealing the sunken corp ses

3 Taphonomy of Juravenator

Instead of having the typical opisthotonic posture ofmany fossil amniotes (MARSHALL FAUX amp PADIAN

Fig 4 Photograph under ultraviolet light of the skeleton of Juravenator starki (JME Sch 200)

2007) the skeleton of Juravenator (Figs 3-5) exhibitstwo distinct torsions at the level of the neck and mostunusually in front of the pelvis While the skull andtail are primarily exposed in right lateral view thetorso shows its left lateral side The pelvic bones aresomewhat disarticulated ndash the ilia are displaced andexhibit their right sides ndash and the hindlimbs aresplayed ventrally each showing its laterocaudal sur -face

The completeness and extensive articulation ofthe skeleton of Juravenator strongly suggests that thecarcass floated for a very short period of time before itsunk into the hostile bottom of the basin soon after theanimalrsquos death Explaining the unusual death postureof the skeleton is however more problematic Now -here in the skeleton is there apparent evidence of pre-dation that can explain the observed rotation of thepelvis with respect to the torso The inferred bottomconditions point at the absence of benthonic sca -vengers that could have disrupted the skeleton andeven the presence of very small marine woodlouses(Isopoda) (Fig 6) ndash that probably scavenged thecorpse before it settled on the basinrsquos bottom ndash canhardly explain the aforementioned rotation All thissuggests that the peculiar torsion of the skeleton ofJuravenator is most likely the result of taphonomicfactors A possible scenario might be that the carcasssettled into the basinrsquos floor with the head lying on itsleft side the torso leaning towards its right side withthe right forelimb tucked under the body the pelvis ina more upright position and the legs splayed If so thecomplete rotation between torso and pelvis may beexplained by the strong compaction of the bindstonesduring diagenesis (VIOHL amp ZAPP 2006)

4 Systematic paleontology

Dinosauria OWEN 1842Theropoda MARSH 1881

Tetanurae GAUTHIER 1986Coelurosauria V HUENE 1914

Juravenator starki GOumlHLICH amp CHIAPPE 2006

Holotype JME Sch 200 a nearly complete and articu -lated skeleton ndash preserving portions of soft tissue ndash missingonly the distal third of its tail (Figs 3-5) The specimen wascollected in two main blocks that were glued together duringpreparation (Fig 3) Some parts of the skull and neck werecollected as slab and counterslab and subsequently gluedtogether and prepared from one side

Horizon and local i ty Silicified laminated limestoneLate Jurassic (Upper Malm Late Kimmeridgian Beckeri

Zone Ulmense Subzone) in the local stratigraphy (ZEISS

2001 SCHWEIGERT 2007) (Fig 2) JME Sch 200 was collec-ted from the Stark Quarry a quarry owned by the Starkfamily and situated a few hundred meters west of the villageof Schamhaupten (district of Eichstaumltt Southern FranconianAlb) Bavaria Germany (Fig 1)

Diagnosis Small basal coelurosaur (a clade encom -passing Passer domesticus and all taxa sharing a morerecent common ancestor with it than with Allosaurus fragi-lis) with a large skull proportionally longer than in Compso-gnathus longipes (skull femur and skull presacral verte-brae ratios are 15 and 047 and 11 and 025 in Juravenatorand C longipes respectively) low number of maxillaryteeth (less than 10 in Juravenator 10 in Ornitholesteshermanni 12 in Sinosauropteryx prima 15 in C longipes)absence of a premaxillary-maxillary (diastema present inScipionyx samniticus) distinct indentation on the denti -gerous margin of the maxilla (between second and thirdteeth) an antorbital fenestra subequal in length to orbit(antorbital fenestra is nearly half the orbit in S primaOrnitholestes hermanni and S samniticus and subequalto the orbit in C longipes and basal tyrannosauroids [XU etal 2004 2006 SERENO et al 2009]) long scapula (hume-russcapula and scapulafemur ratios are 063 and 081 inJuravenator 086 and 060 in S prima and 1 and 054 inC longipes respectively) of which the narrowest portion isat the neck as opposed to near the mid-shaft and short feet(metatarsal III femur is 059 in Juravenator and approxi -mately 075 in S prima and C longipes [CURRIE amp CHEN

2001]) Juravenator is also unique among other basal coe-lurosaurs in having longer and more slender teeth a concave rostral margin of the jugal process of the postorbitalhumeri with very short and triangular-shaped deltopectoralcrests manual claws that are very high proximally and thattaper abruptly around their midpoints and arched bow-likezygapophysial articulations in the mid-caudal vertebrae Weregard these unique features as autapomorphies of Jura -venator starki

5 Anatomical description

The specimen shows features indicating an early onto-genetic age ndash the surface of its bones is intensivelyscarred by small pits and grooves thus revealing apattern of incomplete periosteal formation (egHORNER 1997 SANZ et al 1997 CHIAPPE et al 1998CODORNIUacute amp CHIAPPE 2004) (Fig 7) Evidence ofimmaturity is also present in the lack of fusion be -tween sacral vertebrae and presence of open neuro-central sutures visible on many caudal vertebrae(BROCHU 1996) Based on the preserved length of theaxial skeleton and the estimated length of the missingdistal third of the tail GOumlHLICH amp CHIAPPE (2006)projected the length of the holotype of Juravenatorstarki to 075-80 m This value was considered to bea slight underestimation by THERRIEN amp HENDERSON

(2007) who based on a least-square regression of



Fig 5 Interpretive drawing of the skeleton of Juravenator starki (JME Sch 200) Abbreviations as astragalus cev cervicalvertebrae co coracoid cv 44 caudal vertebra 44 dv dorsal vertebrae fe femur fi fibula ha haemal arches hu humerusil ilium mcI-II metacarpals I-II mt II-V metatarsals II-V ul ulna ra radius sc scapula st soft-tissue ti tibia I-IV pedaldigits I-IV

theropod skulls calculated the length of this specimenas slightly over 1 m The latter may well be accountedfor by the relatively large skull of the holotypebut regardless of the difference between calculationsbased on actual skeletal length and those based onleast-square regressions the juvenile nature of theholotype indicates that the adults of this species musthave substantially exceeded both projected lengthsDetailed osteological measurements are given in theAppendix

51 Skull

Cranial bones ndash The skull (Fig 8) is large in pro-portion to the skeleton (more than 15 times the lengthof the femur) (Figs 3-5) The rostrum is moderatelylong ndash the distance between the lacrimal and therostrumrsquos tip is 53 of the skull length ndash and ex -cavated by a large antorbital fossa

The right premaxilla is preserved in lateral viewand the nasal process of its counterpart is exposeddorsal to the rostral half of the former The body of thepremaxilla is rectangular longer than it is high andbears short maxillary (subnarial) and nasal (inter -narial) processes which have relative lengths morelike those of Compsognathus (OSTROM 1978 PEYER

2006) Scipionyx (DAL SASSO amp SIGNORE 1998) Sino-sauropteryx (CHEN et al 1998 CURRIE amp CHEN

2001) and the basal tyrannosauroid Dilong paradoxus(XU et al 2004) than the long processes of Huaxia-gnathus (HWANG et al 2004) (Figs 8-9) The nasalprocess of the premaxilla extends along the rostro -dorsal margin of the snout bordering dorsally theelliptical external nares for about half the length ofthe nares (Fig 7) The maxillary process extends cau-dally beyond the caudal end of the nasal process andit appears to be thicker than the latter a conditionsimilar to that in Sinocalliopteryx (JI et al 2007a) andopposite to that in Sinosauropteryx (CURRIE amp CHEN

2001) These two processes are also less caudallydivergent than in Sinosauropteryx more similar toSinocalliopteryx (JI et al 2007a) Huaxiagnathus(HWANG et al 2004) Scipionyx (DAL SASSO ampSIGNORE 1998) and Compsognathus (MNHM CNJ79 MICHARD 1991) (Fig 8) There are at least threepremaxillary teeth ndash the last one broken at its base ndashbut we cannot rule out the presence of a fourth andmost anterior tooth (Figs 7 10) Similar uncertaintieshave been expressed for other ldquocompsognathidsrdquoOSTROM (1978) reported a minimum of three pre -maxillary teeth in Compsognathus longipes (BSPGAS I 563) but PEYER (2006) highlighted the presenceof four alveoli in the French Compsognathus (MNHMCNJ 79) CURRIE amp CHEN (2001) were not definitivewhether four or five premaxillary teeth are present inSinosauropteryx (although they believed four to be the


Fig 6 Close-up of isopods preserved as impressions (A) or three-dimensionally (B) within the skeleton of Juravenatorstarki (JME Sch 200)


number) and HWANG et al (2004) were indeterminatebetween three and four In Scipionyx ndash presumablyalso closely related to Compsognathus (DAL SASSOpers comm) ndash there are five premaxillary teeth butonly four in the large compsognathid Sinocalliopteryx(JI et al 2007a)

The right maxilla is exposed in lateral view whatappears to be a displaced small portion of the dorsalramus of the left maxilla is attached to the side ofthe left nasal The maxilla is largely excavated by an

antorbital fossa that contains a large antorbitalfenestra and a small maxillary fenestra (Fig 7) Themaxillary region anterior to the antorbital fossa is veryshort ndash shorter than the length of the body of the pre-maxilla ndash a condition shared with Scipionyx (DAL

SASSO amp SIGNORE 1998) The extent to which thisfeature is related to the early ontogenetic age of theholotypes of Juravenator and Scipionyx is uncertainHowever the condition of these taxa contrasts with theextension of the maxilla in front of the antorbital fossaof Huaxiagnathus (more than three times the length ofthe premaxillary body HWANG et al 2004 fig 2)Dilong and Sinocalliopteryx (more than twice thelength of the premaxillary body XU et al 2004 JI

et al 2007a) Sinosauropteryx (more than 15 timesthe length of the premaxillary body CURRIE amp CHEN

2001 fig 3A) and apparently Compsognathus inwhich the region anterior to the antorbital fossa isslightly longer than the length of the premaxillarybody (Fig 8) The anterior suture of the maxilla ndash itscontact with the premaxilla and nasal ndash is sinusoid Asin Huaxiagnathus and Sinosauropteryx this margin isexcluded from the external nares by the contact ofthe premaxilla and nasal (Fig 8) OSTROM (1978) re -ported that the maxilla of Compsognathus formed thecaudal margin of the external nares however this can-not be verified in either of the two specimens (BSPGAS I 563 MNHM CNJ 79) Three maxillary teeth fitthe dentigerous margin anterior to the antorbital fossandash comparisons with other ldquocompsognathidsrdquo are ham-pered by poor preservation however two teeth appearto be present in Scipionyx and at least three teeth fitthis portion of the maxilla of Sinocalliopteryx Themaxillary of Juravenator preserves an additional fiveteeth along the dentigerous margin beneath the thinventral border of the antorbital fossa ndash the last tooth iscentered along this margin (Figs 7 10) Thus themaxilla of Juravenator bears at least and possibly notmore than eight teeth The number of maxillary teethvaries greatly among other ldquocompsognathidsrdquo andother basal coelurosaurs at least six are present inSinocalliopteryx (JI et al 2007a) seven seem to existin Scipionyx (DAL SASSO pers comm) a minimumof eight are present in both Sinosauropteryx (GMV2124 pers obs) and Huaxiagnathus (HWANG et al2004) 10 in Ornitholestes (OSBORN 1916) 13 in somebasal tyrannosauroids (XU et al 2006 SERENO et al2009) and 14 were reported for Compsognathus(PEYER 2006) (Fig 11) In Juravenator the maxillaryteeth are not separated from those in the premaxilla bya diastema OSTROM (1978) reported the presence of a

Fig 7 Detail of the surface of the left ulna-radius of Jurave-nator starki (JME Sch 200) (A B) as compared to that of aneonate modern bird (C from SANZ et al 2001) Note theintensively scarred surface of the periosteum


Fig 8 Photograph of the skull of Juravenator starki (JME Sch 200) under normal (A) and ultraviolet (B) light andinterpretive drawing (C) Abbreviations an angular AF antorbital fenestra d dentary EN external naris f frontalITF infratemporal fenestra j jugal l lacrimal m maxilla MF maxillary fenestra n nasal OR orbit p parietal plpalatine pm premaxilla po postorbital pr prefrontal pt pterygoid q quadrate qj quadratojugal sa surangularsq squamosal sr sclerotic ring STF supratemporal fenestra Arrows point at dentary teeth


premaxillary-maxillary diastema in the holotype ofCompsognathus longipes (BSPG AS I 563) howeverPEYER (2006) has argued that such a feature (absent inMNHM CNJ 79) may be a preservational artifact Ourexamination of BSPG AS I 563 failed to corroborateOSTROMrsquos claim as he pointed out the ends of thepremaxilla and maxilla are poorly preserved A dia -stema is absent in Sinosauropteryx (CURRIE amp CHEN

2001) Ornitholestes (OSBORN 1916) Scipionyx (DAL

SASSO amp SIGNORE 1998) and Huaxiagnathus (HWANG

et al 2004) and it is most likely that this feature wasalso absent in Sinocalliopteryx and Compsognathus(Fig 11) A distinct indentation is present on theanterior dentigerous margin of Juravenator betweenits second and third maxillary teeth (Figs 8 10) Thisfeature appears to be somewhat exaggerated by boneweathering but we believe it to be real and unique toJuravenator among ldquocompsognathidsrdquo Posterior tothis indentation the ventral margin of the maxilla inlateral view is straight throughout the extension of thetooth row becoming slightly concave behind the lasttooth On the caudoventral corner of the maxilla thecontact between this bone and either the lacrimal orthe jugal cannot be distinguished

The maxilla does not seem to form part of thedorsal margin of the antorbital fossa this marginappears to be formed solely by the nasal and lacrimal(Fig 8) There is no evidence indicating that the antor-bital pneumaticity extended into the nasal ndash this boneappears to just form the dorsal margin of the antorbitalfossa A thin maxillary sheet ndash the medial lamina ofthe ascending ramus of the maxilla ndash lines the anteriorfourth of the antorbital fossa The center of thisportion is punctuated by a round maxillary fenestrahowever due to poor preservation the existence of apromaxillary fenestra cannot be determined A narrowportion of bone extends from the base of the mediallamina of the ascending ramus caudoventrally as along inset that defines the ventral margin of the antor-bital fenestra

The nasals are exposed in laterodorsal (right ele-ment) and dorsal (left element) views These bonesform the caudal and dorsocaudal margins of theexternal nares (Fig 8) The internarial process of thenasal extends for about two-thirds the length of thenares The premaxillary process of the nasal forms abroad arch the base of which forms the caudal marginof the nares and its tip runs between the premaxillaand the maxilla The main body of the nasal becomesgradually wider towards the caudal end of the boneThe lateral and medial margins are straight and

Fig 9 Comparisons of the premaxilla and lacrimal of Jura-venator starki (A) Sinosauropteryx prima (B based onCURRIE amp CHEN 2001 C based on GMV 2124) Compso-gnathus longipes (D based on PEYER 2006) Ornitholesteshermanni (E based on AMNH 619) Scipionyx samniticus(F based on DAL SASSO amp SIGNORE 1998) and Huaxiagna-thus orientalis (G based on HWANG et al 2004) Abbrevia-tions l lacrimal pm premaxilla Drawings not to scale

the latter forms a long suture with its counterpartDorsally the surface of the nasal is essentially flatalthough the central portion of the lateral edge isslightly swollen The caudal end of this bone contactsthe frontal in a more or less transversally orientedsuture although the precise shape of this suture is notclear

The lacrimals are exposed in both lateral (rightelement) and medial (left element) views This bonehas the shape of an inverted ldquoLrdquo and as in otherldquocompsognathidsrdquo it lacks any evidence of a ldquohornrdquoor dorsal protuberance (Fig 8) The inverted ldquoLrdquo shapeof this bone more closely resembles the lacrimals ofScipionyx (DAL SASSO amp SIGNORE 1998) and Comp-


Fig 10 Reconstructions of the skulls ofJuravenator starki (A) Compsognathuslongipes (B based on PEYER 2006)Scipionyx samniticus (C based onDAL SASSO amp SIGNORE 1998) andOrnitholestes hermanni (D based onAMNH 619) Hatched areas indicatemissing portions of the skullDrawings not to scale


sognathus (PEYER 2006) than that of Sinocalliopteryx(JI et al 2007a) Sinosauropteryx (GMV 2124 persobs) and basal tyrannosauroids (XU et al 2004SERENO et al 2009) in which the dorsocaudal cornerextends more caudally (ie it has a more ldquoTrdquo-shapedappearance) (Fig 9) On the lateral surface a minuteswell marks the caudodorsal corner of the bone andanterior to this swell there is a small pit that we inter-pret as a pneumatopore (lacrimal fenestra) a featurewidely distributed among non-avian theropods(RAUHUT 2003) If our interpretation is correct thepresence of a pneumatic lacrimal in Juravenatorwould distinguish this taxon from basal coelurosaurssuch as Compsognathus (PEYER 2006) Sinosauro -pteryx (CURRIE amp CHEN 2001) Scipionyx (DAL SASSO

pers comm) and apparently Sinocalliopteryx (JI et al2007a) in which the lacrimal is apneumatic

The nasal (horizontal) ramus of the lacrimal is sub-equal in length to the jugal (vertical) ramus this con-dition is similar to that of Compsognathus (MNHMCNJ 79) but different from those of Scipionyx (DAL

SASSO amp SIGNORE 1998) and Sinosauropteryx (GMV2124) in which the nasal ramus is either shorter orlonger than the jugal ramus respectively (Fig 9) Thenasal ramus gradually tapers rostrally ending in asharp tip This ramus articulates medially with thenasal and laterally forms the dorsal margin of thecaudal half of the antorbital fenestra The jugal ramusis stout and hourglass-shaped with expanded dorsaland ventral ends This overall appearance of the ramusis comparable to those in Scipionyx and Compso -gnathus but clearly stouter than the slender jugalramus of Sinosauropteryx (Fig 9) The ventral half ofthe jugal ramus is laterally concave the surface isexcavated by an ample trough caudally defined by abroad and prominent rim that forms the caudal marginof the bone Such a trough marks the caudal extensionof the antorbital fossa a condition also present in avariety of other non-avian theropods (eg OSTROM

1978 CURRIE amp ZHAO 1993 XU et al 2004 2006CARPENTER et al 2005a CORIA amp CURRIE 2006) Themedial surface of the lacrimal is essentially flat Thecontact between the lacrimal and the jugal is a sub -horizontal suture

The right jugal is exposed in lateral view a frag-ment of what appears to be the left element probablyin medial view is visible in the caudal corner of theleft orbit The jugal is slender it is dorsoventrallythinner at the center of the orbit (Fig 8) The dorsaland ventral margins of this bone are concave andstraight respectively Caudally the postorbital ramusis well developed and projected dorsocaudally at a 120degree angle with respect to the horizontal body of thejugal (ie the suborbital ramus) The postorbitalramus tapers to a sharp end at the midheight ofthe orbit The anterior margin of this ramus forms aslightly concave contact with the jugal ramus of thepostorbital Caudally the jugal ends in a tapering pro-cess (subtemporal ramus) shortly behind the base ofthe postorbital ramus There is no apparent evidenceof jugal pneumaticity The overall morphology of thejugal agrees well with that described for Compso -gnathus (OSTROM 1978 MICHARD 1991) Sinosaurop-teryx (CURRIE amp CHEN 2001) Scipionyx (CURRIE ampCHEN 2001) and Huaxiagnathus (HWANG et al 2004)However the subtemporal ramus of this bone is shor-

Fig 11 Dentition of Juravenator starki (JME Sch 200)Tooth row (A) and detail (B C) of middle premaxillaryteeth These teeth are interpreted to occupy the second andthird positions of a probably four-toothed premaxilla

ter than those of Huaxiagnathus and SinosauropteryxWe interpret a slender bone rimming the dorsocra-

nial half of the right orbit as the prefrontal Althoughit is unclear how much of this bone is actually pre -served if our interpretation is correct this bone pre-vents most if not all contribution of the frontal to theorbit (Fig 8) Comparable ossifications are preservedrimming the orbits of Compsognathus and Scipionyx(Fig 10)

The two frontals are preserved in dorsal view Thisbone is long and dorsally flat has a transversal contactwith the nasal and a straight suture with its counter-part (Fig 8) Caudally the width of the frontal in -creases gradually however its contact with theparietal is not clear on either side ndash we have inter -preted a transversal line roughly perpendicular to theinterfrontal suture as the frontoparietal contact (Fig8) If this is correct the frontal participates exten -sively of the supratemporal fenestra a condition simi-lar to that apparently present in Huaxiagnathus(HWANG et al 2004) At the same time the fronto -parietal suture interpreted here more closely resem-bles the sigmoid suture of the French Compsognathus(MNHM CNJ 79) than the very forked frontoparietalsuture featured by OSTROM (1978) for the holotype(BSPG AS I 563)

The postorbital is exposed on lateral (right element)and medial (left element) views This T-shaped bonehas a frontal ramus slightly longer than the squamosalramus (Fig 8) These two rami are not at the samelevel while the squamosal ramus points caudally thefrontal ramus is somewhat dorsorostrally directed Thedorsal margin of the right element at the junction ofthese two rami exhibits a small indentation Such anindentation is absent in Scipionyx in which the post -orbital also differs by having the frontal and squamo-sal rami essentially aligned The tapering jugal ramusis longer than the other two rami and it gently curvesrostroventrally its convex caudal margin contactsnearly the entire length of the postorbital ramus of thejugal and its rostral margin forms the concave caudalmargin of the orbit

Portions of both squamosals are preserved Theright element is exposed laterally and partially dor -sally ndash this bone appears to be displaced caudoven -trally ndash and the left element is represented only by itspostorbital process exposed in medial view The squa-mosal is a multibranched bone and its complexmorphology is best interpreted under UV light (Fig8) Two subvertical processes define a broad ventralnotch most probably for the articulation of the qua-

drate The caudalmost of these projections is a pointedprocess that presumably contacted the paraoccipitalprocess although the latter is not preserved The otherprocess the quadratojugal process is a longer andalso tapering projection the tip of which is overlappedby a fragment of a bone interpreted as a remnant of thequadratojugal A third process projects mediodorsallyfrom the caudal margin of the squamosal and appearsto abut the skull roof this process probably forms thecaudal margin of the supratemporal fenestra A fourthprocess is directed rostrally This postorbital process isbest preserved on the left element even if viewedmedially The postorbital process ends in a fork thatreceives the squamosal ramus of the postorbital boneand together with this ramus forms the intertemporalbar (division between supra and infratemporal fene-strae)

Less certain anatomical information is available forthe quadratojugal Only what we interpret as the tipsof the squamosal and jugal processes of the rightelement are preserved (Fig 8) If correctly identifiedthe preserved fragment of the squamosal processwould suggest that the quadratojugal contacted thesquamosal behind the infratemporal fenestra Like-wise if this interpretation is correct it would indicatethat the squamosal process of the quadratojugal ofJuravenator is much thinner than the broad equivalentof Sinosauropteryx (GMV 2124 pers Obs) The por-tion interpreted as the jugal process is slender andtogether with the squamosal process gives the bone anldquoLrdquo shape (Fig 9)

The parietals are also poorly preserved with bothof them in dorsal view The length of this bone appearsto be roughly one-third the length of the frontal (Fig8) Dorsally it is slightly convex as in Compso -gnathus (OSTROM 1978 MICHARD 1991 PEYER 2006)and Scipionyx (DAL SASSO amp SIGNORE 1998) a sagittalcrest running along the skull roof was clearly absentThe caudal margin of both parietals is defined by athick rim that forms the nuchal crest which appears tobe stronger than the weak crest of Compsognathus(BSPG AS I 563)

A long and very poorly preserved bone situated inthe caudolateral corner of the left portion of the skullis interpreted as a portion of the quadrate (Fig 8)This bone exhibits a concave outer surface and underUV light what appears to be a terminal articulationOur identification of this bone is based on both itslocation and its elongate shape

The vaulted rostral portions of the palatines(vomerine processes) are among the few portions of



the palate that can be determined with certainty Thesebones preserved within the antorbital fenestra arebest seen under UV light (Fig 8) The vomerineprocess of the right palatine is in place and situatedwithin the caudal half of the fenestra Its left counter-part is displaced protruding towards the right side ofthe skull at the center of the antorbital fenestra Asin other non-avian theropods (eg MADSEN 1976CURRIE amp ZHAO 1993 DAL SASSO amp SIGNORE 1998NORELL amp MAKOVICKY 2004) the vomerine pro -cesses of the palatines have a strongly concave rostralmargin that forms the caudal border of the choanaand a caudorostrally expanded dorsal (top) marginThese processes most likely contacted each other butthis cannot be ascertained due to the displacement ofthe left one A long and slender bone rostrocaudallycrossing the orbit probably represents a portion of oneof the pterygoids although no morphological detailsof these bones are visible

Very little anatomical information is available forthe braincase which is partially exposed in dorsalview The supraoccipital has a pronounced midlineridge and the paroccipital processes are robust andtall The latter project posteroventrally at approximate-ly 45 degrees and their posterior surface extends overa single plane A thin sliver of bone which may re -present the stapes anteriorly lines each paraoccipitalprocesses

Mandibular bones ndash Only the lateral side of theright mandible is exposed and its dorsal border is over-lapped by the ventral margin of the cranium Howeverportions of some mandibular teeth have been exposedby preparation of the lateral surface of the dentaryThe first exposed tooth is aligned with the brokenbase of the last premaxillary tooth (Figs 8 10) andthe latter covers the tip of the former The last exposeddentary tooth lies in front of the last two maxillaryteeth however it is not possible to determine whetherthe mandibular dentition ended in front (as in manynon-avian theropods) or behind (as in Scipionyx) theupper tooth row Although the number of mandibularteeth cannot be ascertained with precision the size ofthe few exposed suggest a tooth count significantlysmaller than the 18 teeth counted by OSTROM (1978)or the 21-22 counted by MICHARD (1991) and PEYER

(2006) for Compsognathus Assuming the mandibulartooth row to be slightly shorter than the upper toothrow as in Compsognathus (OSTROM 1978) and mostother theropods (eg MADSEN 1976 COLBERT 1989WEISHAMPEL et al 2004) we estimate that the number

of dentary teeth did not exceed 11 a number that moreclosely resembles our estimate of the tooth count ofScipionyx (12-14) Likewise the lower number of den-tary teeth of Juravenator when compared to Compso-gnathus can also be inferred by the fact that the lengthof the dorsal margin of its dentary is significantlyshorter OSTROM (1978) reconstructed the dorsal mar-gin of the dentary of Compsognathus as more thanhalf the length of the skull while it is less than half thelength of the skull in Juravenator (and in Scipionyx)Allometric studies on tyrannosaurids (CURRIE 2003)and other non-avian theropods (eg MADSEN 1976COLBERT 1989) have failed to provide evidence in -dicating an increase in the number of teeth in relationwith either size or age However the number ofdentary teeth increases from 18 to 21-22 in Compso-gnathus when the smaller holotype is compared tothe larger French specimen (OSTROM 1978 MICHARD

1991 PEYER 2006) Individual variations notwith -standing (COLBERT 1989 CURRIE 2003 PEYER 2006)ndash and taking into account that the increment in toothcount in Compsognathus may suggest a similar in -crement in Juravenator ndash we still regard this differ -ence as taxonomically important

The lateral height of the rostral portion of themandible is approximately two-thirds the mandibularheight at the level of the orbit (Fig 8) this ratio isabout one-third in Sinosauropteryx (GMV 2124 persobs) and one-half in Scipionyx and basal tyranno -sauroids (XU et al 2004 2006) The lateral surface ofthe mandible is essentially flat along its rostral halfand slightly convex more caudally The lower jawlacks a mandibular foramen a condition similar tothat in Compsognathus (OSTROM 1978) Huaxiagna-thus (HWANG et al 2004) Sinosauropteryx (CURRIE ampCHEN 2001) Sinocalliopteryx (JI et al 2007a) andScipionyx (DAL SASSO pers comm) The ventralmargin of the dentary is nearly straight (Fig 8) Agently and long caudoventrally slanting suture sepa -rates this bone from the surangular and the angularAlthough the postdentary portion of the mandible ispartially covered by the jugal it can be seen that thedorsoventral width of the surangular is about threetimes that of the angular The ventral margin of thesurangular bears a faint longitudinal rim that forms aledge at the contact with the angular The latter bone isnarrowly exposed laterally The rear portion of themandible is missing beyond the surangular and angu-lar and it is difficult to evaluate whether the caudalportions of these bones are missing as well

Cranial openings ndash The external nares are de -fined by the premaxilla dorsally anteriorly and ven-trally and the nasal posteriorly They are subellipticalin shape with a straight ventral margin and concavedorsal margin (Figs 8) The external nares of Jura -venator end substantially in front (~4 mm) of the an -terior margin of the antorbital fossa This conditionseems to contrast with that of other ldquocompso -gnathidsrdquo The caudal end of the nares of Scipionyx(DAL SASSO amp SIGNORE 1998) and Compsognathus(PEYER 2006) is aligned with the anterior margin ofthe antorbital fossa and HWANG et al (2004) describedthe long nares of Huaxiagnathus as overlappingposteriorly the antorbital fossa (yet this is not entirelyapparent in the illustrations provided by HWANG et al2004)

The antorbital fossa contains what appear to be around and small maxillary fenestra and a large antor-bital fenestra Whether Juravenator also possessed apromaxillary fenestra is uncertain The latter is bor -dered by the lacrimal and nasal dorsally the maxillarostrally and ventrally and the lacrimal caudallywhether or not the jugal contributed to the antorbitalfossa and perhaps the fenestra is unclear (Fig 8) Theantorbital fenestra is suboval although its verticalcaudal margin defines a nearly 90˚ angle with itsventral counterpart The length of this fenestra isapproximately three-fourths the length of the antor -bital fossa ndash proportionally longer than in Scipionyxand Compsognathus (fenestrafossa ratio is about 23)and Sinosauropteryx (fenestrafossa ratio is between23 and 12) (Fig 11)

Both orbits are preserved the left one is incom pleteand the right one is slightly collapsed by the ventraldisplacement of the right frontal However it is clearthat the orbit was round and bordered by equal con -tributions of the lacrimal jugal postorbital and whatwe interpreted as the prefrontal (Figs 8 11) It is notclear if the frontal comprised part of the dorsal marginof the orbit or whether a postorbital-prefrontal contactprevented a frontal contribution Most remarkable isthe fact that the length of the orbit is approximatelythe same as that of the antorbital fenestra a conditionsimilar to that of basal tyranosauroids (XU et al 20042006) but an obvious difference with Sinosauropteryx(CURRIE amp CHEN 2001) Compsognathus (BIDAR etal1972OSTROM 1978MICHARD 1991) Ornitholestes(OSBORN 1916) and Scipionyx (DAL SASSO amp SIGNORE

1998) in which the antorbital fenestra is substantiallyshorter (Fig 11) The rostral half of the sclerotic ringis preserved within the right orbit but neither the

shape nor the number of the scleral oscicles can bedetermined

Both of the temporal fenestrae are distorted Ourinterpretation of the quadratojugal suggests an in -fratemporal fenestra enclosed caudally by a squa -mosal-quadratojugal bar and with a subrectangularshape taller than wide (Fig 8) On the left side theskull roof and the squamosal-postorbital bar surrounda large fenestra This opening is likely the supra -temporal fenestra in dorsal view but its shape and sizemay be modified by the fact that the latter bar isclearly displaced and viewed medially

Dent i t ion ndash As stated above (see Cranial bones)three premaxillary and eight maxillary teeth arepreserved in the upper jaw the premaxilla may havecarried an additional tooth but this is unlikely to be thecase for the maxilla The tooth count of the dentary isunknown (see Mandibular bones) but we estimate it tobe no more than 11

The first preserved premaxillary tooth (here inter-preted as occupying the second position) is less com-pressed than the others (Fig 10) like the ldquoincisiformrdquopremaxillary teeth of other non-avian theropods(SANDER 1997) The third maxillary tooth is the lar-gest of the upper series This tooth aligns with therostral margin of the antorbital fossa (Figs 8 10)Similarly large teeth located at the rostral margin ofthe antorbital fossa are present in Huaxiagnathus andCompsognathus (MNHM CNJ 79) This large tooth isfollowed by two other large teeth thus indicating thatthe largest teeth are located in the middle portion ofthe maxilla similar to the condition in Scipionyx (DAL

SASSO amp SIGNORE 1998) (Fig 10)All preserved teeth (from both the upper and lower

jaws) lack mesial serrations The first preserved pre-maxillary tooth also lacks distal serrations but theseare present in the second preserved premaxillary tooth(the third one is highly incomplete) (Fig 10) Amongbasal coelurosaurs serrated premaxillary teeth areknown in the ldquocompsognathidrdquo Sinocalliopteryx (JI

et al 2007a) and tyrannosauroids (XU et al 2004)although the premaxillary teeth of many basal coelu-rosaurs (and a number of tetanurans) lack serrations(OSTROM 1978 CURRIE amp CHEN 2001) The serrationsof the premaxillary teeth of Juravenator and Sinocalli-opteryx are much finer (~ 13mm) than those of themaxillary teeth (~ 8mm) All other exposed teeth ofJuravenator also have distal serrations As in Compso-gnathus (BSPG AS I 563) basal tyranosauroids (XU

et al 2004 2006 SERENO et al 2009) and possibly



Huaxiagnathus (HWANG et al 2004) and Sinocalli -opteryx (JI et al 2007a) the maxillary and posteriordentary teeth of Juravenator are more recurved thanthe premaxillary teeth (however minimal informationis available for the dentary teeth) This appears to be adifference between these taxa and either Sinosauro -pteryx or Scipionyx where the teeth are more homo -genous in shape

52 Postcranium

Axial skeleton ndash The vertebral column is essen -tially complete and articulated from the skull to thedistal third of the tail The cervical and dorsal seriesare exposed primarily in dorsal view (Fig 12) thesacral region is dislocated due to the rotation of theskeleton at this region and the caudal series is ex -posed on right lateral view

Fig 12 Cervical (A) and dorsal (B) vertebral series of Juravenator starki (JME Sch 200) Arrows point to a very thincervical rib Abbreviations dv dorsal vertebrae hu humerus il ilium ns neural spine pn pneumatic pore sc scapula

The cervical region is not well-preserved espe -cially its cranial half The number of cervicals cannotbe established with precision because no sternum hasbeen preserved (thus the separation between cervicaland dorsal vertebrae cannot be determined on thebasis of the articulation of ribs to this bone) andbecause most cervical ribs are missing Seven verte-brae are preserved between the first one preserving along rib of thoracic (dorsal) morphology (Figs 3-512) and a fragmentary element (presumably a cranial-most cervical) adjacent to the caudal portion of theskull However the possibility of an unaccountedcervical at the junction of the skull block with thebody block cannot be ruled out this region was da -maged as a result of the breakup of the entire slab intwo blocks (Fig 3) All of this suggests the presence ofnine to ten cervicals including the atlas which it is notdiscernable from the mass of bones exposed imme -diately behind the skull

When visible the cervical vertebrae have short andlow neural spines (Fig 12) However those in themiddle (for example the third vertebra in front of thefirst dorsal) appear to have taller neural spines thatthe caudalmost elements The diapophyses are well-developed The lengths of these processes increasegradually from the cranialmost elements to the caudal-

most ones Each zygapophysis (pre- and postzygapo-physes) is well-spaced with respect to its counterpartIn the caudal portion of the series the postzygapo -physes decrease in length towards the back The thirdcervical in front of the first dorsal exhibits a smallround foramen piercing the center of the centrum ndashthis is the only possible evidence of pneumaticity inthe cervical series (Fig 12) There is no informationavailable about the central articular surfaces

There are at least 13 dorsal vertebrae all of themexposed dorsally with the exception of the last onewhich is exposed on its right side The dorsal verte-brae have well-developed transverse processes thatslightly angle posteriorly (Fig 12) The neural spinesof these vertebrae are generally poorly preserved Afew vertical ridges crushed against the center of somecentra are interpreted as remains of neural spines butthese provide minimal anatomical informationNonetheless the craniocaudal extension of the neuralspines of the posterior dorsals appears to be large insome instances these ridges are preserved projectingcaudally between the postzygapophyses The morpho-logy of the neural spine is clearer in the last dorsalvertebra In this vertebra the caudal margin of theneural arch and the posterior portion of the neuralspine are exposed in lateral view apparently over -


Fig 13 Photograph (A) and interpretive drawing (B) of the pelvic region of Juravenator starki (JME Sch 200) Abbre -viations af antitrochanteric facet bf brevis fossa dv dorsal vertebra fe femur il ilium is ischium pp pubic pedicelpu pubis sv sacral vertebra r and l refer to the right and left element


lapping the anteriormost portion of the left preaceta-bular wing Although the neural spine of this vertebrais not exposed completely the observable portionshows the distal expansion characteristic of the neuralspines of the posterior dorsals of Compsognathus(OSTROM 1978 PEYER 2006) Huaxiagnathus (HWANG

et al 2004) Scipionyx (DAL SASSO pers comm) Sino -sauropteryx (CURRIE amp CHEN 2001) Sinocalliopteryx(JI et al 2007a) and the Brazilian compsognathidMirischia (NAISH et al 2004) The caudal articularsurface of the last dorsal centrum is flat thus sug-gesting that dorsal vertebrae had amphiplatyan centraThe exposure of the vertebrae prevents us from de -termining whether there is any evidence of pneu-maticity

The number and general morphology of the sacralvertebrae is also problematic (Fig 13) The sacrumappears to be disarticulated and is partially covered byother bones A number of unidentifiable bones arealso exposed in this area The fact that none of thebones identified as sacral centra are coossified to oneanother suggests that the sacrum was not fused at thetime the animal died ndash the complete lack of fusionbetween sacral vertebrae may be another feature sug-gesting an early ontogenetic age for the holotype ofJuravenator A laterally exposed centrum locatedimmediately caudal to the thirteenth dorsal vertebra isregarded as part of the first sacral vertebra The cen-trum is substantially larger than that of the last dorsaland it has a subquadrangular shape The interpretationof this centrum as that of the first sacral is based onthe fact that (1) it already lies within the realm of theilia and (2) its lateral surface is broadly recessedresembling the condition of a centrum lying betweenthe two ilia (a vertebra more clearly identified assacral) The latter centrum is crushed against themedial surface of the postacetabular wing of the leftilium The lateral surface of this centrum is distinctlyrecessed and a ridge defines the periphery of the cen-trum A disarticulated centrum of similar morphologylies ventral to the postacetabular wing of the left iliumThe lateral recess of this centrum is very distinct andits caudal articular surface is rounded and flat It isunclear whether this centrum corresponds to anothersacral vertebra or to the first caudal A gap betweenthe articulated portion of the tail and the sacral cen-trum crushed against the medial surface of the posta-cetabular wing suggests that the disarticulated cen-trum beneath the left ilium could belong to the firstcaudal (Fig 13) However if this is accepted the cen-trum of the first caudal would have greatly differed

from those of the succeeding tail vertebrae Compari-sons with ldquocompsognathidsrdquo and other basal coeluro-saurs are problematic because pelvic bones oftencover the sacral vertebrae Nonetheless the subqua-drangular shape of the first sacral vertebra of Jura -venator and the fact that its sacrals are substantiallybigger than both dorsal and caudal vertebrae resem-bles the condition present in Scipionyx Howeverthese features contrast with those in Mirischia inwhich the sacrals are smaller than the dorsals (centraare shorter and more depressed) and their centra arenot laterally recessed (NAISH et al 2004)

All preserved caudals are articulated and exposedin right lateral view (Figs 3-5 14) The tail includes44 vertebrae preserved in articulation but it is unclearwhether the centrum of the first caudal is missing (seeabove) Comparisons of the number of caudal verte-brae and relative tail length with other basal coeluro-saurs are hampered by the incomplete nature of thetail of most taxa (eg OSTROM 1978 DAL SASSO ampSIGNORE 1998 KIRKLAND et al 1998 DE KLERK et al2000 HWANG et al 2004 NAISH et al 2004) How -ever meaningful comparisons can be established withthe tail of the holotype of Sinosauropteryx (NIGP127586 CHEN et al 1998 CURRIE amp CHEN 2001) Thenearly complete tail of NIGP 127586 preserves 64caudal vertebrae In this specimen the last 20 verte-brae represent approximately 44 of the total lengthof the tail and the entire length of the tail is approxi-mately 170 of the snout-vent length (measured fromthe tip of the snout to the caudal end of the ilium) Thepreserved portion of the tail of Juravenator is approxi-mately 125 of the snout-vent length but when theproportions observed in Sinosauropteryx (ie that thelast 20 vertebrae constitute 44 of the tail) are extra-polated to the Schamhaupten fossil the length of thetail of Juravenator is estimated to represent about180 of the snout-vent length This extrapolation iswarranted by the fact that the ratio between the com -bined length of the first 43 caudal vertebrae and thesnout-vent length of NIGP 127586 is approximately118 a value slightly shorter than the one observedin Juravenator All of this suggests that the tail ofJuravenator was probably as long if not longer thanthe tail of Sinosauropteryx ndash the proportionally longestknown tail within theropods (CHEN et al 1998)

The length of the caudal centra remains more orless constant (lt 1mm in difference) until the 17th

caudal when the length of the centrum starts increa-sing (Fig 14) At the level of the 27th caudal thelength of the centrum reaches a maximum plateau and

it remains as long with minor variance until the 35th

caudal Distal from this point the length of the caudalcentra becomes gradually shorter but the last pre -served vertebra is still longer than any of the firstseventeen elements of the tail This pattern is similarto that of Huaxiagnathus in which the length of thecaudal centra remains more or less constant until the14th caudal and then increases to a higher plateau forthe subsequent 11 caudals (only 25 caudals are pre -served in this taxon [HWANG et al 2004]) It alsoapproaches that of the compsognathid Sinocalli -opteryx (JI et al 2007a) in which the longest caudalsare between the 20th and the 23rd vertebrae This trendhowever is markedly different from that of Sinosaur-opteryx in which the length of the centra increasesuntil the sixth caudal but then decreases until the endof the tail (CURRIE amp CHEN 2001) It is also differentfrom the one observed in Compsognathus (BSPGAS I 563 and MNHN CNJ 79 preserve approximately

15 [OSTROM 1978] and 30 [MICHARD 1991] elementsrespectively) and Scipionyx (less than 10 vertebraepreserved) in which the length of the centra graduallyincreases throughout the preserved portion of the tail

The centra are elongated and morphologically verysimilar throughout the entire length of the tail (Fig14) The height of the centrum decreases graduallythroughout the tail in such a way that in the proximalcaudals the mid-height of the centrum is approximate-ly half the length of the centrum (with an average of55 for the first five preserved caudals) and in thedistal caudals the height is nearly one-quarter of thecentral length (with an average of 27 for the firstfive preserved caudals) The general depression of thecaudal centra towards the tailrsquos distal end is sharedby Compsognathus (PEYER 2006) but not by Huaxia -gnathus in which the caudal centra remain approxi-mately the same relative height throughout the 25preserved elements of its tail (HWANG et al 2004) The


Fig 14 Caudal vertebral series (A) of Juravenator starki (JME Sch 200) with close-ups of the distal (B) middle (C)and proximal (D) portions of the tail Abbreviations cv caudal vertebrae tp transition point


neural spines are long and caudally oriented in thefirst three caudals of Juravenator ndash the neural spine ofthe third caudal extends back to the middle of thefourth caudal The neural spine of the latter vertebra isshorter but still somewhat inclined its morphologytransitions with the shorter and vertically orientedneural spines of the subsequent vertebrae The neuralspine of the fifth caudal is dorsally expanded and fan-

shaped and this condition appears to remain presentin more distal neural spines (however the morphologyof some of these processes is not clear in every verte-bra) The transverse processes are poorly preservedThey appear to be well-developed in the proximal cau-dals but their presence in more distal caudals cannotbe determined The poor preservation of the transverseprocesses and limited preparation in areas covered bysoft tissue (see Soft tissue below) complicates theassessment of the transition point However remnantsof a moderately tall neural spine are visible down tothe 14th caudal It is difficult to determine if a neuralspine of any significant height was present in the 15th

caudal but it seems clear that the neural spines werereduced from the 16th caudal onwards Thus weregard the transition point to be around the 14th or15th caudal vertebrae (Fig 14) an estimate moreor less comparable to that provided for Compso -gnathus (MICHARD 1991) Huaxiagnathus (HWANG etal 2004) and other basal coelurosaurs (HOLTZ et al2004) None of the centra exhibit evidence of pneu -maticity

The proximal zygapophyses are not well-preservedHowever it seems apparent that they were short Themorphology of the zygapophyses can be better ob -served in the middle and distal portion of the tailAlthough still difficult to determine the morphologyof the zygapophyses of the middle caudals (roughlybetween the 16th and the 35th element) is strikinglyunique (Figs 14 15) These zygapophyses are talland slender relatively short and remarkably archedforming a ldquobowrdquo over the intercentral articulation Theexact contribution of the pre- and postzygapophysesto this bow-like articulation is unclear but bothzygapophyses seem to participate in this joint Themorphology of the zygapophyses becomes more con-ventional from the 36th caudal onwards when the pre-zygapophyses gradually become straighter and longeroverlapping substantially (but less than half) thelength of the preceding centrum

The chevrons are elongate and rod-shaped through-out the tail (Fig 14) There are 25-26 chevrons inarticulation with the first 25-27 vertebrae ndash the lastones are very small and difficult to see but a chevronappears to be present between the 31st and the 32nd

caudal The length of the chevrons decreases gra-dually towards the distal part of the tail The last fewchevrons are approximately one-third the length of theproximal ones (eg the 6th chevron is 88 mm longand the 22nd chevron is 30 mm long) While the pro-ximal chevrons tend to be caudally bowed the middle

Fig 15 Comparisons of middle caudals (vertebrae 19 and20) of Juravenator starki (A) Compsognathus longipes(B based on PEYER 2006) Sinosauropteryx prima (C basedon CURRIE amp CHEN 2001) and Huaxiagnathus orientalis(D based on photographs of CAGS-IG02-301 provided byS Hwang) Abbreviations ha haemal arch

ones are straight These bones are articulated ob -liquely until the 20th chevron but they lay parallel tothe vertebral central onwards The morphology andorientation of the chevrons of Juravenator is com -parable to that of Huaxiagnathus (HWANG et al 2004)Compsognathus (MICHARD 1991) Sinocalliopteryx(JI et al 2007a) and as far as it can be comparedScipionyx (DAL SASSO amp SIGNORE 1998) The chevronsof Huaxiagnathus and Compsognathus differ some-what from the more spatulated ones of Sinosauro -pteryx (CURRIE amp CHEN 2001) ndash the middle chevronsof the latter taxon are also hooked as opposed tostraight as in Juravenator

Only a few cervical ribs are preserved although theone corresponding to the penultimate element of theseries is best exposed These long and extremely thinribs exhibit the characteristic ldquohair-likerdquo condition(Fig 12) of the cervical ribs of Sinosauropteryx(CURRIE amp CHEN 2001) Compsognathus (PEYER

2006) Sinocalliopteryx (JI et al 2007a) and Scipionyx(DAL SASSO pers comm) Twelve pairs of thoracicribs are preserved in near articulation possibly repre-senting the complete ribcage (Figs 3-5 12) The ribs

are composed of a single element (ie without sternalsegments) and they show no evidence of ossified un -cinate processes They become thinner and shortertowards the caudal portion of the ribcage Althoughthose from the left cranial half are straight the gentlycurved outline of the ribcage is better insinuated onthe right side

A full set of partially articulated gastralia is pre -served in front of the pelvis (Figs 3-5 16) There aremore than 13 rows of gastralia a condition appro -aching the 15 or so described for Sinocalliopteryx (JI

et al 2007a) The first row is the thickest and a pair ofsingle gastralia forms it These elements connect attheir medial end forming a short cranial process Twopairs of filament-like segments articulated in a zigzagpattern form subsequent rows of gastralia Like in allother theropods (CLAESSENS 2004) each of these pairsis formed by two overlapping elements When visiblethe medial segments are longer than the lateral one acondition similar to that reported in Sinocalliopteryx(JI et al 2007a) and Huaxiagnathus (HWANG et al2004)


Fig 16 Close-up of the gastralia of Juravenator starki (JME Sch 200) Abbreviations il ilium


Appendicular skeleton ndash Both scapulae are pre -served in approximately their life-orientation ndash nearlyperpendicular to the vertebral column and with theblade diagonally crossing the rib cage (Figs 3-5 17)The left scapula is entirely exposed in lateral view Itscounterpart is beneath several ribs the left scapularblade and the vertebral column and only portions ofits medial surface can be seen between these bonesThe scapula is a slender and elongate bone and theblade expands gradually towards its dorsal end (Fig17) Its length is approximately ten times its width atmid-shaft and 15 times the length of the humerus(Fig 18) ndash while the former proportion is comparableto Sinosauropteryx and Compsognathus the latter ismuch greater than that in Compsognathus (scapulahumerus is 1 in BSPG AS I 563 OSTROM 1978) butsimilar to that in Sinosauropteryx (Fig 19) The pro-portion of the length of the scapula to that of thehumerus is 13 in Sinocalliopteryx a ratio thus inter-mediate between Juravenator and Compsognathus(JI et al 2007a) Unlike these taxa Scipionyx has astouter scapula ndash with an estimated length of aboutseven times the width of the mid-shaft ndash even if theproportion of this bone with respect to the humerus iscomparable to that of Compsognathus The scapula ofJuravenator differs from that of Sinosauropteryx aswell as from that of other ldquocompsognathidsrdquo by thefact that the narrowest portion of the scapula is at theneck as opposed to near the mid-shaft ndash in this respectthe scapula of Juravenator resembles more that of thebasal coelurosaurs Nqwebasaurus (DE KLERK et al2000) and Tyrannosaurus [BROCHU 2003 but notbasal tyrannosauroids such as Dilong (XU et al 2004)in which the scapula is very robust The lateral surfaceof the scapula is virtually flat although it is moreconcave in cross-section along the ventral third of itsblade (its basal portion) Based on what is exposed ofthe right scapula the medial surface (at least in itsbasal portion) appears to be convex In lateral viewthe scapular blade is curved caudally its cranial andcaudal margins are convex and concave respectively(Fig 17) In this respect Juravenator also differs fromother basal coelurosaurs (eg Scipionyx Sinosauro -pteryx Sinocalliopteryx Compsognathus Nqweba-saurus ornithomimids tyrannosauroids) where thescapular blade is straight or less visibly curved (Fig19) The scapula is more abruptly expanded ventrallyto form the glenoid and the coracoidal articular sur -face The former is well-exposed on the left side anddirected caudally as in ldquocompsognathidsrdquo (CURRIE ampCHEN 2001 HWANG et al 2004) and basal theropods

(MADSEN 1976 CURRIE amp ZHAO 1993 WEISHAMPEL

et al 2004) The glenoid is a concave facet about one-fifth of the width of the ventral end of the scapulasurrounded by a distinct rim which is more pro -nounced laterally The acromion is prominent ndash muchmore so than in Sinosauropteryx ndash and subtriangularin shape Unlike Sinosauropteryx Sinocalliopteryxand Huaxiagnathus this process projects abruptlyfrom the scapular blade and its dorsal margin is di -stinctly concave a condition that best approaches thatof Scipionyx (DAL SASSO amp SIGNORE 1998) Thelateral margin of the articulation for the coracoid ispartially weathered offering no information

Neither of the two coracoids is well exposed (Fig17) The left element is incomplete and only a portionof it is articulated with the left scapula The rightcoracoid appears to be better preserved but the leftscapula and ribs unfortunately cover it Nonethelessthe medial surface of this coracoid exhibits a distinctconcavity and a rounded convex caudoventral margin

We interpret two tiny (approximately 5 mm long)S-shaped and rod-like bones preserved in disarticu -lation next to the cranioventral end of the left scapulaas clavicles (Fig 17) The ventralmost of these bonesexhibits a hook-like end that is possibly a preser -vational artifact Although the same end is somewhatrecurved in the dorsalmost preserved element it isclearly not hooked as in its counterpart We interpretthis extremity to be the articulation with the scapulo-coracoid and the opposite medially curved end as themedian contact between these bones If this is correctthese tiny bones would have formed a U-shaped albeitnot fused ldquofurculardquo Fused clavicles forming a furculahave been reported in a great variety of non-aviantheropods (RAUHUT 2003 WEISHAMPEL et al 2004NESBITT et al 2009) including basal coelurosaurs(HWANG et al 2004 JI et al 2007a) but unfused cla-vicles are also known for the ceratosaur Segisaurus(CAMP 1936 CARRANO et al 2005 but see RAUHUT

[2003] for a different interpretation) (unfused cla-vicles are also known for the Cretaceous bird Hesper-ornis [MARSH 1880] and a number of modern birds[BAUMEL amp WITMER 1993]) The absence of clavi -cular fusion in Juravenator could well be ontogeneticalthough the relation between furcular formation andage is not well understood in non-avian theropods(eg the early juvenile Scipionyx has clavicles fusedinto a furcula) Regardless if our interpretation ofthese bones is correct the ldquofurculardquo of Juravenatorwould have been much more U-shaped ndash with a sub-stantially narrower interclavicular angle ndash than those

of other non-avian theropods (eg NORELL et al1997 MAKOVICKY amp CURRIE 1998 DAL SASSO ampSIGNORE 1998 CLARK et al 1999 HWANG et al 2002RAUHUT 2003)

The forelimb is 50 the length of the hindlimb(measured along their longest digit and including theircorresponding claws) (Figs 3-5) a ratio comparableto that of Huaxiagnathus (48 according to HWANG

et al [2004]) and Sinocalliopteryx (JI et al 2007a) and

smaller than that of the basal tyrannosauroid Guan-long (XU et al 2006) This proportion is substantiallygreater than the forelimb hindlimb ratio of Compso-gnathus and Sinosauropteryx (39 and 36 re-spectively according to HWANG et al [2004]) Theratio between humerus + radius and femur + tibiaof Juravenator is 043 slightly greater than that ofHuaxiagnathus (039 HWANG et al 2004) Compso-gnathus (041 for BSPG AS I 563 [HWANG et al 2004]


Fig 17 Photograph (A) and interpretive drawing (B) of the shoulder girdle and clavicles of Juravenator starki (JMESch 200) Abbreviations cl clavicles co coracoid sc scapula Note the intense pitting and grooving of the bone surfaceof the scapula (A) r and l refer to the right and left element


and 038 for MNHN CNJ 79) and Nqwebasaurus(040 using humeral and femoral length estimations ofDE KLERK et al[2000]) and much greater than that ofSinosauropteryx (029 and 031 for NIGP 127586 andNIGP 127587 respectively [JI et al 2007b]) ThusJuravenator has the longest forelimb among ldquocompso-gnathidsrdquo and such a difference does not seem to berelated to the elongation of the hand as in Huaxiagna-thus (HWANG et al 2004) The smaller contribution ofthe hand to the length of the forelimb in Juravenatorcan also be noticed when the length of the manus iscompared to the rest of the forelimb (Figs 18-19)While the ratio between the manus and the (humerus +radius) is 087 for Juravenator 084 for Sinosauro -pteryx (NIGP 127587 HWANG et al 2004) and 074for Compsognathus (holotype HWANG et al 2004) itis 103 for Huaxiagnathus (HWANG et al 2004)

Both humeri are preserved (Fig 18) The left one isexposed in caudal view and the right one in cranial

view however the latter is more flattened and partiallycovered by the ribcage The proximal end of thehumerus bears a convex head (in caudal view) and aweak internal tuberosity The deltopectoral crest isshort and triangular in shape with its caudal surfacebeing distinctly concave Unlike the long deltopecto-ral crest of Sinosauropteryx that originates at the pro-ximal end and extends for more than half the length ofthe humerus (CURRIE amp CHEN 2001 JI et al 2007b)the deltopectoral crest of Juravenator expands for lessthan 15 the length of the humerus and it does notreach the proximal end of the bone The humerus ofCompsognathus is not well preserved in any of thespecimens ndash the French specimen seems to have amore pronounced crest ndash but the deltopectoral crestdoes not appear to be like the short triangular crest ofJuravenator The abbreviated deltopectoral crest ofJuravenator sets this species aside from the majorityof non-avian theropods in which the deltopectoral

Fig 18 Photograph (A) and interpretive drawing (B) of the forelimbs of Juravenator starki (JME Sch 200) AbbreviationsmcII-III metacarpals II-III hu humerus il ilium ra radius sc scapula ul ulna I-III manual digits I-III R and l refer tothe right and left element

crests often extend for a third or more of the length ofthe humerus (eg MADSEN 1976 OSTROM 1978 DE

KLERK et al 2000 CURRIE amp CHEN 2001 BROCHU

2003 RAUHUT 2003) However we cannot be sure asto the extent to which this condition may be relatedto immaturity since the early juvenile holotype ofScipionyx (thus far the only know specimen) alsoexhibits a faint deltopectoral crest

A few millimeters distal to the deltopectoral creston the caudolateral margin of the shaft there is ashort ridge-like structure that could have been theinsertion of the humeroradialis muscle (MADSEN

1976) The distal end expands gradually from the mid-shaft This end exhibits a slight degree of torsion withrespect to the proximal end (both ends are not ex -panded on the same transversal plane) This torsion


Fig 19 Comparisons between the shoulder girdle and forelimb of Juravenator Sinosauropteryx (based on CURRIE amp CHEN

2001) Compsognathus (based on PEYER 2006) and Scipionyx (based on DAL SASSO amp SIGNORE 1998) scaled to the lengthof the humerus


may account for the presence of an inflated margin onthe distal end of the right humerus which constitutesthe medial margin of the bone itself Comparing thedegree of humeral torsion with that of other basalcoelurosaurs is hampered by the bi-dimensional pre-servation of most of these fossils However CURRIE ampCHEN (2001) described the humerus of Sinosauro -pteryx as lacking any kind of torsion and a similarcondition appears to be the case in Scipionyx (DAL

SASSO amp SIGNORE 1998) The caudal surface of thedistal end is slightly excavated by a triangular fossaThe external (ulnar) condyle is well developed andhemispheric the internal (radial) condyle is not suffi-ciently exposed to be described

Both ulnae and radii are preserved (Fig 18) butthose from the right side are largely covered by ribsand gastralia The radius and ulna are straight bonesshorter than the humerus (approximately two-thirdsthe length of the humerus) The shafts of these bonesare essentially of the same width although that of theulna appears to be slightly thicker However althoughjudging from the more weathered right element the

proximal end of the ulna appears to be more expandedthan that of its counterpart Unlike Compsognathusand Sinosauropteryx (OSTROM 1978 CURRIE amp CHEN

2001) the ulna of Juravenator lacks a prominentolecranon thus resembling the condition of Huaxia-gnathus and Sinocalliopteryx whose ulna seems tohave a small olecranon On the distal end the ulna ofJuravenator bears a rounded articular surface

Both manus are preserved in articulation with theleft one completely exposed in dorsolateral view ndashdigit I is mostly exposed on dorsal view and digits IIand III are more laterally exposed (Figs 17 19) Onlythe ends of the digits of the right hand are clearlyvisible ndash the rest is largely covered by gastralia ndash andthey appear to be exposed medially There is no evi-dence of any carpal bone even if the entire forelimb ispreserved in articulation Probably the carpalia havenot been ossified yet ndash another indication of the juve-nileimmature age of the specimen The hand of Jura-venator carries three digits (Fig 19) Metacarpal Iis approximately 40 the length of metacarpal IISuch a ratio agrees well with that of Sinosauropteryx

Fig 20 Photographs of the right hand of Juravenator starki (JME Sch 200) under normal (A) and ultraviolet (B) lightArrows (B) point to the keratinous sheaths of the manual claws Abbreviations mcI-III metacarpals I-III 1-I first phalanxof digit I 2-I second phalanx of digit I 1-II first phalanx of digit II 2-II second phalanx of digit II 3-II third phalanx ofdigit II 1-III first phalanx of digit III 2-III second phalanx of digit III 3-III third phalanx of digit III 4-III fourth phalanxof digit III

(CURRIE amp CHEN 2001) Huaxiagnathus (HWANG etal 2004) Sinocalliopteryx (JI et al 2007a) and thebasal tyrannosauroid Guanlong (XU et al 2006) butit differs from that of Scipionyx which has a sub -stantially lower ratio (~ 27 ) Metacarpal I of Jura -venator more resembles that of Huaxiagnathus Sci-pionyx and Guanlong than that of Sinosauropteryx inthat in the first four taxa this bone is longer than it iswide and similar in width to metacarpal II Yet theproximomedial corner of the metacarpal I of Jura -venator projects into a distinct flange similar to (butless prominent than) the one exhibited by Sinosauro -pteryx ndash a condition apparently unique to these thero-pods (CURRIE amp CHEN 2001) The trochlea of meta-carpal I is asymmetric with its lateral rim projectingdistally more than its medial rim The straight meta-carpal II is thicker than metacarpal III The latter isalso straight but has 75 the length of metacarpal IIa proportion that falls within the range (~70-80 )exhibited by other basal coelurosaurs (eg ScipionyxHuaxiagnathus Sinosauropteryx Guanlong) ndash JI et al(2007a) reported a relatively shorter metacarpal III(64 the length of the metacarpal II) for Sinocallio -pteryx Although the left metacarpal III of Juravenatorappears much thinner than the metacarpal II such adifference may be explained by the fact that the latteroverlaps the former (Fig 20) Examination of the righthand confirms that metacarpal III is thinner thanmetacarpal II but the precise relation between thesewidths is difficult to establish with confidence Ourestimation however is that metacarpal III is about onehalf the width of metacarpal II which agrees well withthe relative widths of these bones in Sinosauropteryxand Huaxiagnathus (CURRIE amp CHEN 2001 HWANG etal 2004) but differs from Scipionyx and Sinocallio -pteryx in which the metacarpal III is proportionallythicker and thinner respectively The metacarpals ofCompsognathus (both specimens) are too poorlypreserved to establish reliable comparisons but ashighlighted elsewhere (eg CURRIE amp CHEN 2001PEYER 2006 GISHLICK amp GAUTHIER 2007) theirmorphology seems to compare well with the overalldesign (eg very short metacarpal I slender meta -carpal III) of the hand of other ldquocompsognathidsrdquo

The three digits of Juravenator are robust andpowerfully clawed (Fig 20) The phalangeal formulais 2-3-4-x-x and when the digits are measured in arti-culation with their metacarpals digit I is the shortestand digit II is the longest Phalanx I-1 is the largest ofall phalanges This straight bone is as long as meta-carpal II and has a similar width The overall morpho-


Fig 21 Comparison between the right ilia of Juravenatorstarki (A) Mirischia asymmetrica (B based on SMNK2349) Sinosauropteryx prima (C based on CURRIE amp CHEN

2001) Ornitholestes hermanni (D based on AMNH 619)Gallimimus bullatus (E based on BARSBOLD amp OSMOacuteLSKA

1990) and Huaxiagnathus orientalis (F based on HWANG etal 2004) Hatched areas indicate missing portions of theskull Drawings not to scale


logy of this bone confirms previous statements (egMICHARD 1991 CURRIE amp CHEN 2001 GISHLICK ampGAUTHIER 2007) indicating that the bone described byOSTROM (1978) as the metacarpal I of Compsognathusis indeed the phalanx I-1 of this animal The relativesize of phalanx I-1 of Juravenator compares well withthat of Huaxiagnathus (HWANG et al 2004) Scipio -nyx (DAL SASSO amp SIGNORE 1998) Compsognathus(GISHLICK amp GAUTHIER 2007) and Sinocalliopteryx(JI et al 2007a) but drastically differs from that ofSinosauropteryx In this latter taxon phalanx I-1 ismuch longer and wider than metacarpal II and it isas long as the radius (CURRIE amp CHEN 2001) Thisunusually large phalanx I-1 appears to be unique toSinosauropteryx since such proportions ndash in parti -cular a length subequal to that of the radius ndashare notshared by other basal coelurosaurs (eg Huaxia -gnathus Compsognathus Sinocalliopteryx Jurave -nator Scipionyx Nqwebasaurus tyrannosauroids or -nithomimids) (Fig 18) The claw supported by pha-lanx I-1 is subequal in length to the latter ndash this claw is

the largest of the hand Like all other manual claws itscurvature is moderate and its proximal end is veryhigh In lateral view this and other manual claws taperabruptly around their midpoints The abrupt transitionbetween the proximal and distal ends of the manualclaws markedly differs from the more gradual taperingof the unguals of Huaxiagnathus SinosauropteryxSinocalliopteryx Scipionyx Nqwebasaurus Ornitho-lestes and other basal coelurosaurs (Fig 19) Thelateral surfaces of the manual claws are excavated by adeep longitudinal groove and the horny sheaths ex -tending the length of the unguals are visible under UVlight (Fig 20) Phalanx II-1 is approximately 80 thelength of phalanx II-2 The latter is slightly bowedalong the dorsoventral plane The dorsal surface of thedistal end of phalanx II-1 has a marked central depres-sion and the sides of the distal end of this phalanx andof phalanx II-2 are excavated by shallow fossae forthe attachment of collateral ligaments The length ofthe ungual phalanx of digit II is subequal to that ofphalanx II-2 and about a third longer than the claw

Fig 22 Photograph of the hindlimbs of Juravenator starki (JME Sch 200) under UV light (A) and an inset of the left anklephotographed under normal light (B) Abbreviations as astragalus fe femur fi fibula mt III metatarsal III mt Vmetatarsal V ti tibia

of digit III The latter digit has two short proximalphalanges (III-1 and III-2) that have approximatelyalmost the same length and that are about two-thirdsthe length of phalanx III-3 The claw (III-4) is thelongest phalanx of this digit (see Appendix I for speci-fic measurements of manual phalanges)

The pelvis is complete with its sides exposedalthough its bones are disarticulated (Fig 13) Theright ilium is exposed in lateral view and the left iliumis exposed in medial view however other bones covermost of the left ilium The ilium is essentially astraight laterally concave long and low bone withits dorsal and ventral margins gently tapering caudal-ly The preacetabular process is slightly shorter andmuch higher ndash more than twice in depth ndash than thepost acetabular process its cranial margin is roundlacking the hooked outline of some other theropods(eg Allosaurus Mirischia Gallimimus Ornithole-stes Sinocalliopteryx Guanlong) (Fig 21) The cau-dal end of the postacetabular process is distinctlysquare in lateral view a design similar to that of manybasal tetanurans (including other basal coelurosaursWEISHAMPEL et al 2004) This design howevershows a significant departure from the pointed endof Ornitholestes (CARPENTER et al 2005b) and moreadvanced coelurosaurs (eg dromaeosaurids therizi-nosaurids WEISHAMPEL et al 2004) The differencebetween the heights of the pre- and postacetabularprocesses is strikingly different from the more sub -equal height of these processes in HuaxiagnathusSinocalliopteryx Sinosauropteryx and GuanlongThis feature is difficult to ascertain in the poorlypreserved pelves of Scipionyx Mirischia and the twospecimens of Compsognathus

The acetabular contribution of the ilium is amplebetween one-third and one-fourth the entire length ofthe bone The dorsal margin of the acetabulum pro -jects laterally forming a distinct ledge over the head ofthe femur (Fig 13) similar supracetabular crests arefound in a variety of other theropods including basalcoelurosaurs The pubic peduncle is robust projectingcranioventrally and distally a bit beyond the distalprojection of the ischiadic peduncle The lateral sideof the pubic peduncle is craniocaudally concave andits craniodistal end approximates the mid-level of thepreacetabular process The proportion between thesize of the pubic peduncle and that of the preacetabu-lar process resembles more the condition in Huaxia-gnathus (HWANG et al 2004) Sinocalliopteryx (JI etal 2007a) Mirischia (NAISH et al 2004) and the


Fig 23 Close-ups (A C) and interpretive drawings (B D)of the right (A B) and left (C D) foot of Juravenator starki(JME Sch 200) Abbreviations mt II-V metatarsals II-VI-IV pedal digits I-IV


basal tyrannosauroid Guanlong (XU et al 2006) thanthat of Sinosauropteryx (CURRIE amp CHEN 2001) orOrnitholestes (CARPENTER et al 2005b) (Fig 21) Asmall and slightly recessed area at the cranial base ofthe pubic peduncle corresponds to a minute preaceta-bular (ldquocuppedicusrdquo) fossa (HUTCHINSON 2001) Thisfossa is also small in Mirischia (NAISH et al 2004)The ischiadic peduncle is much smaller than the pubicpeduncle Its acetabular portion develops into a flatantitrochanteric facet The ventral margin of the post -acetabular process is excavated by the brevis fossa Inside view the lateral (iliac blade) and medial (medialshelf) margins of this fossa are straight and concaverespectively but the caudal end of the medial shelftapers distally instead of having a squared off end likethe iliac blade Although the compression of the iliumprevents us from assessing the width of the brevisfossa it is clear that the base of the medial shelf pro-jects ventrally more than the iliac blade

The other elements of the pelvis are of difficultinterpretation (Fig 21) ndash this area has been affected bythe postmortem rotation of the hindquarters withrespect to the presacral portion of the skeleton (Figs3-5) The long and slender shafts of two bones lyingparallel to the ilium and pointing backwards are inter-preted as proximal halves of the pubes possibly ex -posed in caudolateral view This interpretation isprimarily based on the absence of a distinct obturatorprocess If correct this interpretation suggests a sub-stantial postmortem rotation of the bones towards theback ndash assuming that the pubes were oriented eitherrostroventrally or vertically as in other basal coeluro-saurs (eg CURRIE amp CHEN 2001 HWANG et al 2004WEISHAMPEL et al 2004 CARPENTER et al 2005b XU

et al 2006 JI et al 2007a) These two bones are stillpartially connected to one another but their distalportions are missing The proximal end of the boneinterpreted as the right pubis is expanded and bears adistinct swell An elongated bone running along thecaudal margin of the right femur and dorsal to thisswell cannot be identified with precision but it may bea portion of an ischium Portions of other bones areexposed beneath the right pubis and between thecranial margin of the right femur and a pair of verte-bral centra located in front of this femur The identifi-cation of these bones remains unclear

The hindlimbs are robust and twice as long as theforelimb (Appendix 1) (Figs 3-5) Both femora arepreserved and exposed in lateral-caudolateral view butthe pelvis covers the proximal half of the left elementThe femur is slightly bowed craniocaudally convex

cranially and concave caudally (Fig 22) A notablestep on the proximocranial edge of the right femur isinterpreted as the dorsal margin of the lesser tro -chanter This margin is approximately 5 mm distal tothe proximal end of the bone Opposite to the base ofthe lesser trochanter is a crest-like structure perhapscorresponding to the posterior trochanter amongldquocompsognathidsrdquo a knob-like posterior trochanter ispresent in at least Mirischia The distal end bears twodistinct condyles that are exposed in caudolateralview There is no evidence of an ectocondylar tuberand there does not seem to be any significant grooveseparating both condyles The caudal margin of theright femur does not show any definitive evidence of afourth trochanter but the more lateral exposure of thisbone may prevent this structure from being visible Acrest-like ridge present on the distal half (dorsal to theproximal end of the tibia) is possibly a preservationalartifact This structure is too distal to be considered afourth trochanter and its crest-like appearance seemsenhanced by the overlapping of the tibia on the caudalmargin of the distal fourth of the bone

The complete tibiae are exposed in caudal viewand articulated to the fibulae (Fig 22) The tibiaexceeds the length of the femur by 10 This ratioresembles that of Huaxiagnathus (HWANG et al 2004)Sinocalliopteryx (JI et al 2007a) Sinosauropteryx(CURRIE amp CHEN 2001) and the basal tyrannosauroidDilong (XU et al 2004) but is greater in Nqweba -saurus (tibia is 20 longer than femur DE KLERK etal 2000) and Compsognathus (tibia is 20-30 longerthan femur OSTROM 1978 MICHARD 1991) The tibiais essentially straight although the medial margin of itsproximal half is slightly concave this conditionappears to be absent in other basal coelurosaurs (egKIRKLAND et al 1998 DE KLERK et al 2000 CURRIE

amp CHEN 2001) The caudal extensions of the proximalarticulations are visible and their shape suggests thatas in other theropods the articulations were convexNo other details of the proximal articular surface canbe observed Likewise nothing can be said about theshape of the fibular crest which is in both cases over-lapped by the fibula However the presence of a longi-tudinal depression in the proximal third of the lateralmargin of the right tibia next to the fibula suggeststhat the crests were developed Both fibulae are alsocomplete and caudally exposed They are straight andrelatively strong and their shafts gradually tapertowards the distal end There is no obvious sign of afibular tubercle but the development of this structuremay be obscured by the caudal exposition of the

bones Proximally the width of the fibula is aboutthree-quarters the width of the tibia on its distal halfit is one-fourth to one-third the width of the tibiaDistally the fibula ends about 3 mm before the distalend of the tibia Very little can be said about the proxi-mal tarsals which are partially exposed in caudalview (Fig 22) These bones are most clearly visible onthe left hindlimb They do not seem to be fused to eachother and the astragalus is clearly not coossified withthe tibia What is preserved of the astragalus suggestsa rather shallow bone

Both pedes are complete in articulation andlaterocaudally exposed (Figs 22-23) The foot is pro-portionally shorter than in other ldquocompsognathidsrdquoFor example the length of the metatarsal III plus itsdigit (including claw) is approximately 10 longerthan the tibia in Juravenator 12 in Compsognathus

16 in Sinosauropteryx (NIGP 127587) and 27 inHuaxiagnathus This ratio is mostly influenced by thesize of the metatarsal III which while in Juravenatoris approximately half the length of the tibia it is morethan 60 in all these other taxa All five metatarsalsare preserved and neither is coossified to each other(Fig 23) Metatarsal V is short distally pointed andsomewhat bent medially This metatarsal is approxi-mately one-fifth of metatarsal III thus being relativelyshorter than Huaxiagnathus (one-fourth HWANG et al2004) and Compsognathus (one-fourth to one-thirdOSTROM 1978 MICHARD 1991) The shorter nature ofthis metatarsal also gives it a more stout appearancewhen compared to the more gracile metatarsals of thelatter two taxa Metatarsal IV is completely exposedon both feet of Juravenator This metatarsal is robuststraight shorter than metatarsal III but longer than


Fig 24 Soft tissue structures of Juravenator starki (JME Sch 200) under normal (A C) and ultraviolet (B) light Insethighlighting details of the soft tissue (C) Arrows point at vertical stripes in between chevrons


metatarsal II The lateral rim of its distal trochlea pro-jects laterocaudally into a distinct flange A weakcollateral fossa excavates the lateral surface of thetrochlea of the left metatarsal IV but this fossa is notevident on the right element Metatarsal IV largelyoverlaps metatarsal III However it can be seen thatthe latter is the longest reaching the proximal end ofthe metatarsus (ie lacking the arctometatarsaliancondition of many other coelurosaurs) The distaltrochlea of this metatarsal lacks the lateral flange ofmetatarsal IV but its lateral collateral fossa is deeperMetatarsal II is also partially overlapped by the othermetatarsals Unlike Huaxiagnathus (HWANG et al2004) Nqwebasaurus (DE KLERK et al 2000) Sino-sauropteryx (CURRIE amp CHEN 2001) Compsognathus(OSTROM 1978 MNHN CNJ 79) and Ornitholestes(CARPENTER et al 2005b) the metatarsal II of Jurave-nator is distinctly shorter than the metatarsal IV (Fig23) Distally the trochlea of this metatarsal forms aginglymus in which the lateral rim appears to be moredistally projected than the medial rim The lateralsurface of this trochlea is excavated by a well-deve -loped collateral fossa The short metatarsal I articu -lates at the mid-shaft of metatarsal II as in Compso-gnathus (OSTROM 1978) and apparently Huaxiagna-thus (HWANG et al 2004) this metatarsal articulatesmore distally in Nqwebasaurus (DE KLERK et al2000)

All pedal phalanges are preserved in articulation(Fig 23) Digit III is the longest followed by digitsIV II and I The phalangeal formula is 2-3-4-5-x asis common for most dinosaurs All non-ungual pha-langes exhibit deep collateral fossae on the sides oftheir trochleae and the unguals bear a single lateralgroove on each side The two phalanges of digit I arelonger than their metatarsal this digit ends in a verysmall claw On digit II the proximal phalanx isslightly longer than the intermediate phalanx Theclaw of digit II is much greater than the others and ithas a prominent flexor tubercle The proximal phalanxof digit III is distinctly longer than the other non-ungual phalanges of this digit Its claw also bears aprominent flexor tubercle Like in digit III the proxi-mal phalanx of digit IV is longer than the other non-ungual phalanges However the claw of this digit doesnot appear to carry a strong flexor tubercle

53 Soft tissue

Soft tissue is preserved surrounding several parts ofthe skeleton and is particularly evident under ultra -

violet illumination (GOumlHLICH et al 2006) (Fig 24)Under ultraviolet light soft tissue preservation isvisible dorsal to the rostrum along both tibiae andbetween the 8th and the 22nd caudal vertebrae where itdefines the outline of the tail Furthermore the ultra-violet illumination perfectly visualizes the long hornysheets especially on the manual but also on the pedalclaws In the caudal region the preserved soft tissue isalso apparent under normal light (Fig 24) The caudalintegument is formed of uniformly sized smoothtubercles (about 15 tubercles per 25 mm2) similar inappearance to the small conical and non-imbricatedtubercles of many other non-avian dinosaurs (BROWN

1916 CZERKAS 1997 ANDERSON et al 1998 CORIA

amp CHIAPPE 2007)GOumlHLICH amp CHIAPPE (2006) reported the absence

of both feathers and skin follicles in the caudal portionof the skeleton of Juravenator thus suggesting thatunlike some other basal coelurosaurs (CHEN et al1998 JI et al 1998 XU et al 1999a 1999b 20042009 NORELL amp XU 2005 ZHANG et al 2006) thecentral portion of the tail of Juravenator was devoidof plumage However new technically improveddetailed ultraviolet photography by H TISCHLINGER

(Stammham Germany) has revealed evidence of tinyfilament-like structures that may be similar to thoseinterpreted as the plumage of Sinosauropteryx (CHEN

et al 1998) and other non-avian theropods (XU etal 1999a 1999b 2004 2009 NORELL amp XU 2005ZHANG et al 2006) These structures apparently re -presenting the tips of very slender and parallel-arranged filaments that are angled slightly caudallyfrom the longitudinal mid axis of the vertebralcolumn are visible along the edge of the soft tissuepreserved around caudal vertebrae 18-20 (Fig 25) Ifthese structures were found to be traces of plumagethe holotype of Juravenator starki would stand outamong other specimens of feathered non-aviantheropods in preserving both plumage and extensiveportions of a scaly integument covering main portionsof the body (eg tail) (XU 2006) The coexistence ofscaly skin and bristle-like integumentary structureshas been reported for other non-avian dinosaurs ndashalbeit not theropods (eg Psittacosaurus MAYR et al2002) ndash but it remains highly unusual Nonethelesssuch coexistence is not at odds with a wealth oflaboratory experiments that have demonstrated thatsmall groups of stem cells in scales can be stimulatedby molecular perturbation to form active featherfollicles that develop feathers (DHOUAILLY et al1980WIDELITZ et al 2000 2003 CHANG et al 2009) In

fact such ldquofeathery scalesrdquo are regularly envisioned asan intermediate stage of models of feather evolution(CHANG et al 2009)

The apparent presence of filament-like integu -mentary structures in Juravenator is consistent withsimilar structures often regarded as primitive feathersin the basal coelurosaurs Sinosauropteryx (CHEN et al1998 JI et al 2007b) and Dilong (XU et al 2004) andthe combination of these structure with large portionsof scaly skin along its body suggest that these structu-res may not have completely surrounded the bodies ofthese theropods

Besides integumentary structures internal ele-ments of soft tissue are preserved along the proximalportion of the tail Under ultra-violet illumination adistinct longitudinal band of about 2 mm in width isvisible which runs immediately ventral to the tips ofthe chevrons of the mid-proximal caudals (betweenthe14th and 23rd caudal) (Fig 24) In between thechevrons of the 17th and 21st caudals vertical illumi-nating stripes arranged in regular intervals of about3 mm are observable These vertical structures mightreflect external or internal segmentation of the tail(Fig 24B) Under normal light a series of fibers alsoventral to the chevrons of the 10th to 14th caudals andparallel to the axis of the tail is also visible These

structures may either represent tendinal elements ofthe m ilioischiocaudalis or remnants of a reinforcedseam of the vertical septum ventral to the chevrons acondition present in the terminal portion of the croco-dilian tail (FREY 1988) DAL SASSO amp SIGNORE (1998)similarly interpreted a series of comparable softstructures associated with the skeleton of Scipionyxalthough such structures could also correspond tobundles of subcutaneous collagen fibers (LINGHAM-SOLIAR 2003)

6 Conclusions

The exquisitely preserved holotype of Juravenatorstarki is one of the most complete non-avian theropodskeletons described to date for Europe and a signifi-cant addition to the scant worldwide record of small-bodied Late Jurassic theropods This specimen pro -vides evidence of morphologies ndash from details of theskull to the epidermis ndash that are poorly known in othertheropods interpreted as basal coelurosaurians andthus it contributes significantly to our understandingof the anatomical disparity and evolutionary historyof this clade Juravenator exhibits a number of simi -larities (eg hair-like cervical ribs extremely longtail) to other coelurosaurians often regarded as havingevolved either at the base or near the divergence ofthis clade The discovery of Juravenator augments thediversity of predatory dinosaurs known for the LateJurassic basins of southern Germany (Fig 26) and itdocuments that during this time the nearby recentlyemerged lands were inhabited by diverse small-bodiednon-avian coelurosaurians of presumably similar life-styles

Acknowledgements

We are especially grateful to the finders of the specimenHANS and KLAUS-DIETER WEISS (Kelkheim-Fischbach) toPINO VOumlLKL (Jura-Museum Eichstaumltt) for the spectacularpreparation of the specimen and to HELMUT TISCHLINGER

(Stammham) and GEORG JANSSEN (Bayerische Staatssamm-lung fuumlr Palaumlontologie und Geologie Munich) for the in -valuable UV and normal light photographs respectively Weare also very grateful to MARTINA KOumlLBL-EBERT (Jura-Museum Eichstatt) and FRANZ STARK (Schamhaupten) foraccess to the specimen We thank STEPHANIE ABRAMOWICZWILLIAM EVANS and EMMA FREEMAN (Natural HistoryMuseum of Los Angeles County) for preparing the illus -trations and for editorial assistance respectively JAMES

CLARK (George Washington University) CRISTIANO DAL

SASSO (Museo Civico di Storia Naturale di Milano) EBER-HARD FREY (Staadliches Museum fuumlr Naturkunde Karls -ruhe) DANIEL GOUJET (Museacuteum National drsquoHistoire Natur -


Fig 25 Extremely thin filaments at the edge of the pre -served soft tissue (inside box) are visible at high magni -fication and under ultraviolet light in some areas of themid-caudal region


elle Paris) SUNNY HWANG (American Museum of NaturalHistory New York) MARTINA KOumlLBL-EBERT REINHOLD

LEINFELDER (Museum fuumlr Naturkunde Berlin) MARKUS

MOSER (Muumlnchen) MARK NORELL (American Museum ofNatural History) OLIVER RAUHUT (Bayerische Staatsamm-lung fuumlr Palaumlontologie und Geologie) MARTIN ROumlPER

(Museum Solnhofen) HELMUT TISCHLINGER GUumlNTER

VIOHL (Jura Museum Eichstaumltt) DAVID VARRICCHIO (Mon-tana State University) PINO VOumlLKL PETER WELLNHOFER

and WINFRIED WERNER (Bayerische Staatssammlung fuumlrPalaumlontologie und Geologie) and XU XING and ZHOU

ZHONGHE (Institute of Vertebrate Paleontology and Paleo -anthropology Beijing) provided access to specimens orunpublished photographs logistics andor valuable dis -cussions and STEPHEN BRUSATTE (American Museum ofNatural History) and ROGER BENSON (Cambridge Uni-versity) provided invaluable assistance during their reviewsof the manuscript Finally LUIS CHIAPPE wishes to thankvery especially MARTINA KOumlLBL-EBERT REINHOLD LEIN-FELDER OLIVER RAUHUT PETER WELLNHOFER and WIN-FRIED WERNER for the wonderful hospitality offered duringhis visits to the Jura-Museum Eichstaumltt and the BayerischeStaatssammlung fuumlr Palaumlontologie und Geologie This re -search was made possible by the Alexander von HumboldtFoundation (Friedrich Wilhelm Bessel Award) and sup -ported also by the Jurassic Foundation Synthesys ProgramAntorchas Foundation Bayerische Staatssammlung fuumlrPalaumlontologie und Geologie Department of Geo- and En -vironmental Sciences (Munich University) Jura-MuseumEichstaumltt and Natural History Museum of Los Angeles

County The Alexander von Humboldt Foundation is alsoacknowledged for funding the color printing of thefigures

References

ANDERSON B G LUCAS S G BARRICK R E HECKERTA B amp BASABILVAZO G T (1998) Dinosaur skinimpressions and associated skeletal remains from theupper Campanian of Southwestern New Mexico Newdata on the integument morphology of hadrosaurs ndashJournal of Vertebrate Paleontology 18 739-745

BARTHEL K W SWINBURNE N H M amp CONWAY MORRISS (1990) Solnhofen ndash A study in Mesozoic palaeonto-logy ndash 236 pp Cambridge (Cambridge UniversityPress)

BARSBOLD R amp OLMOSKA H (1990) Ornithomimosauriandash In WEISHAMPEL D B DODSON P amp OSMOacuteLSKA H(Eds) The Dinosauria 225-244 Berkeley (Universityof California Press)

BAUMEL J J amp WITMER L M (1993) Osteologia ndash InBAUMEL J J KING A S BREAZILE J E EVANS HEamp VANDEN BERGE J C (Eds) Handbook of aviananatomy Nomina Anatomica Avium ndash Publications ofthe Nuttall Ornithological Club 23 45-132

BIDAR A DEMAY L amp THOMEL G (1972) Compsogna-thus corallestris nouvelle espegravece de dinosaurien theacutero-pode du Portlandien de Canjuers (Sud-Est de la France)ndash Annales du Museacuteum drsquoHistoire Naturelle Nice I (1)3-34

Fig 26 Reconstructed skeletons of Juravenator starki compared to Compsognathus longipes (based on MNHM CNJ 79PEYER 2006 GISHLICK amp GAUTHIER 2007) These animals the only non-avian theropod skeletons known from the LateJurassic Solnhofen Archipelago are among the most informative non-avian theropods from the European fossil recordDrawings not to scale

BROCHU C A (1996) Closure of neurocentral suturesduring crocodilian ontogeny implications for maturityassessment in fossil archosaurs ndash Journal of VertebratePaleontology 16 49-62

ndash (2003) Osteology of Tyrannosaurus rex Insights from anearly complete skeleton and high-resolution computedtomographic analysis of the skull ndash Journal of Verte -brate Paleontology 22 (Suppl 4) 1-138

BROWN B (1916) Corythosaurus casuarius Skeletonmus culature and epidermis ndash American Museum ofNatural History Bulletin 35 709-716

BUTLER R amp UPCHURCH P (2007) Highly incompletetaxa and the phylogenetic relationships of the theropoddinosaur Juravenator starki ndash Journal of Vertebratepaleontology 27 (1) 253-256

CAMP C L (1936) A new type of small theropod dinosaurfrom the Navajo Sandstone of Arizona ndash Bulletin of theUniversity of California Department of GeologicalSciences 24 39-65

CARPENTER K MILES C amp CLOWARD K (2005a) Newsmall theropod from the Upper Jurassic Morrison For-mation of Wyoming ndash In CARPENTER K (Ed) TheCarnivorous Dinosaurs 23-48 Bloomington (IndianaUniversity Press)

CARPENTER K MILES C OSTROM J H amp CLOWARD K(2005b) Redescription of the small maniraptoran thero-pods Ornitholestes and Coelurus from the Upper Juras-sic Morrison Formation of Wyoming ndash In CARPENTERK (Ed) The Carnivorous 49-71 Bloomington (IndianaUniversity Press)

CARRANO M T HUTCHINSON J R amp SAMPSON S D(2005) New information on Segisaurus halli a smalltheropod dinosaur from the Early Jurassic of Arizona ndashJournal of Vertebrate Paleontology 25 (4) 835-849

CHANG C WU P BAKER R E MAINI P K ALIBARDI Lamp CHUONG C-M (2009) Reptile scale paradigm Evo-Devo pattern formation and regeneration ndash Inter -national Journal of Developmental 53 813-826

CHEN P-J DONG Z-M amp ZHEN S-N (1998) An ex -ceptionally well-preserved theropod dinosaur from theYixian Formation of China ndash Nature 391 147-152

CHIAPPE L M (2007) Glorified Dinosaurs ndash 263 ppNew York (Wiley amp Sons)

CHIAPPE L M CORIA R A DINGUS L JACKSON FCHINSAMY A amp FOX M (1998) Sauropod dinosaurembryos from the Late Cretaceous of Patagonia ndashNature 396 258-261

CHOINIERE J N XU X CLARK J M FORSTER C AGUO Y amp HAN F (2010) A basal alvarezsauroid thero-pod from the Early Late Jurassic of Xinjiang China ndashNature 327 571-574

CLARK J M NORELL M A amp CHIAPPE L M (1999) Anoviraptorid skeleton from the Late Cretaceous of UkhaaTolgod Mongolia preserved in an avianlike broodingposition over an oviraptorid nest ndash American MuseumNovitates 3265 1-36

CLAESSENS L (2004) Dinosaur gastralia Origin morpho-logy and function ndash Journal of Vertebrate Paleontology24 (1) 89-106

COLBERT E H (1989) The Triassic dinosaur Coelophysisndash Bulletin of the Museum of Northern Arizona 571-160

CODORNIUacute L amp CHIAPPE L M (2004) Early juvenilepterosaurs (Pterodactyloidea Pterodaustro guinazui)from the Lower Cretaceous of central Argentina ndash Cana-dian Journal of Earth Sciences 41 9-18

CORIA R A amp CHIAPPE L M (2007) Embryonic skinfrom Late Cretaceous sauropods (Dinosauria) of AucaMahuevo Patagonia Argentina ndash Journal of Paleonto -logy 81 (6) 1528-1532

CORIA R A amp CURRIE P J (2006) A new carcharodonto-saurid (Dinosauria Theropoda) from the Upper Creta-tious of Argentina ndash Geodiversitas 28 (1) 71-118

CURRIE P J (2003) Cranial anatomy of tyrannosauridsfrom the Late Cretaceous of Alberta Canada ndash ActaPalaeontologica Polonica 48 191-226

CURRIE P J amp CHEN P-J (2001) Anatomy of Sinosauro -pteryx prima from Liaoning northeastern China ndashCanadian Journal of Earth Sciences 38 1705-1727

CURRIE P J amp ZHAO X (1993) A new carnosaur (Dino-sauria Theropoda) from the Jurassic of XinjiangPeoplersquos Republic of China ndash Canadian Journal of EarthSciences 30 2037-2081

CZERKAS S (1997) Skin ndash In CURRIE P J amp PADIAN K(Eds) Encyclopedia of Dinosaurs 669-675 San Diego(Academic Press)

DAL SASSO C amp SIGNORE M (1998) Exceptional soft-tissue preservation in a theropod dinosaur from Italy ndashNature 392 383-387

DE KLERK W J FORSTER C A SAMPSON S D CHIN-SAMY A amp ROSS C F (2000) A new coelurosauriandinosaur from the Early Cretaceous of South Africa ndashJournal of Vertebrate Paleontology 20 (2) 324-332

DESMOND A (1982) Archetypes and Ancestors ndash 287 ppChicago (University of Chicago Press)

DHOUAILLY D HARDY M H amp SENGEL P (1980) For -mation of feathers on chick foot scales A stage-depen-dent morphogenetic response to retinoic acid ndash Journalof Embryology and Experimental Morphology 5863-78

FREY E (1988) Anatomie des Koumlrperstammes von Alli -gator mississippiensis DAUDIN ndash Stuttgarter Beitraumlgezur Naturkunde (A) 424 1-106

GAUTHIER J (1986) Saurischian monophyly and the originof birds ndash In PADIAN K (Ed) The origin of birds andthe evolution of flight ndash Memoires of the CaliforniaAcademy of Sciences 8 1-55

GISHLICK A D amp GAUTHIER J (2007) On the manualmorphology of Compsognathus longipes and its bearingon the diagnosis of Compsognathidae ndash ZoologicalJournal of the Linnean Society 149 569-581

GOumlHLICH U B amp CHIAPPE L M (2006) A new carni-vorous dinosaur from the Late Jurassic Solnhofen archi-pelago ndash Nature 440 329-332

GOumlHLICH U B TISCHLINGER H amp CHIAPPE L M (2006)Juravenator starki (Reptilia Theropoda) ein neuerRaubdinosaurier aus dem Oberjura der suumldlichen Fran-kenalb (Suumlddeutschland) Skelettanatomie und Weich-teilbefunde) ndash Archaeopteryx 24 1-26



HOLTZ T R MOLNAR R E amp CURRIE P J (2004) BasalTetanurae ndash In WEISHAMPEL D B DODSON P ampOSMOacutePLSKA H (Eds) The Dinosauria 71-110 Ber -keley Los Angeles amp Oxford (California UniversityPress)

HORNER J R (1997) Behavior ndash In CURRIE P amp PADIANK (Eds) Encyclopedia of Dinosaurs 45-50 San Diego(Academic Press)

HUTCHINSON J R (2001) The evolution of pelvic osteo -logy and soft tissues on the line to extant birds (Ne -ornithes) ndash Zoological Journal of the Linnean Society131 123-168

HUENE F V (1914) Das natuumlrliche System der Saurischiandash Zentralblatt fuumlr Mineralogie Geologie und Palaumlonto -logie (B) 1914 154-158

HWANG S H NORELL M A QIANG J amp KEQIN G(2002) New specimens of Microraptor zhaoianus(Theropoda Dromaeosauridae) from northeastern Chinandash American Museum Novitates 3381 1-44

HWANG S H NORELL M A QIANG J amp KEQIN G A(2004) Large compsognathid from the Early CretaceousYixian Formation of China ndash Journal of SystematicPaleontology 2 (1) 13-30

JI Q amp JI S 1996 On discovery of the earliest bird fossilin China and the origin of birds ndash Chinese Geology 10(233) 30-33

JI S JI Q LUuml J amp YUAN C (2007a) A New GiantCompsognathid Dinosaur with Long Filamentous Inte-guments from Lower Cretaceous of Northeastern Chinandash Acta Geologica Sinica 81 (1) 8-15

JI S GAO C LIU J MENG Q amp JI Q (2007b) Newmaterial of Sinosauropteryx (Theropoda Compsogna-thidae) from Western Liaoning China ndash Acta GeologicaSinica 81 (2) 177-182

JI Q CURRIE P J NORELL M A amp JI S-A (1998) Twofeathered theropods from the Upper JurassicLowerCretaceous strata of northeastern China ndash Nature 393753-761

KIRKLAND J I BRITT B B WHITTLE C H MADSEN SK amp BURGE D L (1998) A small coelurosauriantheropod from the Yellow Cat Memeber of the CedarMountain Formation (Lower Cretaceous Barremian) ofEartern Utah ndash In LUCAS S G KIRKLAND J L ampESTEP J W (Eds) Lower and Middle Cretaceous Ter -restrial ecosystems ndash Science Bulletin of the NewMexico Museum of Natural History 14 239-248

LINGHAM-SOLIAR T (2003) Evolution of birds ichthyo-saur integumental fibers conform to dromaeosaur proto-feathers ndash Naturwissenschaften 90 428-432

MADSEN J H (1976) Allosaurus fragilis A revised osteo-logy ndash Bulletin of the Utah Geological and MineralSurvey 109 1-163

MAKOVICKY P J amp CURRIE P J (1998) The presence of afurcula in tyrannosaurid theropods and its phylogeneticand functional implications ndash Journal of VertebratePaleontology 18 143-149

MARSH O C (1880) Odontornithes a monograph on theextinct toothed birds of North America ndash United StatesGeological Exploration of the 40th Parallel ndash 201 ppWashington DC (US Government Printing Office)

ndash (1881) Principal characters of American Jurassic dino-saurs Part V ndash American Journal of Sciences Series 321 417-423

MARSHALL FAUX C amp PADIAN K (2007) The opisthotonicposture of vertebrate skeletons postmortem contractionor death throes ndash Paleobiology 33 (2) 201-226

MARTILL D M FREY E SUES H-D amp CRUICKSHANK AR I (2000) Skeletal remains of a small theropod dino-saur with associated soft structures from the LowerCretaceous Santana Formation of northeast Brazil ndashCanadian Journal of Earth Sciences 37 891-900

MAYR G PETERS D S PLODOWSKI G amp VOGEL O(2002) Bristle-like integumentary structures at the tailof the horned dinosaur Psittacosaurus ndash Naturwissen-schaften 89 361-365

MICHARD J-G (1991) Description du Compsognathus(Saurischia Theropoda) de Canjuers (JurassiqueSupeacuterieur du Sud-Est de la France) Position phylogeacutene-tique relation avec Archaeopteryx et implications surlorigine Theacuteropodienne des oiseaux ndash UnpublishedPhD thesisndash 328ppParis (Museacuteum National drsquoHistoireNaturelle)

NAISH D MARTILL D M amp FREY E (2004) Ecologysystematics and biogeographical relationships of dino-saurs including a new theropod from the Santana For-mation (Albian Early Cretaceous) of Brazil ndash HistoricalBiology 2004 1-14

NESBITT S J TURNER A H SPAULDING M CONRAD JL amp NORELL M A (2009) The theropod furcula ndashJournal of Morphology 270 856-879

NORELL M A MAKOVICKY P J amp CLARK J M (1997) AVelociraptor wishbone ndash Nature 389 447

ndash (2004) Dromaeosauridae ndash In WEISHAMPEL D BDODSON P amp OSMOacuteLSKA H (Eds) The Dinosauria p196-209 Berkeley Los Angeles amp Oxford (CaliforniaUniversity Press)

NORELL M A amp XU X (2005) Feathered Dinosaurs ndashAnnual Revue of Earth and Planetary Sciences 33 277-299

OSBORN H F (1916) Skeletal adaptations of Ornithole-stes Struthiomimus Tyrannosaurus Bulletin of theAmerican Museum of Natural History 19 459-464

OSTROM J H (1978) The osteology of Compsognathuslongipes WAGNER ndash Zitteliana 4 73-118

OWEN R (1842) Report on British fossil reptiles part II ndashReport of the British Association for the Advancementof Science 1841 60-204

PEYER K (2006) A reconsideration of Compsognathusfrom the Upper Tithonian of Canjuers southeasternFrance ndash Journal of Vertebrate Paleontology 26 (4)879-896

RAUHUT O W M (2003) The interrelationships and evo -lution of basal theropod dinosaurs ndash Special Papers inPalaeontology 69 1-213

RENESTO S amp VIOHL G (1997) A sphenodontid (ReptiliaDiapsida) from the Late Kimmeridgian of Schamhaup-ten (Southern Franconian Alb Bavaria Germany) ndashArchaeopteryx 15 27-46

SANDER M P (1997) Teeth and claws ndash In CURRIE P J ampPADIAN K (Eds) Encyclopedia of Dinosaurs 717-725San Diego (Academic Press)

SANZ J L CHIAPPE L M PEacuteREZ-MORENO B PMORATALLA J J HERNAacuteNDEZ-CARRASQUILLA FBUSCALIONI A D ORTEGA F POYATO-ARIZA F JRASSKIN-GUTMAN D amp MARIacuteNEZ-DELCLOacuteS X (1997)A nestling bird from the Lower Cretaceous of SpainImplications for avian skull and neck evolution ndashScience 276 1543-1546

SANZ J L CHIAPPE L M FERNAacuteNDEZ-JALVO Y ORTEGAF SAacuteNCHEZ-CHILLOacuteN B POYATO-ARIZA F J amp PEacuteREZ-MORENO B P (2001) An early Cretaceous pellet ndashNature 409 998-999

SCHWEIGERT G (2007) Ammonite biostratigraphy as a toolfor dating Upper Jurassic Lithographic limestones fromSouthern Germany ndash first results and open questions ndashNeues Jahrbuch fuumlr Geologie und PalaumlontologieAbhandlungen 245 (1) 117-125

SENTER P (2007) A new look at the phylogeny of Coeluro-sauria (Dinosauria Theropoda) ndash Journal of SystematicPalaeontology 5 (4) 429-463

SERENO P TAN L BRUSATTE S L KRIEGSTEIN H JZHAO X amp CLOWARD K (2009) Tyrannosaurid skeletaldesign first evolved at small body size ndash Science 326418-422

THERRIEN F amp HENDERSON D M (2007) My theropod isbigger than yours hellip or not Estimating body size fromskull length in theropods ndash Journal of VertebratePaleontology 27 (1) 108-115

TISCHLINGER H GOumlHLICH U B amp CHIAPPE L M (2006)Borsti der Dinosaurier aus dem Schambachtal Erfolgs-story mit Hindernissen ndash Fossilien 2006 (5) 277-287

VIOHL G (1999) Fund eines neuen kleinen Theropoden ndashArchaeopteryx 17 15-19

VIOHL G amp ZAPP M (2006) Die Fossil-LagerstaumltteSchamhaupten (oberstes Kimmeridgium Suumldliche Fran-kenalb Bayern) ndash Archaeopteryx 24 27-78

ndash (2007) Schamhaupten an outstanding Fossil-Lager -staumltte in a silicified Plattenkalk around the Kimmerid -gian-Tithonian boundary (Southern Franconian AlbBavaria) ndash Neues Jahrbuch fuumlr Geologie und Palaumlonto-logie Abhandlungen 245 127-142

WAGNER A (1861) Neue Beitraumlge zur Kenntnis der urwelt-lichen Fauna des lithographischen Schiefers V Compso-gnathus longipes WAGNER ndash Abhandlungen der Bayeri-schen Akademie der Wissenschaften 9 30-38

WEISHAMPEL D B DODSON P amp OSMOacuteLSKA H (2004)The Dinosauria (2nd ed) ndash 861 pp Berkeley (Uni -versity of California Press)

WELLNHOFER P (2008) Archaeopteryx ndash Der Urvogel vonSolnhofen ndash 256 pp Muumlnchen (Pfeil)

WIDELITZ R B JIANG T-X LU J-F amp CHUONG C-M(2000) Beta catenin in epithelial morphogenesisConversion of part of avian foot scales into feather budswith a mutated beta catenin ndash Developmental Biology219 98-114

WIDELITZ R B JIANG T-X YU M WU P YUE Z ampCHUONG C-M (2003) Molecular biology of feathermorphogenesis A testable model of Evo-Devo researchndash Journal of Experimental Zoology 298B 109-222

XU X (2006) Scales feathers and dinosaurs ndash Nature440 287-288

XU X CLARK J M FORSTER C A NORELL M AERICKSON G M EBERTH D A JIA C amp ZHAO Q(2006) A tyrannosauroid dinosaur from the Late Juras-sic of China ndash Nature 439 715-718

XU X CLARK J M MO J CHONIERE J FORSTER C AERICKSON G M HONE D E SULLIVAN C EBERTHD NESBITT S ZHAO Q HERNANDEZ R JIA C HANF amp GUO Y (2009) A Jurassic ceratosaur from Chinahelps clarify avian digital homologies ndash Nature 459940-944

XU X NORELL M A KUANG X WANG X ZHAO QAND C JIA (2004) Basal tyrannosauroids from Chinaand evidence for protofeathers in tyrannosauroids ndashNature 431 680-684

XU X WANG X-L amp WU X-C (1999a) A dromaeo -saurid dinosaur with a filamentous integument from theYixian Formation of China ndash Nature 401 262-266

XU X TANG Z-L amp WANG X-L (1999b) A therizino-sauroid dinosaur with integumentary structures fromChina ndash Nature 399 350-354

XU X ZHENG Z amp YOU H (2009) A new feather type ina non-avian theropod and the early evolution of feathersndash Proceedings of the National Academy of Sciences 106(3) 832-834

ZEISS A (2001) Wenig bekannte Ammoniten aus demGrenzbereich OberkimmeridgiumUntertithonium derSuumldlichen Frankenalb ndash Archaeopteryx 19 57-70

ZHANG F ZHOU Z amp DYKE G (2006) Feathers and lsquofea-ther-likersquo integumentary structures in Liaoning birds anddinosaurs ndash Geological Journal 41 1-6

Manuscript received July 23rd 2009Revised version accepted by the Stuttgart editor July 8th2010

Addresses of the authors

LUIS M CHIAPPE The Dinosaur Institute Natural HistoryMuseum of Los Angeles County 900 Exposition BoulevardLos Angeles CA 90007 USAe-mail chiappenhmorg

URSULA B GOumlHLICH Naturhistorisches Museum WienGeologisch-palaumlontologischeAbteilung Burgring 7 A-1010Vienna Austriae-mail ursulagoehlichnhm-wienacat



Appendix Measurements of the skeletonof Juravenator starki (in mm)

Left Right

Skull and MandibleSkull length ndash 82Orbit length ndash 184Orbit height ndash 14Antorbital fossa length ndash 24Antorbital fossa height ndash 9Antorbital fenestra length ndash ~195Supratemporal fenestra length 75 ndashMandible length ~77 ndash

Shoulder and ForelimbCoracoid length ndash ndashScapula length 42 ndashHumerus length 27 275Ulna length 205 205Radius length ndash ndashMetacarpal I length 45 ndashMetacarpal II length 115 ndashMetacarpal III length 9 ndashPhalanx I-1 length 105 ndashPhalanx II-1 length 8 ndashPhalanx II-2 length 10 10Phalanx III-1 length 4 ndashPhalanx III-2 length 45 45Phalanx III-3 length 55 ndashPhalanx I-2 (claw) length ~12 ndashPhalanx II-3 (claw) length 9 10Phalanx III-4 (claw) length 55 7

Pelvis and HindlimbIlium length ndash 40Femur length ndash 52Tibia length 581 581Fibula length 553 56Metatarsal I length 46 45Metatarsal II length 265 ndashMetatarsal III length 34 32Metatarsal IV length 296 298Metatarsal V length 8 68Phalanx I-1 length 58 6Phalanx II-1 length 104 114Phalanx II-2 length 9 8Phalanx III-1 length 119 115Phalanx III-2 length 81 8Phalanx III-3 length 74 77Phalanx IV-1 length 74 7Phalanx IV-2 length 55 65Phalanx IV-3 length 53 45Phalanx IV-4 length 42 4Phalanx I-2 (claw) 60 35Phalanx II-3 (claw) 107 115Phalanx III-4 (claw) 74 66Phalanx IV-5 (claw) 72 58

Lengthmid-height

caudal centraCaudal 1 ndash ndashCaudal 2 62 39Caudal 3 65 38Caudal 4 65 33Caudal 5 64 32Caudal 6 69 35Caudal 7 65 35Caudal 8 64 32Caudal 9 64 31Caudal 10 61 28Caudal 11 68 ndashCaudal 12 ndash ndashCaudal 13 ndash ndashCaudal 14 65 ndashCaudal 15 ndash ndashCaudal 16 ndash ndashCaudal 17 68 25Caudal 18 72 24Caudal 19 75 26Caudal 20 75 27Caudal 21 76 24Caudal 22 76 23Caudal 23 79 27Caudal 24 77 ndashCaudal 25 80 22Caudal 26 80 26Caudal 27 84 ndashCaudal 28 84 26Caudal 29 83 23Caudal 30 83 21Caudal 31 82 23Caudal 32 83 23Caudal 33 83 ndashCaudal 34 83 23Caudal 35 83 27Caudal 36 80 25Caudal 37 79 19Caudal 38 77 17Caudal 39 76 20Caudal 40 76 19Caudal 41 75 17Caudal 42 75 22Caudal 43 72 22Caudal 44 ndash ndash

claws were measured from the dorsoproximal corner tothe osseous tip


implication of the putative monophyletic nature ofa group that includes Compsognathus longipes(WAGNER 1861 OSTROM 1978 PEYER 2006) Huaxia-gnathus orientalis (HWANG et al 2004) Sinosauro -pteryx prima (JI amp JI 1996 CHEN et al 1998 CURRIE

amp CHEN 2001) Scipionyx samniticus (DAL SASSO ampSIGNORE 1998) and Juravenator starki (GOumlHLICH amp

CHIAPPE 2006) among other non-avian theropodsusually regarded at or near the base of the coeluro -saurian clade The morphological diversity of thesetheropods has remained as a poorly understood chap-ter in the evolutionary history of Late Jurassic-EarlyCretaceous theropod dinosaurs (CURRIE amp CHEN

2001 HOLTZ et al 2004 PEYER 2006) However

Fig 1 Geographic location of Juravenator starki (JME Sch 200) Compsognathus longipes and selected specimensof Archaeopteryx lithographica (reconstructions not to scale) Barred and mottled patterns indicate the approximatedepositional areas during the Late Jurassic












































51 Skull








































































52 Postcranium








































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height














































51 Skull








































































52 Postcranium








































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height









































51 Skull








































































52 Postcranium








































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height





























51 Skull








































































52 Postcranium








































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height



























51 Skull








































































52 Postcranium








































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height






















51 Skull








































































52 Postcranium








































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height






51 Skull








































































52 Postcranium








































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height




51 Skull








































































52 Postcranium








































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height



































































52 Postcranium








































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height



























































52 Postcranium








































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height

























































52 Postcranium








































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height

















































52 Postcranium








































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height













































52 Postcranium








































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height





































52 Postcranium








































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height


























52 Postcranium








































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height


















52 Postcranium








































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height





52 Postcranium








































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height








































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height


































































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height


























































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height




















































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height












































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height




































































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height





























































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height























































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height


















































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height










































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height


































53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height




























53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height
























53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height



















53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height











53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height






53 Soft tissue












6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height







6 Conclusions


Acknowledgements
















References








































































































Left Right




Lengthmid-height












References








































































































Left Right




Lengthmid-height




































































































Left Right




Lengthmid-height



































































Left Right




Lengthmid-height


































Left Right




Lengthmid-height





Left Right




Lengthmid-height



Date post:	01-Jan-2017
Category:	Documents
Upload:	vanhuong
View:	215 times
Download:	0 times

Anatomy of Juravenator starki (Theropoda: Coelurosauria) from the ...

Documents