the skeletal morphology and phylogenetic position of...

19
The skeletal morphology and phylogenetic position of Adocus amtgai, an adocid turtle from the Late Cretaceous of Mongolia E.V. Syromyatnikova a, * , I.G. Danilov a , V.B. Sukhanov b a Zoological Institute of the Russian Academy of Sciences, Universitetskaya Emb. 1, 199034 St. Petersburg, Russia b Borissyak Paleontological Institute of the Russian Academy of Sciences, Profsoyuznaya 123,117997 Moscow, Russia article info Article history: Received 14 February 2013 Accepted in revised form 20 July 2013 Available online Keywords: Adocidae Asia Testudines Turtles abstract This paper presents the description of the skeletal morphology of Adocus amtgai, an adocid turtle from the Late Cretaceous of Mongolia, based on an almost complete skeleton from the upper part of the Bainshire Formation (late TuronianeSantonian) of the Bayshin Tsav locality. Examination of this spec- imen, which is the best preserved among Asian Adocus species, reveals some previously unknown and misunderstood characters of A. amtgai and expands our understanding about variation within Adocus. The phylogenetic analysis of Adocusia (Adocidae þ Nanhsiungchelyidae) places A. amtgai within the Adocus clade as a sister taxon to A. aksary from the Late Cretaceous of Uzbekistan. Crown Copyright Ó 2013 Published by Elsevier Ltd. All rights reserved. 1. Introduction Adocidae Cope, 1870 are a group of freshwater cryptodiran turtles, known mainly from the Cretaceous and Paleogene of Asia and North America (Hutchison, 2000; Sukhanov, 2000; Danilov et al., 2011). Although the adocid record is rather extensive (see Danilov et al., 2011 for the latest review of the Asiatic record of the Adocidae), it is represented mainly by shell material, whereas skulls and elements of non-shell postcrania are known only for four taxa. These are Adocus sp. from the Late Cretaceous of the USA, repre- sented by most parts of the skeleton (Meylan and Gaffney, 1989), A. aksary Nessov in Nessov and Krasovskaya, 1984 from the Late Cretaceous of Uzbekistan, represented by isolated shell fragments and the skull (Danilov and Parham, 2005; Syromyatnikova and Danilov, 2009), Ferganemys verzilini Nessov and Khosatzky, 1977 from the earlyemiddle Albian of Kyrgyzstan, represented by several incomplete skulls and numerous shell fragments (Nessov, 1977; Nessov and Khosatzky, 1977), and Shachemys laosiana Lapparent de Broin, 2004 from the AptianeAlbian of Laos, repre- sented by several partial shells, skulls, posterior cervical vertebrae, remains of the girdles, fore- and hindlimbs (Lapparent de Broin, 2004). All these taxa are important for phylogenetic studies of adocids and turtles in general (Meylan and Gaffney, 1989; Lapparent de Broin, 2004; Danilov and Parham, 2005, 2006; Joyce, 2007; Danilov and Syromyatnikova, 2009a, b; Tong et al., 2009; Sterli, 2010; Syromyatnikova, 2011; Anquetin, 2012). One more adocid taxon known from an almost complete skeleton is A. amtgai Narmandakh, 1985 from the Late Cretaceous of Mongolia (Danilov et al., 2011). This species was described based on a partial shell, including a plastron with an incomplete carapace (Narmandakh, 1985). Later, based on the shell morphology of a second specimen, represented by an almost complete skeleton, this species was placed in the monotypic genus Adocoides Sukhanov and Narmandakh, 2006 (Sukhanov, 2000; Sukhanov and Narmandakh, 2006). To date, only an incomplete carapace of the second specimen has been gured (Sukhanov, 2000, g. 17.20B). Some preliminary results of the study of the second specimen of A. amtgai were reported by Syromyatnikova et al. (2011) and were used in the phylogenetic analysis of other Adocus species (Syromyatnikova and Danilov, 2013; Danilov et al., 2013). In this paper, we present a detailed description of this specimen, include A. amtgai in the phylogenetic analysis of Adocusia Danilov and Parham, 2006 (a clade uniting Adocidae and Nanhsiungchelyidae Yeh, 1966) and discuss the systematic position of this species. 2. Material and methods In addition to the material of Adocus amtgai described below, our study relies on published data and personal observations on the following taxa of the Adocidae: Adocus agilis Cope, 1868a, A. beatus (Leidy, 1865)(Hay, 1908; White, 1972; IGD personal * Corresponding author. E-mail addresses: [email protected] (E.V. Syromyatnikova), [email protected] (I.G. Danilov), [email protected] (V.B. Sukhanov). Contents lists available at ScienceDirect Cretaceous Research journal homepage: www.elsevier.com/locate/CretRes 0195-6671/$ e see front matter Crown Copyright Ó 2013 Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.cretres.2013.07.006 Cretaceous Research 45 (2013) 155e173

Upload: lamdang

Post on 23-Aug-2019

227 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

lable at ScienceDirect

Cretaceous Research 45 (2013) 155e173

Contents lists avai

Cretaceous Research

journal homepage: www.elsevier .com/locate/CretRes

The skeletal morphology and phylogenetic position of Adocus amtgai,an adocid turtle from the Late Cretaceous of Mongolia

E.V. Syromyatnikova a,*, I.G. Danilov a, V.B. Sukhanov b

a Zoological Institute of the Russian Academy of Sciences, Universitetskaya Emb. 1, 199034 St. Petersburg, RussiabBorissyak Paleontological Institute of the Russian Academy of Sciences, Profsoyuznaya 123, 117997 Moscow, Russia

a r t i c l e i n f o

Article history:Received 14 February 2013Accepted in revised form 20 July 2013Available online

Keywords:AdocidaeAsiaTestudinesTurtles

* Corresponding author.E-mail addresses: [email protected]

[email protected] (I.G. Danilov),(V.B. Sukhanov).

0195-6671/$ e see front matter Crown Copyright � 2http://dx.doi.org/10.1016/j.cretres.2013.07.006

a b s t r a c t

This paper presents the description of the skeletal morphology of Adocus amtgai, an adocid turtle fromthe Late Cretaceous of Mongolia, based on an almost complete skeleton from the upper part of theBainshire Formation (late TuronianeSantonian) of the Bayshin Tsav locality. Examination of this spec-imen, which is the best preserved among Asian Adocus species, reveals some previously unknown andmisunderstood characters of A. amtgai and expands our understanding about variation within Adocus.The phylogenetic analysis of Adocusia (Adocidae þ Nanhsiungchelyidae) places A. amtgai within theAdocus clade as a sister taxon to A. aksary from the Late Cretaceous of Uzbekistan.

Crown Copyright � 2013 Published by Elsevier Ltd. All rights reserved.

1. Introduction

Adocidae Cope, 1870 are a group of freshwater cryptodiranturtles, known mainly from the Cretaceous and Paleogene of Asiaand North America (Hutchison, 2000; Sukhanov, 2000; Danilovet al., 2011). Although the adocid record is rather extensive (seeDanilov et al., 2011 for the latest review of the Asiatic record of theAdocidae), it is representedmainly by shell material, whereas skullsand elements of non-shell postcrania are known only for four taxa.These are Adocus sp. from the Late Cretaceous of the USA, repre-sented by most parts of the skeleton (Meylan and Gaffney, 1989),A. aksary Nessov in Nessov and Krasovskaya, 1984 from the LateCretaceous of Uzbekistan, represented by isolated shell fragmentsand the skull (Danilov and Parham, 2005; Syromyatnikova andDanilov, 2009), Ferganemys verzilini Nessov and Khosatzky, 1977from the earlyemiddle Albian of Kyrgyzstan, represented byseveral incomplete skulls and numerous shell fragments (Nessov,1977; Nessov and Khosatzky, 1977), and Shachemys laosianaLapparent de Broin, 2004 from the AptianeAlbian of Laos, repre-sented by several partial shells, skulls, posterior cervical vertebrae,remains of the girdles, fore- and hindlimbs (Lapparent de Broin,2004). All these taxa are important for phylogenetic studies ofadocids and turtles in general (Meylan and Gaffney, 1989;

(E.V. Syromyatnikova),[email protected]

013 Published by Elsevier Ltd. All

Lapparent de Broin, 2004; Danilov and Parham, 2005, 2006;Joyce, 2007; Danilov and Syromyatnikova, 2009a, b; Tong et al.,2009; Sterli, 2010; Syromyatnikova, 2011; Anquetin, 2012). Onemore adocid taxon known from an almost complete skeleton isA. amtgai Narmandakh, 1985 from the Late Cretaceous of Mongolia(Danilov et al., 2011). This species was described based on a partialshell, including a plastron with an incomplete carapace(Narmandakh, 1985). Later, based on the shell morphology of asecond specimen, represented by an almost complete skeleton, thisspecies was placed in the monotypic genus Adocoides Sukhanovand Narmandakh, 2006 (Sukhanov, 2000; Sukhanov andNarmandakh, 2006). To date, only an incomplete carapace of thesecond specimen has been figured (Sukhanov, 2000, fig. 17.20B).Some preliminary results of the study of the second specimen ofA. amtgai were reported by Syromyatnikova et al. (2011) and wereused in the phylogenetic analysis of other Adocus species(Syromyatnikova and Danilov, 2013; Danilov et al., 2013). In thispaper, we present a detailed description of this specimen, includeA. amtgai in the phylogenetic analysis of Adocusia Danilov andParham, 2006 (a clade uniting Adocidae and NanhsiungchelyidaeYeh, 1966) and discuss the systematic position of this species.

2. Material and methods

In addition to the material of Adocus amtgai described below,our study relies on published data and personal observations onthe following taxa of the Adocidae: Adocus agilis Cope, 1868a,A. beatus (Leidy, 1865) (Hay, 1908; White, 1972; IGD personal

rights reserved.

Page 2: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

E.V. Syromyatnikova et al. / Cretaceous Research 45 (2013) 155e173156

observations of YPM 782, holotype of A. punctatus Marsh, 1890),A. bossi Gilmore, 1919, A. kirtlandius Gilmore, 1919, Adocus sp.(Meylan and Gaffney, 1989; photos of skull CCM 60-15 by J.F.Parham; hereafter Adocus sp. 1), Adocus sp. (EVS personal obser-vation of UCMP 129732; hereafter Adocus sp. 2), A. syntheticusCope, 1870 from the Late Cretaceous, A. annexus Hay, 1910 (Hay,1910; Gilmore, 1919) and A. substrictus Hay, 1908 from the Paleo-cene of the USA (Hay, 1908); A. aksary Nessov in Nessov andKrasovskaya, 1984 from the Late Cretaceous of Uzbekistan (seeSyromyatnikova and Danilov, 2009); A. bostobensisSyromyatnikova and Danilov, 2009 from the Late Cretaceous ofKazakhstan (Syromyatnikova and Danilov, 2009, 2013), A. foveatusNessov and Khosatzky in Khosatzky and Nessov, 1977 from theLate Cretaceous of Tajikistan (see Syromyatnikova and Danilov,2009); A. kizylkumensis Nessov, 1981 from the Late Cretaceous ofUzbekistan (see Syromyatnikova and Danilov, 2009); “Adocus”orientalis Gilmore, 1931 from the late Eocene of China andMongolia (Gilmore, 1931; Danilov et al., 2011); A. planus(Sukhanov and Narmandakh, 2006) from the Late Cretaceous ofMongolia (see Syromyatnikova et al., 2012); Adocus sp. from theLate Cretaceous of Canada (EVS personal observations of RTM99.63.1; hereafter Adocus sp. 3); Yehguia tatsuensis (Yeh, 1963)from the Late Jurassic of China (Danilov and Parham, 2006).

The phylogenetic analysis of the clade Adocusia was performedbased on a modified character/taxon matrix of Danilov andSyromyatnikova (2009a,b) with additions from Syromyatnikova(2011) and Syromyatnikova et al. (2012). The following modifi-cations to this character/taxon matrix were made: we addedAdocus amtgai and one additional character: 78, marginals over-lapping onto the costals: (0) beginning with marginal 5; (1)beginning with marginals 3 or 4 (see Appendix 1 for distributionof this new character and Appendix 2 for characters coded forA. amtgai). The final data matrix includes 78 osteological charac-ters for 27 taxa. Our updated matrix was assembled using NDE0.5.0 (Page, 2001) and analyzed with NONA ver. 2 and Wincladaver. 1.00.08 by Ratchet algorithm with 1000 iterations. Characterswere left unordered and considered reversible and of equalweight. Bremer supports were calculated using Autodecay 4.0.1(Eriksson, 1998).

Institutional abbreviations. CCM, Carter County Museum, Ekalaka,Montana, USA; PIN, Borissyak Paleontological Institute of theRussian Academy of Sciences, Moscow, Russia; RTM, Royal TyrrellMuseum of Paleontology, Drumheller, Canada; UCMP, University ofCalifornia Museum of Paleontology, Berkeley, USA; YPM, Yale Pea-body Museum, New Haven, USA.

3. Systematics section

Testudines Batsch, 1788Cryptodira Cope, 1868bAdocusia Danilov and Parham, 2006Family Adocidae Cope, 1870Genus Adocus Cope, 1868a

Adocus amtgai Narmandakh, 1985Figs. 1e13

1985 Adocus amtgai Narmandakh: Narmandakh, p. 86, fig. 1;2000 Adocoides amtgai (Narmandakh): Sukhanov, p. 335, fig. 17.20

(nomen nudum);2006 Adocoides amtgai (Narmandakh): Sukhanov and Narmandakh,

p. 124;2009 Adocus amtgai Narmandakh: Syromyatnikova and Danilov,

p. 76; Syromyatnikova et al., p. 77;

2011 Adocus amtgai Narmandakh: Danilov et al., p. 103, 105, 127;Syromyatnikova et al., p. 202; 2013 Adocus amtgai Narman-dakh: Syromyatnikova and Danilov, 2013, p. 198e200.

Holotype. PIN 3640e2, a partial shell, including plastron withincomplete carapace; Amtgai locality (¼Amtgai Khuduk), DornogovAimag, Mongolia; upper part of the Bainshire Formation, upperTuronianeSantonian.

Referred material. PIN 3640e3, an almost complete skeleton fromthe Bayshin Tsav locality, Dornogov Aimag (Eastern Gobi),Mongolia; upper part of the Bainshire Formation, upper TuronianeSantonian. Previously, this specimen was erroneously indicated tocome from the Amtgai locality (Sukhanov, 2000, p. 335, fig. 17.20b;Syromyatnikova et al., 2009, p. 77; Danilov et al., 2011, p. 105;Syromyatnikova et al., 2011, p. 202).

Emended diagnosis. Adocus amtgai can be differentiated from otherAdocus species by narrower pleurals 2e4 (width is w30% of itslength) and accordingly wide lateral and posterior marginals,shorter pectorals; marginals overlapping onto costals beginningwith marginal 4 (except A. aksary), longer anterior lobe of theplastron (except A. annexus). In addition, it can be differentiatedfrom A. aksary by its relatively longer and wider temporal emargi-nation, transverse prefrontalefrontal suture,more elongated neural6, trapezoid cervical, wider anterior border of epiplastron and gu-lars, and shorter extragulars; from A. planus by extension of gularsonto entoplastron; fromA. beatus bynarrower vertebral 5 andwiderinframarginals. See Tables 1 and 2 for a more detailed comparison.

Description of PIN 3640e3.

Skull (Figs.1, 2). The skull is incomplete and deformed. Part of the leftzygomatic arch ismissing alongwith the vomer, palatine,most of thepalatal parts of the maxilla and premaxillae, most part of the ptery-goids and basisphenoid, basioccipital and posterior end of thesupraoccipital crest. Lateral parts of the otic capsules are damaged ornot preserved. Viewed from above, the anterior part of the skull isslightlyconstricted in thepreorbital region. The similar constriction isobserved inAdocus aksary, whereas inAdocus sp.1 this constriction ismore strongly developed. The orbits seem to be slightly depressedand placed in the anterior third of the skull. They are directed ante-rolaterally, forming an angle ofw30� with the long axis of the skull.Similar shape and orientation of the orbits is known in A. aksary,whereas in Adocus sp. 1 the angle formed by the orbits with the longaxis of the skull is slightly greater. The upper temporal emarginationis relatively long and wide, occupying more than half of the skulllength. Theprocessus trochlearis oticum iswell developed. The lowertemporal emargination is well developed, deepest in its medial partand without a pocket in its anterior part similar to A. aksary.

Skull roof. There is no indication of nasal bones. The prefrontals formthe anterior margin of the skull roof and roof the fossa nasalis.Anteriorly, the prefrontals do not form themedial projection presentinA. aksary. The contact between theprefrontal andmaxilla lies at thelevel of the upper border of the orbit. The ventral process of theprefrontal is broken. The prefrontalefrontal suture is perpendicularto themidline of the skull as in Adocus sp.1, whereas in A. aksary thissutureextends slightlyposterolaterally fromthemidline. The frontalsenter the orbitalmargin forming about one half of its length in dorsalview. In A. aksary, the degree of frontal contribution to the orbitalmargin is similar, whereas in Adocus sp. 1 it is distinctly smaller(approximately one-third of the dorsal orbital margin). The contacts

Page 3: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

Table 1Comparisons of skulls of Adocus aksary (ZIN PH 1/17), A. amtgai (PIN 3640e3) and Adocus sp. 1 (CCM 60-15). ‘?’ denotes impossibility of measuring.

Characters ZISP PH 1/17 PIN 3640e3 CCM 60-15

Anterior part of skullin dorsal view

Narrows gradually Narrows gradually Sharply constricted in preorbital region

Orbits Directed anterolaterally under 30�

to long axis of skullDirected anterolaterally under 30�

to long axis of skullDirected anterolaterally under 40�

to long axis of skullTemporal emargination

at dorsal viewShort and narrow (occupies about50% of skull length, conceals medialportion of processus trochlearis oticum)

Long and wide (occupies more than50% of skull length, medial portionof processus trochlearis oticum visible)

Long and wide (occupies more than 50%of skull length, medial portion of processustrochlearis oticum visible)

Snout profile in lateral view Oblique, projected ventrally Oblique, projected ventrally VerticalPrefrontal-frontal suture Wedge between the prefrontals Transverse TransverseParticipation frontals into

orbital margin.Form about one half of its length Form about one half of its length Form less than half of its length

Groove behind the foramenstapedio-temporale

Present Present Absent

Maxilla Long and low in lateral view Long and low in lateral view High and short in lateral viewCheek emargination Deepest in its medial part,

without pocket anteriorlyDeepest in its medial part,without pocket anteriorly

Deepest anteriorly where it forms a pocket

Maxillary “tooth” Absent ? PresentVentral surface

of basioccipitalWith semicircular depressionand groove

? With semicircular depression with small crest,without groove

Postorbital Does not contribute to zygomatic arch ? Probably contributes to zygomatic arch

E.V. Syromyatnikova et al. / Cretaceous Research 45 (2013) 155e173 157

of the frontals with the postorbitals and the parietals are similar toAdocus sp. 1, except that the sutures in A. amtgai are straight andwithout distinct interdigitations. The dorsal portion of the parietal isrelatively narrow anteriorly and does not conceal the medial portionof the processus trochlearis oticum in dorsal view, similar to Adocussp. 1, but differing from A. aksary, in which this portion is relativelywider. The parietals contact the frontals anteromedially and thepostorbitals anterolaterally along long and short sutures, respec-tively. The parietals contact the prootic posterolaterally and thesupraoccipital posteriorly. The lateral aspects of the descending pa-rietal process (i.e., processus inferior parietalis) support the medialthird of the processus trochlearis oticum. The ventral aspects of thisprocess form the dorsal part of the anterior margin of the foramennervi trigemini. Both postorbitals are preserved. The postorbitalformsparts of the posterior orbitalmargin and the anterior rimof theupper temporal emargination. The contacts of the postorbital withthe frontal and the parietal are at the level of the upper orbitalmargin. The contact of the postorbital with the frontal is equal to itscontact with the parietal on the right side and longer than its contactwith the parietal on the left side. The contact of the postorbital withthe jugal is visible only on the right side. The jugal is almost completeon the right side, whereas on the left side its posterior part is brokenoff. The jugal forms theposteriormarginof the orbit, the anteriorpartof the zygomatic arch and the dorsal part of the cheek emargination.The contacts of the jugal with the maxilla and the quadratojugal arenot clear. The medial process of the jugal is either not visible or notpreserved in both jugals. The quadratojugal and squamosal are pre-served in the right side of the skull, but their sutures are not clear. Thedermal roofing bones are unsculptured as in other Adocus.

Palatal elements. The premaxilla is visible in ventral and anterioraspects; its posterior part is missing. Laterally, the premaxillameets the maxilla to form the triturating surfaces; medially, bothpremaxillae meet on the midline and enter dorsally into the pal-atoquadrate and braincase. The ventral portion of the premaxillaforms the anterior-most portion of the poorly preserved labialridge and triturating surface. The contact of the premaxilla withthe vomer is not observable. The maxilla is preserved on bothsides of the skull. The maxilla articulates with the premaxillaeanteromedially and the prefrontals dorsally, but its contact withthe jugal is unclear. The maxilla is long and low in lateral view,having no tooth-like projection, similar to Adocus aksary, whereas

in Adocus sp. 1 the maxilla is higher and shorter, having the tooth-like projection. A ventral horizontal plate of the maxilla is partlyvisible in ventral view and forms the triturating surface. Thevomer and palatines are missing.

Palatoquadrate and braincase. The quadrate is partially preserved onthe left side of the skull only. The cavum tympani is only partiallypreserved and the incisura columella auris ismissing. The processustrochlearis oticum is well developed and visible on both sides of theskull. It is formedby the quadrate laterally, andmoremedially by theprootic and the inferior process of the parietal. The process istrough-like, strongly concave dorsally and anteriorly, and protrudesinto the adductor fossa. The contribution of the parietal to thisprocess is about one-third of the total, but seems to be greater thanin other Adocus. The opisthotic is preserved on both sides and visibleexternally in dorsal and posterior views. It contacts the prooticanteriorly, the quadrate laterally, the supraoccipital anteromedially,and the exoccipital posteromedially. The opisthotic is relativelyshorter than inAdocus aksaryandAdocus sp.1. The fenestra postoticais not preserved. The foramen nervi trigemini is visible on the leftside,with the prootic forming its dorsalmargin, the parietal formingits anterior margin, and the pterygoid forming its ventral margin.Similar morphology of the foramen nervi trigemini is known inAdocus sp. 1. The occipital bones are represented by the supra-occipital and the exoccipitals, whereas the basioccipital is not pre-served. The supraoccipital forms the supraoccipital crest and dorsalmargin of the foramen magnum. The exoccipitals are not discern-able, but probably formed the lateral margins of the foramen mag-num. The basisphenoid is represented only by two isolatedfragments which do not show any morphological features.

Cranial foramina. The foramen stapedio-temporale is preserved onboth sides of the skull. As in other Adocus, it is large and welldeveloped, formed by the prootic medially and the quadrate later-ally. There is a shallowgroovedirectedposteriorly fromthe foramen.A similar groove is known inA. aksarybut seemsdeeper anddirectedmore medially; this groove is absent in Adocus sp. 1. The foramenposterius canalis carotici interni is missing. Part of the canalis car-oticus internus is exposed on the left side due to a breakage of thebasisphenoidepterygoid region. It is directed from the posterioredge of the pterygoid anteromedially and probably was formed bythe pterygoid. The foramen basisphenoidale is not observable.

Page 4: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

Table 2Comparison of some Late Cretaceous species of Adocus in shell characters (after Syromyatnikova and Danilov, 2009; Syromyatnikova et al., 2012, withmodifications), ‘*’ denotesestimated measurements. ‘?’ denotes impossibility of measuring.

Characters A. aksary A. amtgai A. beatus A. bossi A. kirtlandius A. planus Adocus sp. 1 Adocus sp. 2

Length of the shell 400 mm* 400 mm 500 mm 700 mm* 500 mm* 300 mm* ? 500 mmNuchal emargination Weak Weak Weak or absent Absent ? ? Weak

or absent?

Number of neurals 6 6 6 or 7 6 6 ? ? ?Neural 6 Shortened Elongated Not shortened Shortened Shortened ? Not shortened ?Number of suprapygals 2 2 2 1 2 2 ? ?Cervical Lens- or

trapezoid-shaped,expandedanteriorly orposteriorly

Trapezoid-shaped,expandedanteriorly

Trapezoid-shaped,expandedanteriorly

Trapezoid-shaped,expandedposteriorly

? ? ? ?

Vertebral 5 Narrow Narrow Relatively wide Narrow Relatively wide ? ? ?Pleurals 2e4 ? Narrow (width

makes up 30%of length)

Wide (widthmakes up 60%of length)

Wide (widthmakes up 60%of length)

Wide (widthmakes up 60%of length)

? Wide (widthmakes up 60%of length)

?

Marginals overlappingonto the costals

Beginning withmarginal 3 or 4

Beginningwith marginal 4

Beginningwith marginal 5

Beginningwith marginal 5

Beginningwith marginal 5

? Beginningwith marginal 5

?

Anterior edgeof anterior lobeof plastron

Truncated Truncated Truncated Rounded Rounded ? Rounded Rounded

Ratio of lengthof anterior lobeto length of plastron

0.2 0.3 0.2 0.2 0.2 ? 0.2 0.2

Width of anteriorborder of epiplastron

Less than lengthof epiplastralsymphysis

More than lengthof epiplastralsymphysis

More than lengthof epiplastralsymphysis

More than length2wof epiplastralsymphysis

? ? More thanlengthof epiplastralsymphysis

More thanlength ofepiplastralsymphysis

Dorsal surfaceof epiplastron

Concave Concave ? ? ? ? ? ?

Ratio of lengthof posterior lobeto plastron length

? 37% 38% 32% ? 34% 28% 35%

Ratio of length of bridgeto plastron length

? 38% 40% 41 and 46% ? 35% 49% 40%

Ratio of lengthof xiphiplastron tolength of posteriorlobe

? 75% 73% 79% 72% 76% 67% 72%

Width of anteriorborder of gulars

Less than lengthof epiplastralsymphysis

More than lengthof epiplastralsymphysis

More thanlengthof epiplastralsymphysis

Variable Equal to lengthof epiplastralsymphysis

? More thanlength ofepiplastralsymphysis

Less thanlength ofepiplastralsymphysis

Extension of gularsonto entoplastron

Presentor absent

Present Presentor absent

Present ? Absent Present Present

Ratio of lengthof extragular medialborder to lengthof epiplastralsymphysis

0.9 0.6 0.6 0.7 0.5 ? 0.9 0.6

Pectorals extensiononto the entoplastron

Presentor absent

Present Present Absent Absent Present Absent Absent

Ratio of lengthof pectorals to lengthof plastron

0.12 0.08 0.11 0.12 0.13 ? 0.1 0.09

Ratio of lengthof abdominalsto length of femorals

? 1.6 1.1 2.9 and 1.5 1.3 0.9 2.1 1.4

Numberof inframarginals

? 3 or 4 pairs 4 pairs 3 pairs 4 pairs 4 pairs 4 pairs ?

Inframarginals ? Relatively wide Relatively narrow Relatively narrow Relativelywide

Relativelywide

Relativelynarrow

?

Inframarginal 1 ? Longer thanwide or widerthan long

Wider than long Longer than wide Longerthan wide

Widerthan long

Longerthan wide

Longerthan wide

Inframarginal 3 ? Shortened Elongated Elongated Elongated Shortened Shortened ?

E.V. Syromyatnikova et al. / Cretaceous Research 45 (2013) 155e173158

Lower jaw. The lower jaws are represented by both rami. The leftramus and anterior part of the right ramus were separated from theskull during the preparation, whereas the middle portion of theright ramus was left articulated with the skull. The trituratingsurface of the dentary is only slightly concave in the posterior part,unlike Adocus sp. 1. The presence of the dentary pocket known in

Adocus sp. 1 is unclear. The coronoid process is relatively low andreaches its greatest height in the posterior part of the lower jaw. InAdocus sp. 1, the coronoid process is much higher and reaches itsgreatest height in the middle of the lower jaw. The coronoid, sur-angular and angular are visible in lateral view. Other elements ofthe lower jaw are missing.

Page 5: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

E.V. Syromyatnikova et al. / Cretaceous Research 45 (2013) 155e173 159

Shell (Figs. 3e5). The estimated length of the shell is w400 mm,whereas its width is hard to estimate due to deformation. Theplastron is estimated to approximate 75% of the carapace lengthand almost reach the carapace rim anteriorly. The surface of theshell is covered by typical adocid sculpturing, consisting of

Fig. 1. Skull (AeF) and lower jaw fragment (G,H) ofAdocus amtgai (PIN3640e3):Aedorsal viewview, G e lateral view, H e dorsal view.

relatively small and regular grooves and pits (see Danilov et al.,2011). The scute sulci are narrow and shallow.

Carapace. Although the carapace is missing some parts in its mid-dleeposterior portion, it can be almost fully reconstructed. The

, Be ventral view, Ce right lateral view,De left lateral view, Ee anterior view, Feposterior

Page 6: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

Fig. 2. Skull (AeF) and lower jaw fragment (G, H) of Adocus amtgai (PIN 3640e3), explanation drawings of Fig. 1: A e dorsal view, B e ventral view, C e right lateral view, D e leftlateral view, E e anterior view, F e posterior view; G e lateral view, H e dorsal view. See Fig. 2 for abbreviations and designations. Abbreviations: ac e acetabulum; acrp e acromionprocess; ang e angular; bs e basisphenoid; c e costal; cci e canalis caroticus internus; ce e cervical; ch e caput humeri; cor e coronoid; den e dentary; ectf e ectepicondylar

E.V. Syromyatnikova et al. / Cretaceous Research 45 (2013) 155e173160

Page 7: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

Fig. 3. Shell of Adocus amtgai (PIN 3640e3): A e shell, dorsal view, before preparation, B e shell, ventral view, before preparation, C e carapace, dorsal view, after preparation,D e plastron, ventral view, after preparation, E e carapace, dorsal view, explanation drawing, F e plastron, ventral view, explanation drawing. See Fig. 2 for abbreviations anddesignations.

E.V. Syromyatnikova et al. / Cretaceous Research 45 (2013) 155e173 161

carapace is oval-shaped, widened anteriorly (as wide anteriorly asposteriorly), with a shallow nuchal emargination that formed bythe nuchal and anterior part of the peripherals 1 similar to those inAdocus aksary and A. bossi. The carapace is low-domed, as in otherAdocus, its anterior part in lateral view looks more steep than theposterior one. The anterior part of the carapace has a strongly up-ward free border.

The nuchal is relatively wide (wider than long) hexagonal andslightly emarginated anteriorly. Its free edge is rounded andupturned in the cross-section. The anterior border of the nuchal isrelatively wide (ratio of the nuchal anterior width to its maximalwidth is seven-tenths, that is the same as in A. aksary). The antero-lateral and posterolateral borders of the nuchal are nearly straight orslightly convex and about of the same length. As in other Adocus, theposterior border of the nuchal is concave and contacts with neural 1.

foramen; egu e extragular; ent e entoplastron; epi e epiplastron; ex e exoccipital; f e frostapedial artery; gu e gular; if e intertubercular fossa; im e inframarginal; is e ishium; lp e

metatarsal; mx e maxilla; n e neural; nu e nuchal; op e opisthotic; p e peripheral; pa e pph e phalanx; ple e pleural; pmx e premaxilla; po e postorbital; pp e pectineal process; pqu e quadrate; r e radius; rh e ribhead; sc e scapula; so e supraoccipital; sp e suprapygal;minor; u e ulna; ung e ungual; v e vertebral. Arabic numerals designate element numbers;surfaces are indicated by hatched lines.

Theneurals are represented by six elements, howeverneurals 4e6 are only partially preserved and their shapes can be reconstructedbased on the surrounding plates. All neurals are relatively wide andthin as in some other species of Adocus (A. aksary, A. beatus,A. bostobensis, andA. kizylkumensis). The neural formula is typical formost of the Adocidae: neural 1 is hexagonal short-sided posteriorly,neural 2 is tetragonal, neurals 3e5 are hexagonal and short-sidedanteriorly and neural 6 is heptagonal. Neural 2 is relatively shortas seen in A. beatus and A. aksary. Neural 6 is relatively elongatedwith an angled posterior border which slightly wedges betweencostals 7. Most other species of Adocus (A. aksary, A. beatus, A. bossi,and A. kirtlandius) have six neurals, however, among them, thesimilarly elongated heptagonal neural 6 is present only in A. beatus.

There are two suprapygals. Suprapygal 1 is a relatively small,elongated, oval-shaped element. Suprapygal 2 is a relatively wide,

ntal; fn e foramen orbito-nasale; fst e foramen stapedio-temporale; g e groove fromlateral process; m e marginal; metp e metischial process; mp e medial process; mt e

arietal; pal e palatine; pat e area for pelvic attachment; pe e pectoral; pf e prefrontal;r e prootic; pt e pterygoid; pta e patellar tendon attachment; pu e pubis; py e pygal;ss e skin-scute sulcus; tib e tibial condyle; tmaj e trochanter major; tmin e trochantertentative sutures are shown with dashed lines; matrix is shaded with dark gray; broken

Page 8: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

Fig. 4. Shell of Adocus amtgai (PIN 3640e3): A e carapace, ventral view, B e explanation drawing of the same, C e plastron, dorsal view, D e explanation drawing of the same,E e carapace, anterior view, F e explanation drawing the same, G e carapace, lateral view, H e carapace, explanation drawing of the same. See Fig. 2 for abbreviations and designations.

E.V. Syromyatnikova et al. / Cretaceous Research 45 (2013) 155e173162

octagonal element that contacts suprapygal 1 along concave ante-rior borders, costals 8 along straight anterolateral borders, pe-ripherals 10 along convex lateral borders, peripherals 11 alongslightly convex posterolateral borders, and the pygal along a convexposterior border. Similar suprapygals are known in other Adocus,except A. bossi, which has no suprapygal 1.

The pygal is a slightly longer than wide, trapezoidal (ratio of theanterior pygal width to its posterior width is about three-fifths),and has straight lateral borders. The free edge of the pygal isslightly upturned. The described pygal morphology is similar tothat in other members of the Adocidae.

The costals are represented by eight pairs, which are nearlycomplete, except that most of left costals 6 and 7 are missing. Rib-heads and rib thickenings of the costals areweak as in other adocids,except Yehguia tatsuensis. On the internal surface of costal 1, scars ofthoracic ribs 1 and 2 lie close to each other. Costals 7 contact eachother along the midline. Costals 8 contact each other along themidline at the anterior half of their medial length and separated bysuprapygal 1 at the posterior half of their medial length. Amongadocids, similar posterior-costalmorphology is seen inAdocusbeatus.

The peripheral series is represented by almost complete pe-ripherals 1e11 on the right side and peripherals 1e6, 10 and 11 onthe left side. The anterior peripherals (1e3) appear trapezoidal indorsal view with short medial and long lateral borders. Their freeedges are upturned and angled in the cross-section. The bridgeperipherals (4e8) are mostly rectangular and partly broken along

the free edges. The posterior peripherals are wider than the ante-rior ones with upturned and angled free edges. There is somevariation in peripherals of the right and left sides. Peripheral 1 onthe right side fused with peripheral 2 forming a single wideelement, which at least half as wider than peripheral 1 of the leftside. Peripheral 3 on the left side contacts only with costal 1,whereas on the right side it reaches costal 2 as well. Peripheral 4 onthe left side is longer medially than laterally and contacts withcostals 1e3, whereas on the right side it is shorter medially thanlaterally and contacts only costal 2. Peripheral 5 on the left side isshorter medially than laterally and contacts only costal 3, whereason the right side it is longer medially than laterally and contactscostals 2e3. Peripherals 6e9 have convex medial borders thatslightly wedge between the corresponding costals. Peripheral 10contacts costals 7 and 8 and suprapygal 2 medially. Peripheral 11 isshorter than peripheral 10.

All carapacial scutes are observable, however the shape of thecervical and vertebrals 3e5 are partially reconstructed as narrowand trapezoidal, slightly expanded anteriorly. Among other Adocus,a similar cervical is known in A. aksary and A. beatus. Vertebral 1 istrapezoidal, widened anteriorly and in contact with marginals 2, asin all other Adocus. Vertebrals 2e4 are longer thanwide. Vertebral 2seems to be shorter than vertebral 3 which is longer than others.Vertebral 4 is shorter than vertebrals 2 and 3 and narrowed pos-teriorly. Vertebral 5 is small and ovate as in A. aksary and A. bossi. Inother species of Adocus, vertebral 5 is usually wider (A. kirtlandius

Page 9: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

Fig. 5. Reconstruction of the shell of Adocus amtgai. A e dorsal view, B e ventral view.

E.V. Syromyatnikova et al. / Cretaceous Research 45 (2013) 155e173 163

and A. hesperius) or as wide as more anterior vertebrals (A. beatus).The pleurals are longer than wide and getting narrower from theanterior to the posterior ones. Pleural 1 is anomalous and differs inshape on the left and right sides. The left pleural 1 is wide poste-riorly and overlaps peripheral 4; its lateral border has a strongindentation due to the invasion of marginal 4. The right pleural 1 isnarrowed in posterior part and does not reach peripheral 4; theindentation of its lateral border is absent. Pleurals 2e4 areextremely narrow (ratio of the pleural width to its length is aboutthree-tenths). In other Adocus, the pleurals are relatively wider(ratio of the pleural width to its length is usuallyw0.6; see Table 2).The marginals are represented by a virtually complete set of scutes(12 pairs) except for the left marginals 8e11. Marginals 1e3 arerestricted to the peripherals and cover approximately half of theirexternal surface. Marginal 4 overlaps onto the costals. The leftmarginal 4 overlaps costal 1 only by its anteromedial part; whereasits posteromedial part does not reach the costals. The right mar-ginal 4 overlaps costal 2 by its posteromedial part. Thus, the over-lapping of the marginals onto costals begins with marginal 4. Inmost other species of Adocus, overlapping begins with marginal 5,whereas in A. aksary it also begins with marginals 4 (or even 3).Marginals 5e10 overlap the corresponding costals (2e8), coveringabout one-third of their width. Marginals 11 and 12 cover morethan half of the length of costals 8 and suprapygal 2.

The skin-scute sulcus lies very close to the free edge of thenuchal and in the middle part of the peripherals.

Plastron. The plastron is almost completely preserved. The anteriorlobe of the plastron is wider than long (length of anterior lobe isabout one half of its width), truncated anteriorly and does not reachthe anterior carapace rim. In other species of Adocus, the anteriorlobe is either rounded (A. annexus, A. bossi, A. kirtlandius, andA. substrictus) or truncated (A. aksary, A. beatus, A. syntheticus, and“A.” orientalis) anteriorly. The length of the anterior lobe makes up

30% of the plastron length. In other Adocus, the length of theanterior lobe varies from 22 to 23% (A. beatus and A. kirtlandius) to27% (A. annexus) of the plastron length. The posterior lobe narrowsposteriorly and is longer and narrower at the base than the anteriorone. The length of the plastral bridge is w40% of the plastronlength. The gular and anal notches are absent as in other Adocus.The epiplastron has a relatively wide and straight (truncated)anterior border, which is greater than the length of the epiplastralsymphysis that is also known from A. beatus and A. foveatus. Thedorsal surface of epiplastron is concave. The entoplastron is a largehexagonal element, wider than long and with poorly developedlateral borders. The similar hexagonal entoplastron is characteristicof all Adocus species. On the dorsal surface of the entoplastron, a Y-shaped system of ridges is poorly visible. The hyoplastra andhypoplastra have no peculiarities and make equal contributions tothe bridge length. The left hyoplastron and the right hypoplastronhave a long contact that separates the right hyoplastron from theleft hypoplastron. The xiphiplastron is longer than wide, narrowedposteriorly and with a rounded lateral border. The length of thexiphiplastron is w75% of the posterior lobe length along themidline. Internally, it bears an oval area for the pelvic attachment.

The plastral scutes are represented by a complete set includingtwo pairs of gulars, extragulars, humerals, pectorals, abdominals,femorals, anals and four pairs of inframarginals. The gulars arerelatively wide (anterior width of the gular is more than the lengthof the epiplastral symphysis), like in A. beatus. The gulars slightlyoverlap the entoplastron as in most other species of Adocus (exceptA. planus and sometimes A. aksary and A. beatus). The extragularsare relatively small covering about one-fourth of the external sur-face of the epiplastra, with short medial borders that compriseabout three-fifths of length of the epiplastral symphysis. The shortmedial borders of the extragulars are known in A. beatus, whereasin other species they are usually longer. The lateral borders of theextragulars are strongly elongated reaching the epi-hyoplastral

Page 10: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

E.V. Syromyatnikova et al. / Cretaceous Research 45 (2013) 155e173164

suture posteriorly similar to cf. “Adocus” orientalis (see Danilovet al., 2011). The pectorals are strongly shortened (their mediallength is about eight hundredth of the plastron length), longermedially than laterally and slightly overlapping the entoplastron.They contribute to the rim of the axillary notch, as in other adocids.The pectorals of other species of Adocus are variable: overlapping(A. annexus and A. beatus) or not overlapping the entoplastron(A. agilis, A. bossi, A. kirtlandius, “A.” orientalis, and A. substrictus).The abdominals are long and contribute to the rim of the inguinal

Fig. 6. Cervical vertebrae of Adocus amtgai (PIN 3640e3), photographs: A e cervical 4, B e

notch. The femoraleanal sulcus is S-shaped. The inframarginals arerelatively wide and partly extend onto the peripherals. Their sizeand shape are variable; there are three inframarginals on the rightside and five inframarginals on the left side. On the right side,inframarginals are similar to those of PIN 3640–2: inframarginal 1is large, narrowed medially and widened laterally, and does notcontribute to the pectoral rim; inframarginal 2 is about twice aslonger as inframarginal 1, spans the hyo-hypoplastral suture;inframarginal 3 contributes to the rim of the inguinal notch. On the

cervical 5, C e cervical 7, D e cervical 8. Lateral views of C and D are mirror images.

Page 11: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

E.V. Syromyatnikova et al. / Cretaceous Research 45 (2013) 155e173 165

left side, inframarginal 1 is relatively short and widely contributesto the pectoral rim; inframarginal 2 is longer than the others, spansthe hyo-hypoplastral suture only by its most posterior part; infra-marginal 3 is slightly shorter than inframarginal 2; inframarginal 4seems short and contribute to the rim of the inguinal notch. Anadditional inframarginal scute is present on the left side of theplastron between the abdominal and inframarginals 2 and 3. Itspans the hyo-hypoplastral suture. In other Adocus, the

Fig. 7. Cervical vertebrae of Adocus amtgai (PIN 3640e3), explanation drawin

inframarginals are usually represented by four pairs, except forA. bossiwhich has three. The midline sulcus is strongly sinuous. Theskin-scute sulcus lies close to the free edge of the plastral lobes.

Cervical vertebrae (Figs. 6, 7). The cervical series is incomplete andrepresented only by the partly preserved cervical vertebrae 4, 5, 7and 8. All cervicals are opisthocoelous as in Adocus sp. 1. Centra 4, 5and 7 are at least twice as long as they are wide; centrum 8 is as

gs of Fig. 6: A e cervical 4, B e cervical 5, C e cervical 7, D e cervical 8.

Page 12: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

Fig. 8. Left pectoral girdle of Adocus amtgai (PIN 3640e3): A e scapula, anterior view, B e explanation drawing of the same, C e scapula, posterior view, D e explanation drawing ofthe same, E e coracoid, dorsal view, F e explanation drawing of the same, G e coracoid, ventral view, H e explanation drawing of the same. See Fig. 2 for abbreviations anddesignations.

E.V. Syromyatnikova et al. / Cretaceous Research 45 (2013) 155e173166

Page 13: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

Fig. 9. Left humerus (AeH) and left manus (IeL) of Adocus amtgai (PIN 3640e3): A e posterior view, B e explanation drawing of the same, C e dorsal view, D e explanation drawingof the same, E e ventral view, F e explanation drawing of the same, G e anterior view, H e explanation drawing of the same, I e ventral view, J e explanation drawing of the same,K e dorsal view, L e explanation drawing of the same. See Fig. 2 for abbreviations and designations.

E.V. Syromyatnikova et al. / Cretaceous Research 45 (2013) 155e173 167

Page 14: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

Fig. 10. Pelvic girdle of Adocus amtgai (PIN 3640e3): A e dorsal view, B e explanation drawing of the same, C e left lateral view, D e explanation drawing of the same, E e ventralview, F e explanation drawing of the same. See Fig. 2 for abbreviations and designations.

E.V. Syromyatnikova et al. / Cretaceous Research 45 (2013) 155e173168

Page 15: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

E.V. Syromyatnikova et al. / Cretaceous Research 45 (2013) 155e173 169

wide as long. Centrum 4 has a single convex anterior articularsurface; the posterior surface is partly missing but seems to besingle concave. Centrum 5 has a single convex anterior articularsurface; its posterior surface is missing. Centrum 7 has a doubledconvex anterior articular surface; its posterior surface is missing.Centrum 8 has a doubled concave posterior articular surface; theanterior surface is partly missing but appears to be doubled convex.Partial ventral keels are present on centra 4 and 5. They are rep-resented by small, thin fragments which do not allow their shape tobe restored. The neural arches have widely separated pre- andpostzygapophyses. The postzygapophyses become more robustposteriorly.

Pectoral girdles (Fig. 8). Both right and left pectoral girdles arealmost completely preserved. The scapular and acromion processesof the scapula meet at an angle of w90�. The scapular process isrod-like and rounded in cross-section. The acromion process isslightly flattened and bowed downwards. It is shorter than thescapular process (ratio of its length to length of the scapular process

Fig. 11. Right femur of Adocus amtgai (PIN 3640e3): A e anterior view, B e explanation drawF e explanation drawing of the same, G e posterior view, H e explanation drawing of the

is 0.7), like in Adocus sp. 1. The coracoid is long, flat, and relativelywide (its length is approximately one-third of its width) with atruncated distal edge. It is longer than the acromion process of thescapula and only slightly shorter than the body of the scapula. InAdocus sp. 1, the coracoid is about the same proportion, but has arounded distal edge.

Forelimb (Fig. 9). The left and proximal part of the right humeri arealmost complete. The caput humeri are large and oval in dorsalview. It is located at an angle of w90� to the shaft of the humerus.The medial process of the humerus is about three times larger thanthe lateral process; between them is a well-developed inter-tubercular fossa. The medial process extends posteriorly from theshaft at an angle of w40�, and extends as far medially as does thecaput humeri. The lateral process diverges anteriorly from the shaftat an angle ofw10� and does not extend medially to the level of thecaput humeri. The two distal trochanters, the capitellum andtrochlea, are approximately equal in size. The ectepicondylar fora-men is closed. Dorsally under the caput humeri, there is a concave

ing of the same, C e dorsal view, D e explanation drawing of the same, E e ventral view,same. See Fig. 2 for abbreviations and designations.

Page 16: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

E.V. Syromyatnikova et al. / Cretaceous Research 45 (2013) 155e173170

attachment for mm. latissimus dorsi et teres major. The describedmorphology of the humerus is similar to that of Adocus sp. 1.

Other elements of the forelimb are represented by the completeleft radius, small distal part of the ulna and metacarpal with pha-langes which are encased in the sediment and have shifted fromtheir original natural position. The distal part of the ulna is flatteneddistally and seems to bewidened distally. Themanus is like those ofmany other turtles and has a primitive number of elongatedmetacarpals and phalanges. The carpus is not preserved. Themetacarpals and phalanges are represented partly for 1e4 digits.The phalanges are 2e3e3e3e3 with claws apparently present onall five digits.

Pelvic girdle (Fig. 10). The pelvic girdle is incomplete and has beendistorted. The ilia form the dorsal third of the acetabulum andextend posterodorsally at an angle of nearly 120�. The thelial pro-cesses are represented by the virtually complete right element,whereas the left element is absent. The thelial process is weak as inAdocus sp. 1. The pubes make up one-third of the acetabulum. Theyare slightly narrowed anteriorly with broad interpubic contact thatextend anteriorlywell beyond the pectineal processes.Whether thepubis projected a process into the thyroid fenestra at the midline isunclear, but probably absent in contrast to Adocus sp. 1. The pecti-neal processes are large and elongate, lying at nearly right angles tothe midline; however, such a position is reconstructed and could be

Fig. 12. Right tibia (AeH) and fibula (IeP) of Adocus amtgai (PIN 3640e3): A e dorsal view, Bsame, E e ventral view, F e explanation drawing of the same, G e posterior view, H e exK e medial view, L eexplanation drawing of the same, M eventral view, N e explanation drabbreviations and designations.

incorrect. In Adocus sp. 1, the pectineal processes appear to besomewhat shorter. The ischia comprise the posteroventral third ofthe acetabulum; the ventral portions of ischia are missing. Themetischial processes are long, as in Adocus sp. 1. The thyroidfenestra is not divided.

Hindlimb (Figs. 11e13). The femur is represented by both elements,but their distal ends are damaged. The estimated length of the fe-mur is 73 mm. The femur is slightly curved and expanded medially.There are two large proximal trochanters which are about the samesize and separated by the well-defined and deep intertrochantericfossa. The trochanter minor does not extend proximally as much asthe trochanter major and does not extend medially beyond thehead. The femoral head is oval, longer than narrow. The distal endof the femur is partly damaged and represented only by part of thetibial condyle. The distal articular surface appears to face ventrallyand is divided into two distinct condyles. The tibia is represented bythe complete right and proximal part of the left elements; its lengthis 55 mm. The tibia is expanded to bear the articulation with thefemur proximally and tapers distally. The shaft of the tibia bowsaway from the fibula as in other turtles. The proximal articularsurface has a medial concavity and a weak lateral convexity for thearticulation with the condyles of the femur. The patellar tendonattachment site (¼cnemial crest) is identifiable as a rugose areawith a concavity on the dorsal surface of the tibia. In Adocus sp. 2,

e explanation drawing of the same, C e anterior view, D e explanation drawing of theplanation drawing of the same, I e dorsal view, J e explanation drawing of the same,awing of the same, O e lateral view, P e explanation drawing of the same. See Fig. 2 for

Page 17: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

E.V. Syromyatnikova et al. / Cretaceous Research 45 (2013) 155e173 171

the patellar tendon attachment site looks weakly developed andsmooth. At the proximal part along the shaft of the tibia, there is arugose protuberance for attachment of the flexor tibialis internus.The fibula is preserved by an almost complete right element withlength is 56 mm. The fibula is relatively straight, its distal endexpanded and the proximal one only slightly larger than the shaft.The plane of distal expansion is twisted from the plane of theproximal expansion for w30� as in most other turtles. The distalpart of the fibula shaft is flat on its dorsal side but there is athickened ridge on the ventral side that ends at an expanded part ofthe distal articulation surface. The tarsals and part of metatarsalsare missing from the pes. The preserved metatarsals and phalangesare disarticulated and do not allow identifying its exact position inthe pes. However, it is clear that they are distinctly larger than thecorresponding bones of the manus.

4. Discussion

The newly described material of an almost complete Adocusamtgai skeleton, allows us to emend the diagnosis of this species(see Systematic paleontology section).

Adocus amtgai is assigned to Adocidae based on the followingcharacters (following Danilov and Syromyatnikova, 2008): (1)adocid-type of sculpturing, (2) narrow and shallow scute sulci, (3)weak ribheads, (4) weak rib thickenings of the costals, (5) longerthan wide pygal, (6) relatively short plastral bridges (<50% ofplastron length), (7) relatively long posterior plastral lobe (>30% ofplastron length), (8) presence of the pectoral contribution to theaxillary rim, (9) ventromedial edge of marginal 6 not expanded, and(10) absence of overlapping scutes onto the dorsal surface of plas-tral lobes. Adocus amtgai is assigned to the genus Adocus based onthe marginals that overlap onto the costals in the middle andposterior parts of the carapace. Most of the described characters ofthe skull and shell (see Tables 1 and 2) indicate that A. amtgai is

Fig. 13. Pes of Adocus amtgai (PIN 3640e3): A e dorsal view, B e explanati

most similar to A. aksary. One of these characters is the extension ofthe marginals onto the costals beginning with marginals 4. InA. aksary, such an extension begins with marginals 3 or 4, whereasin most other species of Adocus it begins with marginals 5 (see alsoSyromyatnikova and Danilov, 2013; Danilov et al., 2013). Anotherimportant character of A. amtgai is a very narrow vertebral 5(narrower than anterior vertebrals). Among other species of Adocus,the similar narrow vertebral 5 is known in A. aksary and A. bossi.

Previously, Adocus amtgaiwas placed in its own genus Adocoidesbased on the following characters identified by Sukhanov (2000)and Sukhanov and Narmandakh (2006): reduction of the cervical;presence of relatively wide vertebrals 2e4; significantly widermarginals 4e12; very narrow pleurals 2e4; very narrow vertebral5; larger size of the plastron; smaller size of gulars and extragulars;strongly shortened of pectorals and correspondingly elongatedabdominals, greater width of inframarginals. New data on Adocusamtgai’s morphology allow us to comment on these characters. Ourobservation of PIN 3640e3 does not confirm the presence of an“almost complete, reduction of the cervical and the appearance of acontact between the first marginals” (Sukhanov, 2000, p. 336).Actually, A. amtgai has a narrow cervical that excludes a contactbetween marginals 1. Vertebrals 2e4 of A. amtgai are similar inwidth to those of A. aksary (the length of vertebral 3 is same as itswidth in both species). The narrow vertebral 5 is known in someother species of Adocus (see above). The ratio of the length of theplastron to the length of the shell varies among Adocus: A. aksary e

76%; A. beatus e 65%; A. planus e 80%; Adocus sp. 2 e 70%;A. substictus e 78%. In A. amtgai it is w75% that is almost the sameas in A. aksary. The gulars of A. amtgai are relatively large; theirwidth is more than length of the epiplastral symphysis. Theextragulars of A. amtgai are relatively small but similar in size tothose of A. beatus and Adocus sp. 3. The inframarginals of A. amtgaiare relatively wide that is also known in some other species ofAdocus. Thus, the above-mentioned characters are highly variable

on drawing of the same. See Fig. 2 for abbreviations and designations.

Page 18: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

Fig. 14. Phylogeny of Adocusia showing the hypothesized position of Adocus amtgai.This is a strict consensus of 85 phylogenetic trees resulting from this study (seeDiscussion for description of the tree). Numbers designate Bremer support indices.

E.V. Syromyatnikova et al. / Cretaceous Research 45 (2013) 155e173172

within Adocus and insufficient to assign A. amtgai to a separategenus. On the other hand, other characters of A. amtgai (wide lateraland posterior marginals and very narrow pleurals 2e4; the shorterof the pectorals) are unique for this species and may represent itsautapomorphies. According to criteria used for Adocus, the differ-ences between A. amtgai and other species of this genus (seeTable 2) reflect mere species-level differences.

Our study also presents new observations on Adocusmorphology in general, including the detailed morphology of theskull and elements of the non-shell postcrania. Some elements(femur, tibia, fibula, bones of the pes) were previously unknown forthe genus. Themorphology of the non-shell postcranial elements ofA. amtgai on the whole is similar to that of Adocus sp. 1, which isanother taxon of Adocus with known morphology of the non-shellpostcrania. The only difference between these taxa in themorphology of the non-shell postcrania is the presence of thetruncated distal edge of coracoid in A. amtgai, whereas in Adocus sp.1 it is rounded. As a result, this study makes A. amtgai the best-known Asiatic representatives of the genus and one of the best-known members of the genus in general.

The examination of both specimens of Adocus amtgai (PIN3640e2 and PIN 3640e3) allows us to note variation of this speciesin the following shell characters: presence/absence of the cervical,shape of the extragulars, shape and number of the inframarginals,and degree of sinuosity of the midline sulcus. The cervical in PIN3640e3 is narrow and trapezoidal, slightly expanded anteriorly,whereas in PIN 3640e2 it seems to be absent (Narmandakh, 1985).The extragulars are more elongated posterolaterally and mediallyin PIN 3640e3 than in PIN 3640e2. The inframarginals in PIN3640e2 are represented by three pairs (Sukhanov, 2000, p. 336),whereas in PIN 3640e3 there are three inframarginals on the rightside and five inframarginals on the left side. In PIN 3640e2, theright side the inframarginal 1 almost does not contribute to thepectoral rim and is separated from it by a small contact of thepectoral and marginal 4. The small additional inframarginal sepa-rating abdominal from the rim of the inguinal notch is mistakenlyshown (Narmandakh, 1985, fig. 1g; Sukhanov, 2000, fig. 17.20A). Onthe left side of PIN 3640e2, inframarginal 1 contributes broadly tothe pectoral rim. In PIN 3640e3, the right side of the inframarginalsare similar to those of PIN 3640e2 in the presence of three ele-ments and inframarginal 1 separated from the pectoral rim. How-ever, on the left side of PIN 3640e3 the inframarginals differ fromPIN 3640e2 in presence of five elements. The midline sulcus isstrongly sinuous in PIN 3640e3 along the whole length of theplastron, whereas in PIN 3640e2 it is sinuous only along thepectorals.

All the data obtained on A. amtgai allow us to undertake aphylogenetic analysis. The result of our phylogenetic analysis con-sists of 85 trees with 151 steps, consistency index is 0.56, andretention index is 0.76. The resulting strict consensus tree is givenin Fig. 14. This tree demonstrates that A. amtgai is placed within theAdocus where it forms a clade with A. aksary. This result supportsthe assignment of A. amtgai to the genus Adocus. The clade unitingA. amtgai and A. aksary (both from the Late Cretaceous of Asia) issupported by a single synapomorphy, the overlapping of the mar-ginals onto the costals beginning with marginals 3 or 4. Two otherspecies of Adocus (A. bostobensis and a new species from the LateEocene of China) known to have the overlapping of the marginalsonto the costals beginning with marginals 4 also belong to thisclade (Syromyatnikova and Danilov, 2013; Danilov et al., 2013). Inaddition to the members of this clade, there are other species ofAdocus in Asia characterized by the overlapping of the marginalsonto the costals beginning with marginal 5 (A. foveatus,A. kizylkumensis) or unknown for this and other important char-acters (see Syromyatnikova and Danilov, 2009). In all North

American Adocus, in which this character is observable, the mar-ginals overlap onto the costals beginning with marginal 5. Thecharacter of costal overlap beginning with marginal 5 is consideredto be primitive in our analysis. Adocus planus, the third Asiaticspecies of Adocus in our analysis is placed outside theA. amtgai þ A. aksary clade in polytomy with other species ofAdocus. The topology of the rest part of the tree agrees with pre-vious analyses (Danilov and Syromyatnikova, 2009a,b;Syromyatnikova, 2011; Syromyatnikova et al., 2012).

5. Conclusions

The virtually complete skeleton of Adocus amtgai, the best pre-served among Asian species of the genus, is described in detail. Thismaterial of A. amtgai allows us to emend the diagnosis of the spe-cies. The result of our study confirms the assignment of A. amtgai tothe genus Adocus based on the marginals that overlap onto thecostals in the middle and posterior parts of the carapace. Most ofthe described characters of the skull and shell indicate thatA. amtgai is most similar to A. aksary. Our study also presents newobservations on Adocus morphology in general, especially in thedetailed morphology of the skull and elements of the non-shell

Page 19: The skeletal morphology and phylogenetic position of ...paleot.weebly.com/uploads/1/5/0/7/15071262/syromyatnikova_et_al_2013... · The skeletal morphology and phylogenetic position

E.V. Syromyatnikova et al. / Cretaceous Research 45 (2013) 155e173 173

postcrania. The morphology of some elements (femur, tibia, fibula,bones of the pes) were previously unknown for the genus.

Acknowledgments

The authors thank J.F. Parham for offering photographs of theskull of Adocus sp. 1 (CCM 60-15), Julien Claude (Institut des sci-ences de l’évolution, Montpellier, France) and an anonymousreviewer for reviewing the manuscript and useful comments. Thisstudy was done with financial support from Grants of the Presidentof the Russian Federation to the Leading Scientific Schools (NSh-6560.2012.4), Russian Foundation for Basic Research 12-05-31015_mol_a to EVS.

References

Anquetin, J., 2012. Reassessment of the phylogenetic interrelationships of basalturtles (Testudinata). Journal of Systematic Palaeontology 10, 3e45.

Batsch, G.C., 1788. Versuch einer Anleitung, zur Kenntnis und Geschichte der Thiereund Mineralien. Akademische Buchhandlung, Jena.

Cope, E.D., 1868a. On some Cretaceous Reptilia. Proceedings of the Academy ofNatural Sciences of Philadelphia 20, 233e234.

Cope, E.D., 1868b. On the origin of genera. Proceedings of the Academy of NaturalSciences of Philadelphia 20, 242e300.

Cope, E.D., 1870. On the Adocidae. Proceedings of the American Philosophical So-ciety 11, 547e553.

Danilov, I.G., Parham, J.F., 2005. A reassessment of the referral of an isolated skullfrom the Late Cretaceous of Uzbekistan to the stem-testudinoid turtle genusLindholmemys. Journal of Vertebrate Paleontology 25, 784e791.

Danilov, I.G., Parham, J.F., 2006. A redescription of ‘Plesiochelys’ tatsuensis from theLate Jurassic of China, with comments on the antiquity of the crown cladeCryptodira. Journal of Vertebrate Paleontology 26, 573e580.

Danilov, I.G., Syromyatnikova, E.V., 2008. New materials on turtles of the familyNanhsiungchelyidae from the Cretaceous of Uzbekistan and Mongolia, with areview of the nanhsiungchelyid record in Asia. Proceedings of the ZoologicalInstitute of the Russian Academy of Sciences 312, 3e25.

Danilov, I.G., Syromyatnikova, E.V., 2009a. Phylogenetic analysis of turtles of thesuperfamily Adocoidea. In: Rozanov, A.Yu., Lopatin, A.V., Parkhaev, P.Yu. (Eds.),Modern Paleontology: Classical and New Methods e 2009. PaleontologicalInstitute of the Russian Academy of Sciences, Moscow, pp. 67e82.

Danilov, I.G., Syromyatnikova, E.V., 2009b. Phylogeny of the extinct turtle cladeAdocusia. In: Abstract Volume of the Gaffney Turtle Symposium, pp. 44e51.Drumheller, Alberta.

Danilov, I.G., Sukhanov, V.B., Syromyatnikova, E.V., 2011. New Asiatic materials on tur-tlesof the familyAdocidaewitha reviewof theadocid record inAsia. Proceedingsofthe Zoological Institute of the Russian Academy of Sciences 315, 101e132.

Danilov, I.G., Syromyatnikova, E.V., Skutschas, P.P., Jianhua, J., Kodrul, T., 2013. Thefirst ‘true’ Adocus (Testudines: Adocidae) in the Paleogene of Asia. Journal ofVertebrate Paleontology, in press.

Eriksson, T., 1998. AutoDecay. 4.0.1. Department of Botany, Stockholm University,Stockholm.

Gilmore, C.W., 1919. Reptilian faunas of the Torrejon, Puerco, and underlying UpperCretaceous formations of San Juan County, New Mexico. U.S. Geological SurveyProfessional Paper 119, 1e75.

Gilmore, C.W., 1931. Fossil turtles of Mongolia. Bulletin of the American Museum ofNatural History 59, 213e257.

Hay, O.P., 1908. The Fossil turtles of North America. Carnegie Institute of Wash-ington Publication 75, 1e568.

Hay, O.P., 1910. Descriptions of eight new species of fossil turtles from west of theone hundredth meridian. Proceedings of the United States National Museum38, 307e326.

Hutchison, J.H., 2000. Diversity of Cretaceous turtle faunas of Eastern Asia and theircontribution to the turtle faunas of North America. Paleontological Society ofKorea Special Publication 4, 27e38.

Joyce, W.G., 2007. Phylogenetic relationships of Mesozoic turtles. Bulletin of thePeabody Museum of Natural History 48, 3e102.

Khosatzky, L.I., Nessov, L.A., 1977. Turtles of the genus Adocus from the LateCretaceous in the USSR. Trudy Zoologicheskogo Instituta AN SSSR 74, 116e119.

Lapparent de Broin, F. de, 2004. A new Shachemydinae (Chelonii, Cryptodira) fromthe Lower Cretaceous of Laos: preliminary data. Comptes Rendus Palevol 3,387e396.

Leidy, J., 1865. Cretaceous reptiles of the United States. Smithsonian Contributionsto Knowledge 192, 1e135.

Marsh, O.C., 1890. Notice of some extinct Testudinata. American Journal of Scienceand Arts, Series 3 40, 177e179.

Meylan, P.A., Gaffney, E.S., 1989. The skeletal morphology of the Cretaceous cryp-todiran turtle, Adocus, and the relationships of the Trionychoidea. AmericanMuseum Novitates 2941, 1e60.

Narmandakh, P., 1985. A new species of Adocus from the Late Cretaceous ofMongolia. Paleontologicheskiy Zhurnal 2, 85e93.

Nessov, L.A., 1977. Skull morphology of Early Cretaceous turtles of the familyAdocidae. Trudy Zoologicheskogo Instituta AN SSSR 74, 75e80.

Nessov, L.A., Khosatzky, L.I., 1977. Freshwater turtle from the Early Cretaceous ofFergana. Ezhegodnik Vsesoyuznogo Paleontologicheskogo Obshchestva 20,248e262.

Nessov, L.A., Krasovskaya, T.B., 1984. Changes in the composition of turtles assem-blages of Late Cretaceous of Middle Asia. Vestnik Leningradskogo Gosu-darstvennogo Universiteta 3, 15e25.

Page, R.D.M., 2001. NDE: NEXUS Data Editor 0.5.0. University of Glasgow, Glasgow.Sterli, J., 2010. Phylogenetic relationships among extinct and extant turtles: the

position of Pleurodira and the effects of the fossils on rooting crown-groupturtles. Contributions to Zoology 79, 93e106.

Sukhanov, V.B., 2000. Mesozoic turtles of Middle and Central Asia. In:Benton, M.J., Shishkin, M.A., Unwin, D.M., Kurochkin, E.N. (Eds.), The Age ofDinosaurs in Russia and Mongolia. Cambridge University Press, Cambridge,pp. 309e367.

Sukhanov, V.B., Narmandakh, P., 2006. New taxa of Mesozoic turtles from Mongolia.In: Danilov, I.G., Parham, J.F. (Eds.), Fossil Turtle Research, Vol. 1. ZoologicalInstitute of the Russian Academy of Sciences, St. Petersburg, pp. 119e127.

Syromyatnikova, E.V., 2011. Turtles of the genus Ferganemys Nessov et Khosatzky,1977 (Adocidae): shell morphology and phylogenetic position. Proceedings ofthe Zoological Institute of the Russian Academy of Sciences 315, 38e52.

Syromyatnikova, E.V., Danilov, I.G., 2009. New material and a revision of turtles ofthe genus Adocus (Adocidae) from the Late Cretaceous of Middle Asia andKazakhstan. Proceedings of the Zoological Institute of the Russian Academy ofSciences 313, 74e94.

Syromyatnikova, E.V., Danilov, I.G., Sukhanov, V.B., 2009. Morphology and system-atic position of Adocoides amtgai – turtles of the family Adocidae from LateCretaceous of Mongolia. In: International conference to 40th anniversary of theJoint Russian-Mongolian paleontological expedition, Moscow, Russia, pp. 77e79.

Syromyatnikova, E.V., Danilov, I.G., 2013. New material and phylogenetic position ofAdocus bostobensis, a poorly known adocid turtle from the Late Cretaceous ofKazakhstan. Proceedings of the Zoological Institute of the Russian Academy ofSciences 317, 195e201.

Syromyatnikova, E.V., Danilov, I.G., Sukhanov, V.B., 2011. The morphology andphylogenetic position of Adocus/Adocoides amtgai, an adocid turtle from thelate Cretaceous of Mongolia. Journal of Vertebrate Paleontology 31 (Supple-ment), 202.

Syromyatnikova, E.V., Danilov, I.G., Sukhanov, V.B., 2012. A redescription andphylogenetic position of Adocus planus, an adocid turtle from the Late Creta-ceous of Mongolia. Proceedings of the Zoological Institute of the RussianAcademy of Sciences 316, 380e391.

Tong, H., Claude, J., Naksri, W., Suteethorn, V., Buffetaut, E., 2009. Basilochelysmacrobios n. gen. and n. sp., a large cryptodiran turtle from the Phu KradungFormation (latest Jurassiceearliest Cretaceous) of the Khorat Plateau, NEThailand. Geological Society, London, Special Publications 315, 153e173.

White, R.S., 1972. A recently collected specimen of Adocus (Testudines; Dermate-mydidae) from New Jersey. Notulae Naturae 447, 1e10.

Yeh, H.K., 1963. Fossil turtles of China. Paleontologica Sinica, New Series C 18, 1e112.Yeh, H.K., 1966. A new Cretaceous turtle of Nanhsiung, Northern Kwangtung. Ver-

tebrata PalAsiatica 10, 191e200.

Appendices

1. Details about characters and taxa added to matrix of Danilov andSyromyatnikova (2009) with additions from Syromyatnikova (2011) andSyromyatnikova et al. (2012).

Character 78. Marginals overlapping onto the costals: 0 ¼ beginning with mar-ginal 5; 1 ¼ beginning with marginals 3 or 4.

Codings: Xinjiangchelys levensis, ?; X. tianshanensis, ?; Carettochelys insculpta,d;Apalone ferox, d; Yehguia tatsuensis, ?; Adocus aksary, 1; A. amtgai, 1; A. beatus/Adocus sp. 1, 0; A. bossi, 0; A. kirtlandius, 0; A. planus, ?; “Ferganemys” itemirensis, d;F. verzilini, d; Shachemys ancestralis, d; Sh. baibolatica, d; Sh. laosiana, 0; Khar-akhutulia kalandadzei, d; Zangerlia testudinimorpha, d; “Z”. neimongolensis, d; “Z”.ukhaachelys, d; Hanbogdemys orientalis, d; Anomalochelys angulata, d; Nanh-siungchelys wuchingensis, d; Basilemys gaffneyi, d; B. variolosa, d; B. sinuosa, d;B. praeclara, d.

2. Characters coded for Adocus amtgai and added to the matrix of Danilov andSyromyatnikova (2009a,b) with additions from Syromyatnikova (2011) andSyromyatnikova et al. (2012). Adocus amtgai: 00110000?0 ???????1?1 00010000100100000110 0000000010 0100001000 0010100011 10010111.