Abelisauridae (Dinosauria: Theropoda) from the Late Jurassic of Portugal and dentition-based phylogeny as a contribution for the identification of isolated theropod teeth

Citation:
Hendrickx, C., & Mateus O. (2014).  Abelisauridae (Dinosauria: Theropoda) from the Late Jurassic of Portugal and dentition-based phylogeny as a contribution for the identification of isolated theropod teeth. Zootaxa. 3759, 1-74.

Abstract:

Theropod dinosaurs form a highly diversified clade, and their teeth are some of the most common components of the Mesozoic dinosaur fossil record. This is the case in the Lourinhã Formation (Late Jurassic, Kimmeridgian-Tithonian) of Portugal, where theropod teeth are particularly abundant and diverse. Four isolated theropod teeth are here described and identified based on morphometric and anatomical data. They are included in a cladistic analysis performed on a data matrix of 141 dentition-based characters coded in 60 taxa, as well as a supermatrix combining our dataset with six recent datamatrices based on the whole theropod skeleton. The consensus tree resulting from the dentition-based data matrix reveals that theropod teeth provide reliable data for identification at approximately family level. Therefore, phylogenetic methods will help identifying theropod teeth with more confidence in the future. Although dental characters do not reliably indicate relationships among higher clades of theropods, they demonstrate interesting patterns of homoplasy suggesting dietary convergence in (1) alvarezsauroids, therizinosaurs and troodontids; (2) coelophysoids and spinosaurids; (3) compsognathids and dromaeosaurids; and (4) ceratosaurids, allosauroids and megalosaurids.

Based on morphometric and cladistic analyses, the biggest tooth from Lourinhã is referred to a mesial crown of the megalosaurid Torvosaurus tanneri, due to the elliptical cross section of the crown base, the large size and elongation of the crown, medially positioned mesial and distal carinae, and the coarse denticles. The smallest tooth is identified as Richardoestesia, and as a close relative of R. gilmorei based on the weak constriction between crown and root, the “eight-shaped” outline of the base crown and, on the distal carina, the average of ten symmetrically rounded denticles per mm, as well as a subequal number of denticles basally and at mid-crown. Finally, the two medium-sized teeth belong to the same taxon and exhibit pronounced interdenticular sulci between distal denticles, hooked distal denticles for one of them, an irregular enamel texture, and a straight distal margin, a combination of features only observed in abelisaurids. They provide the first record of Abelisauridae in the Jurassic of Laurasia and one of the oldest records of this clade in the world, suggesting a possible radiation of Abelisauridae in Europe well before the Upper Cretaceous.

Abelisauridae (Dinosauria: Theropoda) from the Late Jurassic of Portugal and dentition-based phylogeny as contribution for the identification of isolated theropod teeth

 

Christophe Hendrickx1,2 & Octávio Mateus1,2

 

1 Universidade Nova de Lisboa, CICEGe, Departamento de Ciências da Terra, Faculdade de Ciências e Tecnologia, Quinta da Torre 2829-516 Caparica, Portugal.

2 Museu da Lourinhã, Rua João Luis de Moura 9 2530-158 Lourinhã, Portugal.

E-mail: c.hendrickx@.unl.pt, omateus@fct.unl.pt

 

Abstract

 

Theropod dinosaurs form a highly diversified clade and their teeth are some of the most common component of the dinosaur fossils record in the Mesozoic. This is the case in the Lourinhã Formation (Late Jurassic, Kimmeridgian-Tithonian) of Portugal where theropod teeth are particularly abundant and diverse. Four isolated theropod teeth are here described and identified based on morphometrical and anatomical data. They were included in a cladistic analysis performed on a data matrix of 140 dentition-based characters coded in 50 taxa, as well as a supermatrix combining our dataset with six recent datamatrices based on the whole theropod skeleton. The consensus tree resulting from the dentition-based data matrix reveals some phylogenetic potential of theropod teeth and this method will help identifying theropod teeth with more confidence in the future.

Based on morphometrical and cladistic analyses, the biggest tooth from Lourinhã is referred to a mesial crown of the megalosaurid Torvosaurus tanneri due to the ovoid cross section of the crown base, the large size and elongation of the crown, medially positioned mesial and distal carinae and the coarse denticles. The smallest tooth is identified as belonging to Richardoestesia and a close relative of R. gilmorei based on a distal constriction between crown and root, the “8-shaped” outline of the base crown and the average of ten symmetrically rounded denticles per five millimetres on the distal carina. Finally, the two medium-sized teeth belong to the same taxon and exhibit pronounced interdenticular sulci between distal denticles, hooked distal denticles for one of them, an stipple enamel texture, and a straight distal margin, a combination of features only observed in abelisaurids. They provide the first record of Abelisauridae in the Jurassic of Laurasia and the second oldest record of this clade in the world, suggesting a possible radiation of Abelisauridae in Europe well before the Upper Cretaceous.

 

Key words: Abelisauridae, Megalosauridae, Torvosaurus, Richardoestesia, teeth, cladistic, Theropoda, Portugal, Late Jurassic

 

Introduction

 

The Upper Jurassic of Portugal has yielded an important fauna of dinosaurs, one of the richest of Europe. Dinosaur bones and teeth have been collected for more than 140 years, mainly from two important paleontological sites both situated in the centre of Portugal (Rauhut 2000; Antunes & Mateus 2003). The first one, the Guimarota Mine, is constituted by several layers of limestone, sandstone, mudstone, marl and coal belonging to the Alcobaça Formation dated of the Kimmeridgian stage (Helmdach 1971; Henkel & Krusat 1980; Schudack 2000; Kullberg et al. 2012) and the exploration campaign in the 1960s and new excavations in the mine from 1972 to 1982 allowed unearthing ornithischian and saurischian dinosaurs, mostly represented by isolated teeth (Krebs 2000; Rauhut 2000). The second, the Lourinhã region, is the richest area for dinosaur fossils in Portugal as bones, teeth, tracks, eggs and embryos of dinosaurs have been uncovered in several localities of the Lourinhã Formation Kimmeridgian-Tithonian in age (Antunes & Mateus 2003, Kullberg et al. 2012).

Most major clades of dinosaurs (ornithopods, thyreophorans, sauropodomorphs and theropods) are represented in the Upper Jurassic of Portugal but theropods are the most diversified group of dinosaurs (Rauhut 2000; Mateus 2006). Material from the Guimarota Mine and the Lourinhã region can be referred to at least 9 theropod taxa including Ceratosaurus dentisulcatus (Mateus & Antunes 2000; Mateus et al. 2006), Torvosaurus tanneri (Mateus et al. 2006), Allosaurus europaeus (Mateus et al. 2006), Allosaurus fragilis (Pérez-Moreno et al. 1999), Lourinhanosaurus autunesi (Mateus 1998), Aviatyrannis jurassica (Rauhut 2003), cf. Compsognathus sp. (Zinke 1998) and cf. Archaeopteryx sp. (Weigert 1995; Wiechmann & Gloy 2000) and theropods belonging to Carcharodontosauridae, Dromaeosauridae, Troodontidae and some uncertain systematic clades (cf. Richardoestesia sp. and cf. Paronychodon sp.) have also been reported (Zinke & Rauhut 1994; Zinke 1998; Mateus 2005). Moreover, theropod embryos, ascribed to Lourinhanosaurus (Ricqlès et al. 2001; Mateus et al. 1998; Hendrickx & Mateus 2012) and a megalosauroid (Araújo et al. 2012), were also collected in Portugal, as well as a diverse ichnological record (Mateus & Milàn 2010).

Theropod teeth are very common in the Lourinhã Formation and some of them have been reported in the literature already. In the 1950s, several theropod teeth found at Porto das Barcas (Lourinhã Formation) near Lourinhã were briefly described by Lapparent & Zbyszewski (1957). The material was collected by Carlos Ribeiro during a geologic cross section on June 20 1863 and those teeth seem to be the earliest dinosaur fossils found in Portugal so far (Antunes & Mateus 2003). Identified by Lapparent & Zbyszewski (1957) as belonging to the species Megalosaurus insignis and the new taxon Megalosaurus pombali, these two taxa are however currently considered as invalid (Antunes & Mateus 2003; Holtz et al. 2004). Later, Antunes (1990) mentioned the presence of a tooth fragment also attributed to the genus Megalosaurus. However, the first thorough study of theropod teeth from the area of Lourinhã was made by Rauhut & Kriwet (1994) who described two large theropod teeth also found in Porto das Barcas and which they attributed cautiously to an indeterminate “carnosaur”. Finally, Mateus (2005) and Mateus et al. (2006) mentioned and briefly described several theropod teeth from the Lourinhã Formation and recognized the presence of Ceratosaurus dentisulcatus and the clades of Carcharodontosauridae and Troodontidae in this unit.

Although theropod teeth are rather simple structure far less informative than mammals teeth (Longrich 2008; Han et al. 2011) or many other part of the skeleton such as the quadrate (Hendrickx et al. 2012), a number of workers have successfully used theropod tooth morphology for taxonomic purposes (e.g., Currie et al. 1990; Fiorillo & Currie 1994; Rauhut & Werner 1995; Baszio 1997; Zinke 1998; Fiorillo & Gangloff 2001; Rauhut 2002; Sankey et al. 2002; Fanti & Therrien 2007; Larson 2008; Longrich 2008; Brinkman 2008; Sankey 2008; Soto & Perea 2008; Larson et al. 2010; Ősi et al. 2010; Han et al. 2011). Tooth measurements were first utilized by Currie et al. (1990) and Farlow et al. (1991) for systematic identification of theropod teeth and different authors later followed or modified this method to document isolated theropod teeth (e.g., Fiorillo & Currie 1994; Baszio 1997; Holtz et al. 1998; Sankey 2001; Sankey et al. 2002; Bakker & Bir 2004; Sankey 2008; Larson 2008; Han et al. 2011). Smith (2005) and Smith et al. (2005) were the first to successfully discriminate theropod teeth to the genus level based on a quantitative methodology and discriminant analyses. Such methodology was later followed by Smith & Vechia (2006), Smith & Lamanna (2006), Van der Lubbe et al. (2009), Torres-Rodríguez et al. (2010) and Ősi et al. (2010) to identify isolated teeth of theropods, and used in a slightly different way by Fanti & Therrien (2007) and Larson (2008). The taxonomic utility of theropod teeth evaluated with cladistics tools has recently been investigated by Hwang (2007) who mostly focused on the enamel microstructure. Hwang (2007) performed a first cladistic analysis by using height dental and 31 enamel characters coded in 52 dinosaur taxa among which 25 theropods, and combined the enamel based characters to the dataset of Makovicky et al. (2005).

The present work aims to evaluate the systematic potential of theropod teeth as well as to investigate the systematic palaeontology of four isolated theropod teeth chosen in the collection of the Museu of Lourinhã based on their completeness, particular shape and interesting features displayed (e.g., interdenticular sulci, transversal and marginal undulations, mesio-distal constriction of the crown). While the taxonomic value of theropod teeth was assessed by following Hwang (2007) methodology, i.e., performing a cladistic analysis on a data matrix including dentition-based characters only, the taxonomic identification of the four teeth from Portugal was investigated by using Smith et al. (2005) methodology with morphometrical data, and Hwang (2007) methodology by incorporating the four isolated teeth into a phylogenetic analysis. This way, it is intended to give a contribution to the identification of isolated theropod teeth, which often fail to be referred to a particular clade of theropod with certainty (e.g., Maganuco et al. 2005; Ősi et al. 2010; Han et al. 2011).

 

Institutional abbreviations

 

AMNH                  American Museum of Natural History, New York, USA

ANSP                     Academy of Natural Sciences of Drexel University, Philadelphia, Pennsylvania, USA

BYUVP                 Brigham Young University Vertebrate Paleontology, Provo, Utah, USA

CM                        Carnegie Museum, Pittsburgh, Pennsylvania, USA

DMNH                  Denver Museum of Natural History, Denver, Colorado, USA

FMNH PR            Field Museum of Natural History, Chicago, Illinois, USA

IGM                       Institute of Geology, Ulaan Baatar, Mongolia

LH PV                   Long Hao Institute of Geology and Paleontology, Hohhot, Nei Mongol, China

MACN-CH           Museo Argentino de Ciencias Naturales “Bernardino Rivadavia”, Buenos Aires, Argentina

MCF-PVPH          Museo Municipal Carmen Funes, Paleontologia de Vertebrados, Plaza Huincul, Argentina

MG GUI                Museu Geológico (Guimarota collection), Lisbon, Portugal

MIWG                   Dinosaur Isle, Isle of Wight Museum Services, Sandown, United Kingdom

ML                         Museu da Lourinhã, Lourinhã, Portugal

MLP                      Museo de La Plata, La Plata, Argentina

MMCN-PV           Museo Municipal “Ernesto Bachmann”, Villa El Chocón, Neuquén, Argentina

MNHN                  Muséum national d’Histoire naturelle, Paris, France

MNN                     Musée National du Niger, Niamey, Niger

MPCA                   Museo Provincial Carlos Ameghino, Cipolletti, Río Negro, Argentina

MSNM                  Museo di Storia Naturale di Milano, Milan, Italy

MUCPv                 Museo de Ciencias Naturales de la Universidad Nacional de Comahue, Lago Barreales, Argentina

MUCPv-CH         Museo de Ciencias Naturales de la Universidad Nacional de Comahue, El Chocón collection, Villa El Chocón, Argentina

MWC                    Museum of Western Colorado, Fruita (or Grand Junction), Colorado, USA

NCSM                   North Carolina Museum of Natural Sciences, Raleigh, North Carolina, USA

NHM                     The Natural History Museum, London, United Kingdom

NMC                     Canadian Museum of Nature, Ottawa, Ontario, Canada

OUMNH               Oxford University Museum, Oxford, UK

PMAA                   Provincial Museum and Archives of Alberta, Paleontological Collections, Drumheller, Alberta, Canada

PVSJ                      Museo de Ciencias Naturales, Universidad Nacional de San Juan, San Juan, Argentina

SBA-SA Soprintendenza per i Beni Archeologici di Salerno Avellino Benevento e Caserta, Salerno, Italy

SMA                      Sauriermuseum Aathal, Aathal, Switzerland

SMNS                    Staatliches Museum für Naturkunde, Stuttgart, Germany

SMU                      Southern Methodist, University, Dallas, Texas, USA

TMP                      Royal Tyrrell Museum of Palaeontology, Drumheller, Alberta, Canada

UA                         Université d’Antananarivo, Antananarivo, Madagascar

UC                         University of Chicago Paleontological Collection, Chicago, USA

UCMP                   University of California Museum of Paleontology, Berkeley, California, USA

UCPC                    University of Chicago Paleontological Collection, Chicago, USA

UMNH VP            Utah Museum of Natural History, University of Utah, Salt Lake City, Utah, USA

USNM VP             United State National Museum Vertebrate Paleontology, Washington, District of Columbia, USA

USNM                   United State National Museum Vertebrate Paleontology, Washington, District of Columbia, USA

SGM                      Ministère de l’Énergie et des Mines, Rabat, Morocco

 

 

Locality, geological and stratigraphical setting

 

The four teeth all come from the Lourinhã Formation and have been found near the town of Lourinhã. The Lourinhã Formation is 600 to 1100 meters thick and mostly appears along the cliffs bounding the Atlantic Ocean, 70 kilometres North of Lisbon. The formation is delimited at its base by the Amaral Formation of Kimmeridgian in age and defined by shallow marine sandstones and oolites, as well as a shallow marine carbonate shelf forming the limit with the Lourinhã Formation. The Cretaceous continental clastic Torres Vedras Formation (or Group) which lies uncoformably on the Lourinhã Formation forms its upper limit.

Lithologically, the Lourinhã Formation consists of continental deposits intercalated with some shallow marine deposits corresponding to an alluvial fan and fluvio-deltaic environments punctuated by periodic marine transgressions (Kullberg et al. 2012; Hill 1988, 1989). Theropod teeth can be found in both Porto Novo and Santa Rita members of the Lourinhã Formation. For more information regarding the sedimentology on those two members, see Hill (1988, 1989).

Several authors (e.g., Manuppella 1996, 1998; Manuppella et al. 1999) have considered the Alcobaça Formation as being similar to the Lourinhã Formation. However the latter is dated Upper Kimmeridgian-Tithonian in age and is therefore slightly younger and also more continental than the Alcobaça Formation (Mateus 2006). Nevertheless, both the Lourinhã and Alcobaça Formation of Portugal are comparable with the Morrison Formation of Northern America and the Tendaguru Beds in Tanzania as the three regions are Kimmeridgian-Tithonian in age and show similar ecosystems, all dominated by dinosaurs (Mateus 2006).

 

Methodology

 

Quantitative methodology

 

The description of teeth follows the dental nomenclature proposed by Smith and Dodson (2003). Both descriptive morphological characters and quantitative morphometric techniques were used to analyse and identify the four theropod teeth. Observations were made with a binocular microscope Leica MZ6 as well as a digital microscope AM411T-Dino-Lite Pro. Photographs were taken with a digital camera for the biggest teeth and the digital microscope for the smaller tooth.

The quantitative methodology was based on numerical data developed by Smith (2005) and Smith et al. (2005), and updated by Smith and Lamanna (2006), Smith and Dalla Vecchia (2006) and Smith (2007). Additional morphometrical data of theropod teeth were collected from Canudo et al. (2006), Soto & Perea (2008), Sereno & Brusatte (2008), Molnar et al. (2009), Van der Lubbe et al. (2009), Torres-Rodríguez et al. (2010), Ősi et al. (2010) and Gianechini et al. (2011b). Morphometrical measurements were also taken on many theropod teeth belonging to the palaeontological collections of 24 museums from Argentina, Europe and United States (see Table 1). Anatomical and morphometric abbreviations follow Smith et al. (2005) except abbreviations related to the number of denticles which have been modified for more coherence.

Measurements and ratios proposed by Smith et al. (2005) were taken with a digital calliper and the following measurements were done: AL, apical length (in millimetres); CBL, crown base length, measured at the base of the crown from its mesialmost to its distalmost extension (excluding the carinae; in millimetres); CBR, crown base ratio, numerical value derived from dividing CBW through CBL (= labiolingual compression); CBW, crown base width, labiolingual extension of the crown at its base (in millimetres); CH, crown height, measured from the basal-distal most point of the crown toward its tip (in millimetres); CHR, crown height ratio, numerical value derived from dividing CH through CBL; DAVG, average distal denticle density on 5 millimetres; DDda, denticle density for distal apical serration, i.e., denticles per 5 mm at the most apical part of the distal carinae; DDdb, denticle density for distal basal serration, i.e., denticles per 5 mm at the most basal part of the distal carinae; DDdm, denticle density for distal mid-crown serration, i.e., denticles per 5 mm at the mid-crown part of the distal carinae; DSDI, denticle size difference index, ratio between the number of denticles per 5 mm of the anterior and posterior carinae (if MAVG OR DAVG ≥ 1 then DSDI = (MAVG + 1)/(DAVG + 1), otherwise DSDI = 0);  DDma, denticle density for mesial apical serration, i.e., denticles per 5 mm at the most apical part of the mesial carinae; DDmb, denticle density for mesial basal serration, i.e., denticles per 5 mm at the most basal part of the mesial carinae; DDmm, denticle density for mesial mid-crown serration, i.e., denticles per 5 mm at the mid-crown part of the mesial carinae; MAVG, average mesial denticle density on 5 millimetres.

 

Cladistic analysis

 

A data matrix of dentition-based characters (see appendix) was created and scored in 50 nonavian theropod taxa in order to evaluate the taxonomic potential of theropod dentitions and assess the phylogenetical relationship of the four teeth from the Lourinhã Formation. Teeth pertaining to most clades of nonavian theropods were examined and coded from first-hand observations and high-resolution photographs in 47 specimens (94%), and by using full descriptions and illustrations of the teeth in the literature in three taxa (Dilophosaurus, Scipionyx and Richardoestesia; see Table 1). Eoraptor lunensis, considered to be either a basal saurischian (Langer & Benton 2007), a basal sauropodomorph (Martinez et al. 2011) or a basal theropod (ex. Nesbitt et al. 2009; Nesbitt 2011; Sues et al. 2011), was treated as the outgroup.

The data matrix encompasses 140 equally weighted morphological characters based on the morphology of the crown, root, mesial and distal carinae, denticles, interdenticular sulci (‘blood groove’ sensu Currie et al. 1990, and ‘caudae’ sensu Abler 1992), crown surface (which can display transversal undulations, short undulations adjacent to carinae, flutes, and other longitudinal grooves), and enamel texture and microstructure. Characters related to the shape, size and number of teeth/alveoli of the premaxilla, maxilla and dentary were also included in the data matrix (see appendix). Only 74 characters are derived from the literature and 66 characters (47%) were revealed by descriptive work on the teeth and our personal observation. Due to the important variation of morphology between premaxillary/mesial dentary teeth and lateral teeth of the maxilla and dentary, the dataset was divided into most mesial and lateral teeth (see Appendix). Among the 140 morphological characters, ten are continuous characters and three are meristic and concerns the number of premaxillary, maxillary and dentary teeth. Both continuous and meristic characters were treated differently following two methodologies. In a first method, all 13 continuous and meristic characters were transformed into discrete characters of no more than five character states by assigning a specific range or value (see appendix), and only two multistate characters, both concerning the crown size (char. 36 and 65, see appendix), were ordered. In a second methodology, the actual number of premaxillary, maxillary and dentary teeth was scored in the data matrix, and all continuous characters were treated as such (see Appendix).

The taxonomic potential of theropod teeth was first evaluated by performing a cladistic analysis on the data matrix of dentition based characters without the isolated teeth from Portugal. In order to constrain all major theropod clades and visualize the dentition-based synapomorphies for each theropod clades, a second analysis was performed on a supermatrix combining our dentition-based data matrix to six recent datasets on non-avian theropods based on the whole skeleton (Xu et al. 2009; Brusatte et al. 2010; Martinez et al. 2011; Senter 2011; Pol & Rauhut 2012; Carrano et al. 2012), and from which all teeth-related characters were removed. The resulting supermatrix includes 1971 characters with 56 treated as ordered (see appendix) in the first methodology, and ten continuous and 54 discrete ordered characters in the second one. The four isolated teeth from the Lourinhã Formation were then incorporated in the dentition based dataset and the supermatrix in order to assess their phylogenetic relationship.

TNT v1.1 (Goloboff et al. 2008) was employed to search for most-parsimonious trees (MPTs). The matrix and supermatrix were analysed under the ‘New Technology Search’ with the ‘driven search’ option (TreeDrift, Tree Fusing, Ratchet, and Sectorial Searches selected with default parameters), and stabilizing the consensus twice with a factor of 75. The consistency and retention indexes as well as the Bremer supports (Bremer 1994) were calculated using the “stats” and “aquickie” commands, respectively, and a bootstrap analysis was performed on the resulted consensus trees with the standard options.

 

Phylogenetic usefulness of nonavian theropod teeth

 

Figure 1

 

The resulting consensus trees recovered from the cladistic analysis performed on the dentition based data matrix (the isolated teeth ML 327, 939, 962, 966 not included) gave two different topologies when using discrete characters only or a combination of discrete and continuous characters (see Fig. 1 and appendix). In the first method using discrete characters only, analysis of the data matrix of 50 taxa yielded 10 most parsimonious trees (MPTs), in which the strict consensus trees (Length = 571 steps, CI = 0.366, RI = 0.503) resulted in many polytomies. Nevertheless, this lack of resolution was due to the instability of the ‘wildcard’ taxon Erectopus superbus (Allain 2005) which has equally parsimonious positions within distant clades. A reduced consensus approach (Wilkinson 1994) was therefore used by excluding this taxon and yielded four MPTs in which the resulting consensus tree (Length = 565 steps; CI = 0.365, RI = 0.558) displays only two polytomies (Fig. 1). Although the strict consensus trees did not recover the general topology of theropod classification with the usual major clades (Ceratosauria, Tetanurae, Avetheropoda, and  Coelurosauria), many theropod clades were found resolved, demonstrating some taxonomic potential of theropod teeth at least at a family level. As noted by Hwang (2007), dentition-based characters can be good at recovering individual clades, but not at resolving the relationship between those clades. This is obviously due to the large amount of convergence in theropod dentition, directly linked to diet, among a clade displaying the most important variation of feeding strategies among dinosaurs (Rayfield 2005; Therrien et al. 2005; Zanno et al. 2010).

Similarly to Hwang (2007: fig. 38), basalmost theropods and derived toothed coelurosaurs are closely related. They seem indeed to have similar dentitions in both their morphology and microstructure, and might therefore have shared similar diet. Morphologically, both non-neotheropods and deinonychosaurs possess small ziphodont (i.e., labio-lingually compressed crown in which both mesial and distal margins possess serrated carinae) and/or lanceolate (i.e., labio-lingually compressed crown with a mesio-distal constriction at the base crown) lateral teeth with large sometimes apically inclined distal denticles and, when present, smaller mesial serrations. The most mesial teeth of these theropods also tend to lack mesial and sometimes distal serrations. Likewise, derived Megalosauridae (Torvosaurus, Megalosaurus and Afrovenator) and some Allosauroidea (Allosaurus, Neovenator, Giganotosaurus, Mapusaurus and Carcharodontosaurus) are nested in a same clade (Fig. 1), witnessing similar feeding strategies among these large tetanurans possessing a robust skull. Their lateral teeth are indeed sometimes very similar as both clades possess large elongated and labio-lingually narrow teeth with strongly displaced distal carina bearing coarse symmetrical to asymmetrical chisel-like denticles with often deep and elongated interdenticular sulci. The lateral teeth of megalosaurids and allosauroids can also display same crown structures: large transversal undulations as well as short, pronounced mesio-distally oriented undulations adjacent to carinae. Tyrannosaurid teeth also display similar structures on the crown but their lateral teeth are much stouter and the mesial carina corresponds to a rather low ridge compared to the protuberant mesial carina of Megalosauridae and Allosauroidea (pers. observ.).

                Due to the variability of their dentition, Dromaeosauridae, Megalosauridae and Carcharodontosauridae were recovered polyphyletic in this analysis. The lateral teeth of the dromaeosaurids Buitreraptor, Dromaeosaurus and Saurornitholestes are indeed quite distinct. Regarding their serration, for instance, the first one lacks mesial and distal denticles whereas Dromaeosaurus bears subquadrangular denticles with a convex margin and Saurornitholestes possesses large and apically hooked denticles, reminding those of Troodontidae. The differences among megalosaurid and carcharodontosaurid dentition are much subtle. In Megalosauridae, Eustreptospondylus and Dubreuillosaurus teeth are smaller (with smaller denticles) than those of Afrovenator, Megalosaurus and Torvosaurus and the crowns do not display any structures such as well-developed interdenticular sulci, short undulations marginal to carinae and well-visible transversal undulations, unlike the crowns of large megalosaurids. In Carcharodontosauridae,  

On the other hand, with to their characteristic and distinctive dentition, ceratosaurids, abelisaurids, spinosaurids, tyrannosauroids and compsognathids clades were found monophyletic in this analysis. Ceratosaurid teeth are characterized by much larger mid-maxillary teeth when compared to anterior maxillary teeth and, in lateral teeth, a wide mesiodistally concave area on the basal part of the labial margin of the crown, a distal carina extending to the cervix and a broad interdenticular space in between distal denticles. Their lateral teeth also have a braided texture of the enamel, and the crown tend to be strongly labiolingually compressed, sometimes bearing well-visible transversal undulations. In addition, the labial surface adjacent to the distal and sometimes mesial carinae is often flat or concave in ceratosaurid lateral teeth (Rauhut, 2004; pers. observ.).

Likewise, abelisaurid teeth are weakly recurved distally, so that their distal profile is either straight or even convex in lateral view, and many abelisaurids tend to have low crowns (Smith 2007; pers. observ.), although elongated crowns can be bore by some abelisaurid taxa such as Aucasaurus garridoi (MCF-PVPH 236) and Skorpiovenator bustingorryi (MMCN-PV 48). Like noasaurid teeth, the enamel surface texture of abelisaurid crown is irregular (here referred to ‘stipple texture’) with no orientation of the enamel texture, unlike the braided or veined texture of many neotheropod enamel (pers. observ.). Abelisaurid lateral teeth also possess mesial and distal carinae centrally positioned on the crown in mesial and distal views, respectively. Both keels always reach the neck of the crown and the denticles often show deep interdenticular sulcus and/or apex pointing apically. Their most mesial teeth are also typical as they have a concave surface adjacent to the mesial carina and sometimes marginal to the distal carina in the most mesial tooth (Fanti & Therrien 2007; Smith 2007; pers. observ.).

Due to their highly specialized skull and dentition to piscivory (e.g., Charig & Milner 1997; Sereno et al. 1998; Sue et al. 2002; Dal Sasso et al. 2005; Hendrickx & Buffetaut 2008; Dal Sasso et al. 2009), several features of spinosaurid teeth have already been used as synapomorphies by many authors (e.g., Sereno et al. 1998; Holtz et al. 2004; Benson 2009; Mateus et al. 2011; Carrano et al. 2012). Their teeth are indeed highly diagnostic as all spinosaurids possess subcircular most mesial and lateral crowns displaying flutes on the lingual and/or labial margin, minute denticles or no serrations at all on both mesial and distal carinae, and the enamel texture of Spinosaurus and baryonychines are deeply veined (or sculptured sensu Hasegawa et al. 2010) and curves towards the base close to the keels. Their spatulated premaxillae bear a minimum of six teeth in which the posterior premaxillary teeth are significantly smaller than the anterior ones, and the premaxillary tooth row extends anterior to the external naris. Moreover, the dentaries also form a terminal rosette where the anteriormost teeth are significantly larger than mid- and posterior dentary teeth.

Tyrannosauroidea dentition is characterized by the following features: the third and fourth premaxillary teeth are distinctively overlapping, the mid-maxillary teeth are larger than the anterior maxillary teeth, and the crowns display a braided enamel texture, on the contrary of most other coelurosaurs. Moreover, the basal cross-section of the most mesial crown is D-shaped, the distal margin faces lingually in most mesial teeth, and the distal denticles are larger than the mesial ones in lateral teeth (DSDI < 1.2). In Compsognathidae, on the other hand, the mesial and distal carinae of most mesial teeth are absent or unserrated, and the lateral teeth do not bear mesial denticles, whereas the distal denticles disappear well beneath the apex of the crown. Absence of mesial denticles in both most mesial and lateral teeth and unserrated most mesial teeth (in the first two most mesial teeth at least) seem indeed to be a condition shared by all compsognathids but Sinocalliopteryx gigas (Ji et al. 2007; Chiappe & Göhlich 2011; Dal Sasso & Maganuco 2011).

Four most parsimonious trees were recovered from the cladistic analysis combining discrete and continuous dentition based characters and the resulting consensus tree (Length = 811.445 steps; CI = 0.470; RI = 0.682) includes two polytomies only. Due to the fact that continuous characters are treated as ordered characters, the general topology significantly differs from the consensus tree obtained when using discrete characters only. The Spinosauridae, Noasauridae and Abelisauridae are the only clades gathered in a same monophyletic group, and all Ceratosauridae, Megalosauridae, Allosauroidea, Tyrannosauroidea, Compsognathidae and Deinonychosauria were recovered paraphyletic. Basal Carcharodontosauridae and the allosauroids Allosaurus and Neovenator are however closely related but, due to the variability of their crown size, members of the clade of Tyrannosauroidea and Deinonychosauria occupy very disparate position on the tree.

The cladistic analysis performed on the supermatrix of discrete characters, and constraining all theropod clades, has shown that only a few unambiguous dentition-based synapomorphies can define theropods clades (see appendix). The large majority of dentition-based characters are indeed homoplastic, witnessing of a high degree of convergence among theropod dentition. Several clades of theropods such as Ceratosauridae, Abelisauroidea, Abelisauridae and Coelurosaurs are characterized by a combination of ambiguous synapomorphies and only Spinosauridae, Tyrannosauroidea and Compsognathidae are defined by both ambiguous and unambiguous synapomorphical characters. With five unambiguous synapomorphies, the clade of Spinosauridae is the best supported in term of dental characters, and both Tyrannosauroidea and Compsognathidae only include two unambiguous synapomorphies.

 

 

Systematic Palaeontology

 

Dinosauria Owen, 1842

 

Saurischia Seeley, 1887

 

Theropoda  Marsh, 1881

 

Ceratosauria Marsh, 1884

 

Abelisauroidea Bonaparte, 1991

 

Abelisauridae Bonaparte & Novas, 1985

 

Gen. and sp. indet.

 

Figures 2 & 3

 

Referred material. ML 327 and ML 966 (Figs 2 & 3).

Type locality and horizon. Cliffs of Lourinhã, Lourinhã, Portugal. Lourinhã Formation, Kimmeridgian-Tithonian, Upper Jurassic.

Description. ML 327 lacks the lowermost part of the crown, a small piece of the mesial carina on the lingual face and a few denticles on the distal keel. However, the crown is well preserved and most of the denticles are intact. The apical part of the distal carina of ML 966 is also missing; otherwise this tooth is relatively well-preserved, with some part of the enamel cracked and missing.

Crown. The teeth are slightly elongated baso-apically (CHR of 1.58 in ML 327 and 1.95 in ML 966) and typically ziphodont in shape, and both crowns are only weakly curved distally and the apex has been worn.

In lateral view, the distal carina is slightly concave, almost straight. The axis passing through the basal part of distal carina is perpendicular to the transversal plane of the crown. The mesial margin of the crown is much more recurved than the distal margin and the curvature is more important apically than basally. The apex is not acute and pointed but slightly rounded. In ML 327, it shows a small subrectangular wear facet on the labial face and a much bigger one forming an elongated tong-shape surface inclined mesio-basally on the two-third surface of the lingual face. The wear facet on the lingual side of the crown in ML 966 is rather subtriangular and only limited to the apex.  Both mesial and distal carinae are serrated from the base to the tip of the crown. The lingual surface of ML 327 bears an important longitudinal depression on its mesial part, 4 mm besides the mesial keel at the mid-crown. This narrow groove of 1.5 millimetres width extends around 8.5 mm above the cervix dentis (or neck of the tooth, here referred to cervix as Smith & Dodson 2003) and ends at a distance of 8 mm from the apex. The longitudinal depression roughly follows the curvature of the crown, is closer to the mesial carina at its basal and apical endings and almost contacts the large wear facet apically. No longitudinal groove is present on the labial face of the crown in ML 327 and on both labial and lingual sides of the tooth in ML 966.

In mesial view, the carina of both teeth is arced and inclined baso-lingually. The carina remains medially positioned on the tip of the crown but twists lingually towards the root more basally and extends mesio-lingually to the cervix. The crown apex remains straight and follows the general curvature of the crown. The lingual surface is slightly baso-apically sigmoid with the basal part of the crown concave and the apical one convex. On the other hand, the entire labial surface of the crown is strongly convex baso-apically.

In distal view, the distal carina is weakly sigmoid with a large bow oriented lingually along the basal two-third of the crown while the apical part of the distal carina is straight. The keel is slightly lingually positioned on the distal margin of the crown but moves medially at the tip.

In apical view, the tip of both crowns is distally positioned, with no curvature on the lingual or labial sides. The labial margin is globally convex but the distal surface is rather flattened or weakly convex. On the contrary, the surface adjacent to the distal carina on the lingual margin is rather slightly concave. In ML 327, the mesial part of the labial face is strongly convex whereas the mesial part of the lingual surface has a double curvature due to the presence of the longitudinal depression. In both teeth, the distal carina is angular whereas the mesial keel forms a low but pointed ridge which strongly displaces lingually towards the root. There is a flattened surface at the base of the mesial margin which is delimited lingually by the mesial carina in ML 327. This flattened surface which appears above the cervix extends on the one thirds of the mesial margin in this tooth. In ML 966 however, the surface at the base of the mesial margin is strongly convex.

In basal view, the cross-section of the crown base is elliptical and slightly piriform in ML 327 whereas ML 966 has a well-marked piriform outline of the crown base. In ML 327, the mesial part is roughly triangular in shape with the tip of the triangle pointed lingua-distally whereas the mesial part of ML 966 is strongly subtriangular with the tip of the triangle medially positioned. In both crowns, the distal margin of the crown forms a semicircle. The distal margin bears the superficial ridge of the distal keel which is mesio-lingually positioned. The labio-lingual width of the base crown is bigger mesially (CBW of 10.69 in ML 327 and 12.94 in ML 966). With their rather flattened bases, the middles of the lingual and labial faces are almost parallel. The middle of the labial surface remains roughly flat towards the tip while the lingual surface becomes strongly convex more apically. In ML 327, the dentine layer is thin (1 mm in the labial margin) and becomes thicker in the distal part of the crown (1.9 mm). Although the lingual margin has been damaged in this tooth, the pulp cavity seems to share the same drop shape than the crown-base but there is a weak labio-lingual constriction of the cavity 8 mm below the extremity of the distal carina.

Denticles. The mesial carina of ML 327 bears 11 denticles per 5 mm at the tip 13 at the mid-crown and 20 near the cervix. In ML 966, the mesial carina shows 19 denticles at the base and 15 at mid-crown, the mesioapical denticles having been worn. In both crowns, the denticles decrease in size towards the root at two-third of the crown and the most basal denticles are minute. In lateral view, the mesial denticles are longer baso-apically than mesio-distally, which give them a subrectangular (or “cartouche-like” sensu Harris 1998) outline. Since the denticles are inclined towards the tip of the crown and the main axis of the denticle is not perpendicular to the mesial margin of the crown, the shape of the denticle is rather plane-parallel. The external margin of the mesial denticles is rounded and sometimes asymmetrically convex, with the concavity positioned slightly apically. In both teeth, the lingual and labial surfaces of the denticles are convex and the interdenticular space is shallow. In mesial view, the denticles are not labio-lingually large, they are roughly chisel-like in shape but their external margin is rounded, and the main body of the denticles is almost cylindrical. There is no interdenticular sulcus between the mesial denticles in both teeth.

The distal carina of ML 327 counts 12 denticles per 5 mm at the apex 13 at the mid-crown and around 15 at the crown base (but not near the cervix, this part being missing) so that they are similar in size to mesial denticles (DSDI of 1.14). In ML 966 19 12 and 14 denticles per 5 mm can be observed at the base, mid-crown and apex of the crown respectively and this tooth also share a DSDI close to one. Unlike the mesial denticles, the distal denticles of both crowns are longer mesio-distally than baso-apically, except in the apical denticles which are squared-like in shape, and the main axis of the denticles is perpendicular to the distal margin. In lateral view, some distal denticles of ML 327 show an external margin pointing slightly towards the tip of the crown (Fig. 2E), so that the apical margin of the denticles is weakly concave whereas the basodistal margin is strongly convex. In all other distal denticles of ML 327 and all distal denticles of ML 966, the external margin is asymmetrically convex, with the denticle apex slightly apically positioned. In both teeth, the labial and lingual surfaces of the denticle body are convex. The distal denticles also have a deeper interdenticular space than the mesial ones and their external margin is more acute, giving them a real chisel-like shape in distal view. In ML 327, the enamel layer is thicker than in the mesial denticles and most of denticles show an elongated interdenticular sulcus diagonally oriented basally in both teeth. These shallow grooves parallel to each other extend from the base of the interdenticular space and runs on both labial and lingual faces of the crown. They are shorter in the apical denticles, and completely absent in the most apical one, both on the labial and lingual surface. Their inclination also tends to be reduced towards the root with interdenticular sulci being almost perpendicular to the distal margin in the basal denticles.

Surface. The enamel surface of both crowns is very well preserved and shows perfectly a granular and irregular texture on both lingual and labial faces. Besides the large longitudinal depression present on the lingual face, transversal and shallow undulations are present on both lingual and distal surfaces in ML 327. On the labial face of this tooth, they form large parabolic furrows curving apically near the distal carina, disappearing on the mesial part of the labial face due to the strong curvature of the crown. On the lingual face of this crown, they are visible distally, near the distal carina, and also in the middle of the crown, in the mid-crown surface. The undulations are absent on both convex surface adjacent to the mesial carina and the longitudinal depression. Unlike the labial wrinkling, these undulations do not bent towards the tip of the crown near the keel. In ML 966, the transversal undulations are also clearly visible on both sides of the crown. They are particularly pronounced close to the distal carina on the labial margin where they also curved apically adjacent to the distal carina (Fig. 3H). As in ML 327, the transversal undulations are large, parabolic and shallow on the lingual side of the crown and they do not curved toward the apex close to the carinae. In both teeth, these undulations are parallel and irregularly spaced and there are approximately 3 to 4 wrinkles per 5 mm on both faces of those crowns.

Discussion. Since the root is absent, ML 327 and ML 966 are most likely shed teeth. The labio-lingually compression of those moderately large teeth (CH > 30 mm), associated with serrated mesial and distal carinae and curvature of the tip distally is a plesiomorphic condition seen in theropod dinosaurs. Amongst known large terrestrial Jurassic groups of vertebrates, this combination of characters is only seen in theropods.

Although ML 966 is slightly bigger than ML 327, both teeth can confidently be associated to the same taxon  as they share same outline, CBR, DSDI, and the following features: presence of well-developed interdenticular sulci pointing basally, transversal undulations on both labial and lingual faces, a mesial carina offset, strongly twisted lingually towards the root and reaching the cervix, a distal carina slightly sigmoid and lingually positioned, a lingual face baso-apically concave and a labial surface baso-apically sigmoid, and a tear-drop outline of the base-crown in cross-section. Nevertheless, some denticles of ML 327 differ from ML 966 as their external margin are pointing apically and are not asymmetrically convex on their entire distal margins. However, denticle recurvature can vary in tooth row (Fanti & Therrien 2007; see below). ML 966 is also slightly more elongated than ML 327 (CHR of 1.95 and 1.58) but elongation of the crown also varies importantly along the tooth crown in theropods (e.g., Ceratosaurus, Allosaurus, Proceratosaurus, Tyrannosaurus).

One of the most striking features in those two isolated teeth is the presence of tenuous to well-marked transversal undulations (‘enamel wrinckles’ sensu Brusatte et al. 2007) on the crown. Though to be a possible tetanuran synapomorphy (Brusatte et al. 2007), transversal undulations are present on the crown of many theropods, from basal to derived forms, as well as metriorhynchid crocodylomorphs (Andrade et al. 2010) and rauisuchian crurotarsans (Brusatte et al. 2009b), and this feature cannot therefore be considered as a reliable tool alone for identifying teeth. In theropod, they have indeed been observed in basalmost theropods such as Sanjuansaurus gordilloi (PVSJ 605) and Eodromaeus murphi (PVSJ 561), ceratosaurs such as Ceratosaurus nasicornis (USNM VP 4735), Berberosaurus liassicus (MNHN Pt369), Genyodectes serus (MLP 26-39), Aucasaurus garridoi (MCF-PVPH 236) and Majungasaurus crenatissimus (FMNH PR 2278), many basal tetanurans (see Brusatte et al. 2007), and some deinonychosaurs like Troodon formosus (DMNH 22337) and Dromaeosaurus albertensis (AMNH 5356).

ML 966 also displays pronounced undulations adjacent to the distal carina. Short and marginal undulations close to carinae are a well-known feature of carcharodontosaurids teeth (Sereno et al. 1996; Coria & Currie 2006) as they appear on the teeth of Carcharodontosaurus saharicus (SGM Din-1; UC PV6), Mapusaurus roseae (MCF-PVPH 108) and Giganotosaurus carolinii (MUCPv-CH-1). However, marginal undulations have also been reported among non-carcharodontosaurid theropods such as the abelisaurid Skorpiovenator bustingorryi (Canale et al. 2009). They actually seem to be present in a large range of non-coelurosaur avetheropods as they have also been noticed in other ceratosaurs such as Ceratosaurus nasicornis (USNM 4735) and Majungasaurus crenatissimus (FMNH 2100), megalosaurids like Afrovenator abakensis (UC UBA1), Megalosaurus bucklandii (NHM R.234; OUMNH J.23014) and Torvosaurus tanneri (ML 1100), spinosaurids such as Baryonyx walkeri (NHM R.9951), Suchomimus tenerensis (MNN G35-9), and Irritator challengeri (SMNS 58022), and other allosauroids like Allosaurus fragilis (USNM 8335), Neovenator salerii (MIWG 6348) and Acrocanthosaurus abakensis (NCSM 14345).

Both teeth also possess a slightly curved distal profile of the crown, with the apex of the teeth located just apical to the most distal point of the crown at the cervix. This feature was considered to be a potential synapomorphy for Abelisauridae by Smith (2007) as a straight or slightly curved distal profile of the crown exists in Majungasaurus crenatissimus, Indosuchus raptorius, Rugops primus, Kryptops palaios, Aucasaurus garridoi (Smith & Vechia 2006; Smith & Lamanna 2006; Smith 2007; Candeiro 2007; pers. observ.) and many indeterminate abelisaurids (e.g., UNPSJB-PV247; UCPC 10; MNHN MRS 1619, MRS 1620). Although the distal profile of the crown displays a strong curvature in most other theropods (Ezcurra 2009; pers. observ.), a weak curvature of the distal profile can also occur in some teeth of basalmost theropods (PVSJ 512), ceratosaurids (USNM 4735; MLP 26-39), noasaurids (PVL 4061), carcharodontosaurids (SGM Din1; MCF-PVH 108.43), tyrannosauroids (MIWG 1997.550; USNM 12814; FMNH PR2081) and some coelurosaurs (Currie et al. 1990: fig. 8.5A; Sankey et al. 2002: fig. 4.10), therefore the systematic utility of this feature requires association with other characters.

Nevertheless, the presence of strongly developed and elongated interdenticular sulci between distal denticles seem to be a condition genuinely shared by non-maniraptoriform avetheropods as they have been observed in the abelisaurids Kryptops palaios (MNN GAD1−1) and Majungasaurus crenatissimus (FMNH PR 2100, 2278), the megalosaurids Megalosaurus bucklandi (OUMNH J13506) and Torvosaurus tanneri (ML 1100), the carcharodontosaurids Giganotosaurus carolinii (MUCPv-CH-1) and Mapusaurus roseae (MCF-PVPH-108), and the tyrannosaurid Tyrannosaurus rex (FMNH PR2081). On the other hand, an irregular stipple texture of the enamel (i.e., no specific orientation of the enamel wrinkling texture) seems to be present in most non-tetanurans theropods such as Coelophysoidea and Abelisauroidea, and many coelurosaurs like Tyrannosaurus, Compsognathidae and Deinonychosauria (pers. observ.). On the contrary, a braided/veined oriented texture of the enamel has been observed in Ceratosauridae, Megalosauroidea, Allosauroidea and most Tyrannosauroidea and it is therefore unlikely that ML 327 and ML 966 belong to one of those clades.

A peculiar anatomical feature of ML 327 is the also presence of distal denticles with an apex pointing towards the tip, a feature present in the teeth of some abelisauroids such as Masiakasaurus knopfleri (FMNH PR 2221, 2296), Kryptops palaios (MNN GAD1−1), Rugops primus (MNN IGU1), Majungasaurus crenatissimus (FMNH PR 2008, 2100, 2278) and other abelisaurid taxa (e.g., MUCPv 482; MUCPv 641). Among large theropods such as ceratosaurids, megalosauroids, allosauroids and tyrannosauroids, the denticles are symmetrically rounded or slightly asymmetrically convex in lateral view but never hooked apically (contra Bakker & Bir 2005 for ceratosaurids and allosaurids, and Smith 2007 for tyrannosaurids; Currie et al. 1990; Abler 1992; pers. observ.). Slightly to strongly hooked distal denticles can also be observed in many Troodontidae (e.g., Currie 1987; Currie et al. 1990; Holtz et al. 1998; Longrich 2008) and Dromaeosauridae (e.g., Currie et al. 1990; Currie & Varricchio 2004; Baszio 1997; Longrich 2008). These theropods, however, possess either very large and well-separated serrations, as in troodontids, or a number of denticles per five millimetre superior to 14 on the distal carina (Smith et al. 2005). Both dromaeosaurids and Masiakasaurus also tend to have larger distal denticles when compared to mesial serrations (Currie et al. 1990; Currie & Varricchio 2004; Norell et al. 2006; Longrich 2008; pers. observ.) and, to our knowledge, neither noasaurids nor deinonychosaurs display pronounced and elongated interdenticular sulci and short marginal undulations on the crown.

Interestingly, ML 966 lacks hooked denticles on the distal carina. This would therefore suggest that apically recurved denticles might not be present in all teeth along the tooth row. Denticles recurvature seems indeed to vary in the dentition of Majungasaurus crenatissimus as strongly recurved denticles are present in lateral and mesial dentary teeth and slightly recurved to symmetrically rounded denticles exist in some lateral and premaxilla teeth (Fanti & Therrien 2007; Smith 2007; pers. observ.).

The presence of an elongated and deep groove adjacent to the mesial carina on the lingual side of the crown in ML 327 is another peculiar feature that, to our knowledge, has not been observed in any teeth belonging to a large theropod (crown with CH > 30 mm), and might therefore represent an autapomorphy. A concave surface adjacent to the mesial carina can be observed in the most mesial teeth of many abelisaurids such as Rugops primus (MNN IGU1), Indosuchus raptorius (AMNH 1753) and Majungasaurus crenatissimus (FMNH PR 2100), but also in Allosaurus fragilis (AMNH 851), some tyrannosauroids such as Proceratosaurus bradleyi (NHM R 4860) and Eotyrannus lengi (MIWG 1997.550) and some dromaeosaurids like Dromaeosaurus albertensis (AMNH 5356). However, the surface adjacent of the mesial carina in ML327 is convex and the concave area formed by the longitudinal groove is narrow. Longitudinal grooves running along the crown surface can also be observed in several theropod taxa such as Scipionyx samniticus (Dal Sasso & Maganuco 2011), Buitreraptor gonzalezorum and Austroraptor cabazai (Gianechini et al. 2011b), but those grooves are double and separated by a large medial ridge (Gianechini et al. 2011a; Gianechini et al. 2011b; pers. observ.). Likewise, the mesial groove present in ML 3287 cannot be confused with the large medial concavity (‘supradental groove’ of Gong et al. 2010) present on the crown of many theropods like Orkoraptor burkei (Novas et al. 2008) and Sinornithosaurus (Gong et al. 2010), or the numerous flutes visible on the teeth of Coelophysis bauri (Buckley 2009), Masiakasaurus knopfleri (Carrano et al. 2002), Ceratosaurus nasicornis (Madsen & Welles 2000), spinosaurids (e.g., Charig & Milner 1997; Sereno et al. 1998; Sues et al. 2002), Paronychodon lacustris (e.g., Cope 1876; Sankey et al. 2002; Baszio 1997; Sankey 2008) or Velociraptor mongoliensis (AMNH 6515).

On the basis of the combination of several important features in ML 966 and ML 327, a large crown (CH > 30 mm), an almost straight distal profile of the tooth, transversal and short marginal undulations on the crown, denticles with strongly developed interdenticular sulci, a DSDI close to one, an irregular enamel texture and the presence of apically pointed denticles on the distal carina in ML 327, these two teeth are confidently assigned to a member of the Abelisauridae. Within this clade, ML327 and ML 966 only differ from other abelisaurids by having a strongly twisted mesial carina. However, this feature is also present in some basal abelisaurids such as Abelisaurus (MPCA 685).

ML 327 and ML 966 resemble very much the tooth MG GUI D 191 illustrated by Zinke (1998: fig. 8e-f) and identify as belonging to a probable allosaurid. The three teeth share a similar outline, same curvature of the mesial margin and absence of curvature of the distal profile, mesial denticles asymmetrically convex, and the presence of interdenticular sulci and a longitudinal groove adjacent to the mesial carina, like in ML 327. Nevertheless, ML 327 differs from MG GUI D 178 by the absence of a second longitudinal groove close to the distal carina on the lingual surface and a medial ridge mentioned by Zinke (1998) which rather forms a large convex surface medially positioned on the lingual face of ML 327. Unlike ML 327 and ML 966, the tooth illustrated by Zinke (1998) also possesses symmetrically rounded denticles on the distal keel rather than asymmetrically convex or pointed denticles. The presence of two concave areas adjacent to the carinae, clearly visible in the illustration of the cross-section of the mid-crown given by Zinke (1998: fig. 8i) seems to make its affiliation to an allosaurid theropod doubtful. To our knowledge, the lateral (and most mesial) teeth of Allosaurus do not display two concave area delimited by a convex surface on the lingual surface (e.g., USNM 8335; UMNH VP 4166; SMA 0005/02). Such feature is rather present in most mesial teeth of some abelisaurids like Majungasaurus crenatissimus (Fanti & Therrien 2007). Although the CBR of these teeth tends usually to be above 0.5 and often between 0.7 to 1 (pers. observ.) some most mesial teeth of abelisaurids can also have values close to 0.5 (e.g., MPCA 267) corresponding to the values obtained by Zinke (1998) for MG GUI D 191-194. Therefore, we also see those teeth as belonging to a probable abelisaurid.

 

Figures 4 to 7

 

Bivariate plots of CBR and CHR reveal that ML 966 and ML 327 mainly occupy the same area of values than Abelisauridae (Majungasaurus + indeterminate abelisaurids), Ceratosaurus, Allosaurus, Acrocanthosaurus and Gorgosaurus teeth (Fig. 4). However, bivariate plots with MAVG or DAVG clearly show that the two teeth possess smaller mesial and distal denticles than any abelisaurids represented, with a number of denticles per five millimetres situated among the values of Allosaurus, Acrocanthosaurus and Berberosaurus (Figs 5–7). The number of denticles per five millimetres of ML 966 and ML 327 are indeed situated between 13 to 16, a higher number than in Majungasaurus, Indosuchus, Rugops and UCPC 10 (Smith 2007; Sereno & Brusatte 2008; pers. observ.) but comparable to that of the most basal abelisaurid Kryptops (Sereno & Brusatte 2008) and Abelisaurus (pers. observ.).

Due to their piriform shape and a relatively important labiolingual compression of the base crown (CBR close to 0.5), ML 966 and ML 327 are most likely lateral teeth and have therefore been coded as such in our datasets.

 

Figure 8 & 9

 

When ML 327 and ML 966 are included in the dentition-based data matrix of discrete characters, the cladistic analysis yielded 9 most-parsimonious trees where the resulting consensus tree retrieved both teeth in a relatively well supported clade (Bremer support of 3) encompassing all other Abelisauridae (Fig. 8). ML 966 was found grouped with Majungasaurus and ML 327 forms the sister taxon of this clade. The monophyletic group formed by ML 966, ML 327 and Majungasaurus is supported by 4 synapomorphies: the hooked distal denticles (char. 87), the long and well-developed interdenticular sulci of basal and mid-crown denticles on the distal carina (char. 105 & 106), and the presence of tenuous transversal undulations on the crown (char. 109). A similar result was found when using both discrete and continuous characters but ML 327 and ML 966 were grouped together among the clade of Abelisauridae, forming a polytomy with Skorpiovenator and a clade including several derived abelisaurids and the troodontid Troodon (Fig. 9). When incorporated into the supermatrix of discrete and continuous characters as well as discrete characters only, both analyses resulted in a same topology. ML 327 and ML 966 were both nested among Abelisauridae, ML 966 being more closely related to Majungasaurus than ML 327 (Fig. 10).

 

Chronological and biogeographical implications. Results of both cladistic and morphometrical analyses indicate that ML 327 and ML 966 belong to a member of the Abelisauridae, representing the earliest record of this clade in Laurasia and the first record of abelisaurids in the Kimmeridgian-Tithonian.

Abelisauridae have often been considered as one of the dominant terrestrial predators in most Gondwanian landmasses during the Cretaceous (Carrano & Sampson 2008). Their presence is now attested in the Jurassic of Gondwana as a newly described abelisaurid, Eoabelisaurus mefi, comes from the Middle Jurassic of Argentina, extending the lineage of this clade by more than 40 million years (Pol & Rauhut 2012). Abelisaurid teeth have also been reported in the Middle Jurassic of the Mahajanga basin of Madagascar by Maganuco et al. (2005). Although there is no doubt that many of the teeth described and illustrated by these author can be referred to an Abelisauridae (contra Flynn et al. 2006; e.g., morphotype 1 is highly diagnostic to most mesial teeth of abelisaurid taxa based on the outline of basal cross-section, hooked denticles and convex mesial profile), the Bathonian age given to the material seems to be unsure as the presence of similar to identical teeth belonging to the abelisaurid Majungasaurus crenatissimus are abundant in the Late Cretaceous sediment of the Mahajanga basin (e.g., Dépéret 1896; Smith 2007; Fanci & Therrien, 2007). The dinosaur material reported from the “Bathonian, faciès mixte – dinosauriens” sensu Besairie (1972) by Maganuco et al. (2005, 2007) may therefore come from the Upper Cretaceous of the Mahajanga basin, perhaps from the Maevarano Formation of Maastrichtian in age (e.g., Rogers et al. 2000; Abramovich et al. 2002; Rogers et al. 2007), and the abelisaurid teeth may belong to Majungasaurus. Pending on a better knowledge of the stratigraphy of the region where the material was collected, we prefer to consider the Jurassic age of the abelisaurid teeth from Madagascar as uncertain.

With the exception of Eoabelisaurus, oldest records of abelisaurids come from the Hauterivian-Barremian of Argentina (Rauhut et al. 2003) and the Aptian-Albian of Niger (Sereno & Brusatte 2008) as all potential abelisaur remains from the Middle and Late Jurassic of Gondwana and Laurasia pertained to Abelisauroidea (e.g., Rauhut 2005; Allain et al. 2007; Carrano & Sampson 2008; Ezcurra & Agnolín 2012). In Europe, abelisaurid remains are scarce and have only been collected from the Upper Cretaceous of France (Buffetaut et al. 1988; Le Loeuff & Buffetaut 1991; Carrano & Sampson 2008; Tortosa et al. 2010). The two abelisaurid teeth discovered in the Late Jurassic of Portugal therefore indicate a first radiation of this clade in the European archipelagos well before the Late Cretaceous. With Allosauridae, Ceratosauridae, Megalosauridae, Tyrannosauroidea, Compsognathidae, Dromaeosauridae and Archaeopterygidae previously documented, the theropod fauna of the Lourinhã Formation included elements specific to Europe (Compsognathus and Archaeopteryx) and also those known in Northern America (Allosaurus, Ceratosaurus, Torvosaurus) but the presence of Abelisauridae adds for the first time a typical Gondwanian element to the large diversity of the Laurasian theropod in the Late Jurassic of the Iberian Peninsula. As it was already suggested by Buffetaut (1989) and Le Loeuff (1991) for Cretaceous theropods, the European Jurassic theropod fauna may have been a mixture of Gondwanian and Laurasian elements where the typical Gondwanian abelisaurids are in minority.

 

Tetanurae Gauthier, 1986

Megalosauroidea Fitzinger, 1843

 

Megalosauridae Fitzinger, 1843

 

Torvosaurus tanneri Galton & Jensen, 1979

 

Figure 11

 

Referred material. ML 962 (Fig. 11).

Type locality and horizon. Cliffs of Praia da Area Branca North, Praia da Area Branca, Lourinhã, Portugal. Bombaral Member, Lourinhã Formation, Tithonian, Upper Jurassic.

Description. ML 962 is an elongated tooth lacking the mesial part of the tip. Although most of the mesial and distal denticles are damaged and missing, their base is still present so that it was possible to count the number of denticles basally, apically and at the mid-crown.

Crown. The tooth is particularly large, (CH of 85 mm) and the general shape of the tooth resembles the ‘typical’ blade-like theropod tooth by being labiolingually compressed, distally curved and serrated carinae. However, the base is particularly narrow mesio-distally (CBL of 31.5 mm) and quite large labio-lingually (CBW of 20.2 mm) so that the crown-base has an ovoid cross-section.

In lateral view, the mesial and distal margins of the root and basal half of the crown are roughly straight whereas the distal half to the crown bent distally. The curvature of the crown is larger mesially than distally and the base of the crown is slightly larger than the mid-crown mesio-distally.

In distal view, the distal carina is medially positioned, slightly curved and bowed labially. The keel bears denticles all along the crown edge, from the preserved tip of the crown to the cervix.

In mesial view, the mesial carina, on the other hand, appears at the mid-crown, approximately 30 mm from the cervix, the basal part of the crown remaining smooth and rounded. The carina is labially positioned and weakly offset apically but slightly curves lingually towards the root, becoming medially positioned on the mesial margin of the crown. Both lingual and labial surface are baso-apically concave and the root surface remains almost straight.

In apical view, the tip is weakly labio-lingually oriented and medially positioned in the crown. The mesial carina forms just a low ridge whereas the distal keel is more acute, and bends lingually towards the root.

In cross section, the basal crown is elliptical with both mesial and distal parts rounded. The labial face shows a short flattened surface in its centre whereas the lingual margin is weakly convex. Both labial and lingual surfaces are strongly mesio-distally convex all along the crown. The dentine layer is thin (0.6 mm in the lingual part) and its thickness is bigger on the distal part of the crown (1.7 mm), the mesial part being absent. The length of the pulp cavity is 17.8 mm labio-lingually and around 28 mm mesio-distally.

Denticles. The mesial carina counts 8 denticles at the mid-crown and the number of denticles near the apex is unknown due to the missing part of tip of the crown. The size of the denticles decreases towards the root from approximately the two-third of the crow, a tendency also observable on the distal carina but on a much longer distance.

The distal keel bears around 7 denticles per 5 mm at the apex, 8 at the mid-crown and 11 at the base of the crown, the latter being minute near the cervix. The biggest denticles can be found 20 mm below the apex of the crown and are the only denticles entirely preserved on the apical part of the distal keel. They are chisel-like in shape, mesio-distally longer than baso-apically and their main axis is perpendicular to the distal margin. A transversal section of the denticles would reveal a triangular shape as their bases are labio-lingually large and their tips are angular.

The labial and lingual surfaces of the denticles are slightly convex or completely flattened baso-apically, and only their basal and apical borders are rounded and curved to form the limits of the interdenticular spaces. The latter are deep and narrow and often filled with sediments. Their width tends to decrease towards the tip of the denticles which is slightly wider baso-apically than the base. This enlargement of the denticles towards their tips is clearly visible in distal view and the denticles show on both apical and basal surfaces a rounded ridge medially positioned between the interdenticular spaces and labio-lingually oriented. These ridges link each denticle on their two-third rather than on their summit.

The external margin of the denticles is symmetrically slightly convex and does not point towards the tip of the crown. The denticle surface is cover by enamel but the layer of enamel has disappeared in the middle of the several denticles surfaces. This might however be due to erosion rather than initial wear. A few other denticles are also preserved on the basal part of the distal carina. They are quite different from the apical denticles by having a much more rounded external margin. The denticles are symmetrically rounded in lateral view and their labial and lingual surfaces are strongly convex. The interdenticular space is shallower and also slightly wider than in the apical denticles.

The mesial and distal denticles also differ in their elongation; the few preserved denticles on the mesial carina are longer baso-apically than mesio-distally. The baso-apical width of the denticles does not really increase towards the tip of the crown and the interdenticular space is narrow and deep. The external margin of the denticle is slightly convex, almost flat, and the denticles connect each other on their summit.

Short interdenticular sulci appear between the distal denticles, but not in the most apical and basal ones. These shallow grooves running on both labial and lingual surfaces of the crown are inclined towards the root and more pronounced on the lingual face. They are however totally absent between the mesial denticles.

Surface. The crown surface is rugged and show many irregularities. May be due to erosion and wear, the enamel texture of the crown is completely smooth and does not show any microscopic sculpturing. Two large transversal undulations appear on both labial and lingual surface of the basal part of the crown but those deep structures does not correspond to the numerous and shallow transversal undulations illustrated by Brusatte et al. (2007) and might be due to deformation.

Discussion. Since most of the root is missing and the pulp cavity is excavated and filled with sediment, we interpret ML 962 as a shed tooth (Bakker & Bir 2004). A very large and fairly straight crown showing a labio-lingually compression, distinct serrations on mesial and distal carinae, and a slight curvature of the tip distally is a combination of character observed in theropod dinosaurs only (Buffetaut & Ingavat 1986), especially in the Upper Jurassic of Portugal (pers. observ.).

With a crown height of more than height centimetres (CH of 85.8 mm), ML 962 is a particularly large crown belonging to a particularly large theropod. Although size is a plastic feature and must be used carefully for systematic purpose, this feature has already demonstrated to be useful for discriminating teeth of different theropod taxa (Smith 2005; Smith et al. 2005; Han et al. 2011). Indeed, to our knowledge, crowns of more than height centimetres are only bore by non-maniraptoriform averostrans, as they can be found in Ceratosauridae (Ceratosaurus, Genyodectes), Megalosauroidea (e.g., Torvosaurus and Spinosaurus), Allosauroidea (e.g., Carcharodontosaurus, Mapusaurus, Giganotosaurus)  and Tyrannosauridae (e.g., Tyrannosaurus, Tarbosaurus).

The denticles of ML 962 are also particularly coarse and an average of 8 denticles per 5 mm on both carinae is a condition present in particularly large basal tetanurans. Such feature can indeed be observed in Megalosauridae (Rauhut & Werner 1995; Smith 2007; pers. observ.), Carcharodontosaurinae (Rauhut & Werner 1995; Veralli & Calvo 2004; Corria & Currie 2006; pers. observ.) and Tyrannosauridae (Rauhut & Werner 1995; Smith 2005; pers. observ.). To our knowledge, less than 9 denticles on both mesial and distal carinae is a feature absent in basal Megalosauroidea (e.g., Piatnitzkysaurus), some Megalosauridae (e.g., Eustreptospondylus, Dubreuillosaurus), non-carcharodontosaurine Allosauroidea (e.g., Allosaurus, Neovenator, Acrocanthosaurus), and all Ceratosauridae and Spinosauridae (pers. observation). Indosuchus raptorius (AMNH 1753, 1955, 1960) is the only abelisaurid possessing less than 8 denticles per 5 mm on both carinae but the teeth are typical of abelisaurids as their crowns are low and weakly recurved distally. It is therefore unlikely that ML 962 belongs to one of these groups of theropods.

With an elliptical outline of the base-crown in cross-section and an important elongation, ML 962 is also very peculiar. In most carnivorous theropods, the lateral teeth are usually strongly medio-laterally flattened, giving a lenticular or piriform outline of the base crown in cross-section, and an elliptical outline of the crown base is usually present in most mesial teeth, i.e., the premaxillary and most mesial teeth of the dentary and maxilla (pers. observ.). Among basal tetanurans but Spinosauridae (which possess conical and fluted crowns along the tooth row), an ovoid subcircular outline of the crown base can clearly be observed in most mesial teeth of megalosaurids such as Duriavenator hesperis (NHM R.332), Dubreuillosaurus valesdunensis (MNHN 1998-13) and Torvosaurus tanneri (Britt 1991) and allosauroids like Acrocanthosaurus atokensis (NCSM 14345) and Giganotosaurus carolinii (MUCPv-CH-1; Candeiro 2007). Some tetanurans like Acrocanthosaurus, Giganotosaurus and Tyrannosaurus can also have an ovoid cross-section of the base crown more distally in the jaws (Smith 2005; Candeiro 2007; pers. observ.). Nevertheless, the lateral teeth of those theropods are much more massive and incrassate, the labiolingual width of the crown base being sometimes equal or larger than its mesiodistal length in Tyrannosauridae, giving them the typical ‘banana’ shape (Smith 2005; pers. observ.). We therefore interpret ML 962 as a most mesial tooth of a basal tetanurans.

This large crown also possesses a mesial carina medially positioned on the mesial margin of the crown, running slightly diagonally and terminating at the mid-crown, well above the cervix. Among most mesial teeth of tetanurans, such combination of features can be observed in Megalosauridae such as Torvosaurus tanneri (BYUVP 2003), Duriavenator hesperis (NHM R.332) and Dubreuillosaurus valesdunensis (MNHN 1998-13) as well as the carcharodontosaurid Acrocanthosaurus atokensis (NCSM 14345). In Allosauridae and Tyrannosauroidea, the mesial carina extends to the cervix of the crown, or very close to it, and either twists lingually like in Allosaurus fragilis (AMNH 851; CM 21703; SMA 0005/02) and Proceratosaurus bradleyi (Rauhut et al. 2010) or faces entirely lingually in more derived tyrannosauroids, giving the typical D-shape cross-section of the base-crown (Smith 2005; Sereno et al. 2009; pers. observ.). The distal carina of ML 962 is also centrally positioned on the distal margin of the crown, a similar feature visible in the most mesial teeth of megalosaurids such as Eustreptospondylus oxoniensis (OUMNH J.13558), Dubreuillosaurus valesdunensis (MNHN 1998-13) and Duriavenator hesperis (NHM R.332). On the other hand, the distal carina of most mesial teeth of carcharodontosaurids such as Acrocanthosaurus atokensis (NCSM 14345) and Giganotosaurus carolinii (MUCPv-CH-1) is slightly to strongly displaced labially on the distal margin of the crown (a similar feature found in Genyodectes and Dromaeosaurus for instance; Currie et al. 1990; Rauhut 2004; pers. observ.), so that the mesial and distal carinae are not aligned on a same plan like in megalosaurid theropods (pers. observ.). It is therefore more likely that ML 962 belongs to a Megalosauridae than a Carcharodontosauridae.

Among Megalosauridae, a very large and strongly elongated crown (CHR > 2.5) with large chisel-like and symmetrically rounded denticles (less than 9 denticles on the carina) seems to be a combination of characters only seen in Torvosaurus (pers. observ.). The general shape and outline of ML 962 also resemble very much those of one probable Torvosaurus tanneri shed tooth illustrated by Jensen (1985: fig. 5e) and the first dentary tooth of Torvosaurus (BYUVP 2003). These two teeth share with ML 962 same curvature and elongation as well as a lateral face particularly convex. In addition, the outline of the basal crown seems to fit with the alveoli of the mesial alveoli of the dentary of Torvosaurus (Britt 1991: fig. 3f), the premaxillary alveoli being more elongated mesio-distally (or labio-lingually for the first alveolus).

Both morphological and cladistic analyses support the identification of ML 962 to the taxon Torvosaurus. Bivariate plots of MAVG and DAVG (Fig. 6) show that ML 962 possesses the same number of denticles per five millimetres than Carcharodontosaurus, Tyrannosaurus and Indosuchus, and close values of denticles than Torvosaurus. However, bivariate plots of CHR and DAVG clearly illustrates the same values of ML 962 and Torvosaurus teeth (Fig. 5), on the opposite of bivariate graphs with CBR, as CBR values of ML 962 and Torvosaurus teeth are significantly different (Figs 4, 7). This can be explained by the absence of mesial teeth of Torvosaurus in our dataset. As it has already been mentioned previously, most mesial teeth of many theropods are usually labiolingually thicker than lateral teeth, and this is clearly the case in Torvosaurus and Megalosauridae where most mesial teeth have an elliptical to rounded cross-section at the crown base instead of a lenticular outline typically present in the lateral teeth of these taxa. Following this observation, characters on most mesial teeth were only coded in ML 962 in our data matrix.

When using discrete characters only, the cladistic analysis of dentition-based characters recovered ML 962 into a clade grouping the carcharodontosaurid Carcharodontosaurus with the three megalosaurids Afrovenator, Megalosaurus and Torvosaurus (Fig. 8). This lack of resolution can be explained by the total absence of most mesial teeth in Afrovenator, Megalosaurus and Carcharodontosaurus. A similar topology was found when the cladistic analysis was performed on the supermatrix of discrete characters and continuous characters as well as discrete characters only, but the polytomy this time only includes megalosaurid taxa (Fig. 10). Interestingly, ML 962 and Torvosaurus are recovered in a same clade in the resulting consensus tree from the analysis of the dentition based dataset when continuous characters are treated as such (Fig. 9).

Following the results of both cladistic and morphological analyses, we identify ML 962 as a mesial tooth, perhaps a dentary tooth, belonging to the species Torvosaurus tanneri. Material of Torvosaurus tanneri are not rare in the Kimmeridgian – Tithonian of Europe and North America and have already been reported several times in the Lourinhã Formation previously (Antunes & Mateus 2003; Mateus 2005; Mateus et al. 2006). Therefore, this referral to Torvosaurus is coherent both biogeographically and stratigraphically.

 

 

Neotetanurae Sereno et al., 1994

 

Coelurosauria von Huene, 1914

 

Richardoestesia Currie et al., 1990

 

cf. Richardoestesia gilmorei Currie et al., 1990

 

Figure 12

 

Referred material. ML 939 (Fig. 12).

Type locality and horizon. Cliffs of Valmitão South, Lourinhã, Portugal. Amoreira-Porto Novo Member, Lourinhã Formation, Tithonian, Upper Jurassic.

Description. The crown is entirely preserved but shows an important wear facet extending on the distal part of the mesial margin of the tooth. A small piece and some denticles of the distal carina are missing but most of them are intact and well-preserved. The tooth lacks the root.

Crown. The crown is small (CH of 5.1 mm), slightly elongated (CBH of 1.82) and strongly compressed labio-lingually (CBR of 0.5). The tip is strongly recurved distally and the apex is pointed, mostly due to the wear facet. The mesial carina is missing and might have been worn on the tip of the crown. The distal carina is serrated and bears denticles from the cervix to the apex.

In lateral view, the crown has a straight crown along the basal part which then abruptly curves distally at the two-third of their height at an angle of 55° to the vertical, forming an acute backward tip. The most basal part of the crown is slightly constricted mesio-distally but the constriction only occurs on the mesial margin of the crown, the distal margin being straight along the one-fourth of the crown. The distal carina is globally concave but the keel is curved above the straight basal margin and the distal part of the carina is straight. The mesial margin is convex above the cervix and only at the half of the crown, the other half remaining straight due to the wear facet. A convex surface delimited by a longitudinal groove mesially and a flattened or slightly concave surface distally appears on both lingual and labial faces. This large mesial ridge follows the same curvature of the crown and its mesio-distal width decreases towards the tip. It starts one-third of the crown on the labial face and from the apical root on the lingual surface. Both lingual and labial grooves are narrow and reach the wear facet at the tip.

In mesial view the crown tip is straight and curves neither labially nor lingually. Both labial and lingual faces are weakly convex and the crown-base width is slightly narrower than the mid-crown width. The crown remains however strongly compressed labio-lingually all along its height and the crown width slightly decreases from the mid-crown to the tip.

In distal view, the most basal part of the serrated carina is straight and vertical but then curved all along the rest of the crown with the bow directed lingually. The distal carina is slightly oriented lingually (we regarded the lingual face of the crown as the face towards which the distal carina was displaced) and the lingual face adjacent to the carina is flat whereas the labial surface near the keel is concave.

In apical view, the basal part of the mesial margin is strongly convex and the wear facet situated on the distal part forms a narrow flat surface revealing the enamel and the dentine layers. In basal view, the crown-base forms an “8-shape” in cross section due to a labio-lingual constriction. This concave surface on the lingual face is shallow, triangular in shape and extends on one-third of the crown whereas the one on the labial face is slightly deeper and ends at the cervix level.

In basal view, the mesial part of the crown is labio-lingually wider (1.2 mm) than the distal part (1 mm). The dentine layer is thicker in the level of the constriction, giving an even well-pronounced “8-shaped” to the pulp cavity, thinner distally.

Denticles. Only the distal carina is serrated and the morphology of the denticles varies along the keel. With 10 denticles per 1 mm basally and at the mid-crown and 9 apically, the denticles increase in size near the apex. The basal denticles are indeed longer mesio-distally than baso-apically whereas the apical denticles are larger baso-apically than mesio-distally. In lateral view, the basal denticles are tong-shaped with their external margin strongly convex, parabolic and symmetrically rounded or slightly pointing towards the tip of the crown. The interdenticular space is elongated, deep and usually filled with sediment.

The interdenticular sulci are absent or very short, shallow and straight extending perpendicular to the distal margin on the labial and lingual faces from between the denticles. In apical view, the lingual and dorsal surfaces of the denticles body are convex and the denticle tip is chisel-like in shape. The interdenticular space is large and the denticle body is baso-apically narrow. The basal denticles become mesio-distally shorter towards the root and the mid-crown. The apical denticles are short and baso-apically larger than the basal ones. The denticles of the basal series have symmetrically rounded external margin forming almost semicircle on the keel. The most apical ones are cartouche-shaped with their external margin pointing towards the apex. The denticles at the apex are mesio-distally short and just form a small symmetrical bump in lateral view. The interdenticular sulci are absent in the apical denticles.   

Surface. The enamel surface of the crown show irregular and finally wrinkled structures on both sides. Except the presence of those microscopic sculptures, there is no other structure on the crown surface.

Discussion. ML 939 is interpreted as a shed tooth as it lacks most of the root and the pulp cavity is slightly excavated.

The presence of a basal constriction between crown and root has been observed in basal most theropods like Eoraptor lunensis (Sereno et al. 1993) and many coelurosaurs such as the ornithomimosaur Pelecanimimus (Pérez-Moreno et al. 1994), alvarezsaurids (Perle et al. 1993), basal oviraptorosaurs (Osmólska et al. 2004), therizinosaurs (e.g., Russell & Dong 1993; Zhao & Xu 1998; Kirkland et al. 2005), troodontids (e.g., Currie et al. 1990; Baszio 1997; Norell et al. 2000; Currie & Dong 2001; Sankey et al. 2002; Averianov & Sues 2007), the dromaeosaurids Saurornitholestes (Currie et al. 1990; Baszio 1997) and Microraptor (Xu et al. 2000), and many basal avialans such as Archaeopteryx and Cathayornis (Hou 1997; Feduccia 2002).

Nevertheless, the presence of an eight-shape outline of the crown-base in cross-section has only been recorded in some derived deinonychosaurs such as Saurornitholestes (Currie et al. 1990; Sankey et al. 2002), Tsaagan (Norell et al. 2006), Pyroraptor (Allain & Taquet 2000; Gianechini et al. 2011b), Buitreraptor (Gianechini et al. 2011b) and the enigmatic theropod Richardoestesia gilmorei (Currie et al. 1990).

ML 939 serrations are particularly minute and the distal carina bears nine to ten denticles per one millimetre. Among deinonychosaurs, such condition only exists, to our knowledge, in the taxa Richardoestesia gilmorei (e.g., Sankey 2001; Sankey et al. 2002; Baszio 1997; Larson 2008). Likewise, the external margins of the denticles are symmetrically rounded or slightly curved towards the tip of the crown, a condition also shared by Richardoestesia gilmorei (Baszio 1997; Larson 2008). Although the presence of a longitudinal groove mesially positioned on the crown has never been noticed in Richardoestesia gilmorei, this feature seems to be present in this species (see Baszio 1997: Plate IV), and longitudinal groove have already been observed in the genus Richardoestesia (Currie et al. 1990; Sankey 2001; Rauhut 2002).

With a strongly labiolingually compressed profile of the crown, ML 939 was coded as a lateral tooth. Both cladistic analyses performed on the dentition-based dataset of discrete and continuous characters and discrete characters only recovered ML 939 as a close relative of Richardoestesia gilmorei (Figs 8 & 9). The clade encompassing those taxa (ML 939 + Richardoestesia gilmorei), which forms a polytomy with the coelurosaurs Buitreraptor and Compsognathus, is define by three ambiguous synapomorphies: an important constriction occurring at the base crown on the distal side of the tooth (characters 63 and 64) and a wide mesiodistally concave area on the basal part of the labial margin of the crown (character 71). The analysis performed on the data matrix also yielded a clade gathering ML 939 and Richardoestesia gilmorei but, this time, the monophyletic group is either retrieved among coelurosaurs, forming a polytomy with Compsognathidae and Deinonychosauria (Fig. 10), or nested among Dromaeosauridae, closely related to Bambiraptor, when continuous characters are treated as such (see appendix).

Richardoestesia gilmorei is a common species in the Late Cretaceous of Northern America and teeth of this taxon have been identified in the Santonian Milk River Formation, the Campanian Belly River Group, the Campanian-Maastrichtian Horseshoe Canyon Formation, and the Maastrichtian Scollard Formation of Alberta, the Frenchman Formation of Saskatchewan (Canada), the Hell Creek Formation of Montana and the Lance Formation of Wyoming (e.g., Currie et al. 1990; Baszio 1997; Longrich 2008; Sankey 2008; Larson 2008).

Small theropod teeth from the Upper Jurassic of Portugal have already been assigned with caution to the genus Richardoestesia by Zinke (1998). Nevertheless, they strongly differ from ML 939 by being extremely elongated and weakly recurved, resembling the elongated and subtriangular teeth assigned to Richardoestesia sp. by Baszio (1997), and Richardoestesia isosceles by Sankey (2001). Following the cladistic analysis and the diagnosis of teeth belonging to Richardoestesia sp. (and R. gilmorei in particular) given by Currie et al. (1990), Baszio (1997) and Longrich (2008), and since the presence of teeth similar to those of Richardoestesia isosceles has been reported in the Late Jurassic of Portugal (Zinke 1998), ML 939 is ascribed to the enigmatic taxon Richardoestesia which the stratigraphic range may therefore have extended back to the Jurassic. ML 939 is nearly the same as R. gilmorei teeth but this taxon has only been recorded in the Late Cretaceous of North America, more than 90 million years after the Jurassic/Cretaceous boundary, and we therefore consider that ML 939 belongs to a close relative of Richardoestesia gilmorei.

 

Conclusion

 

The description and identification of four theropod teeth from the Lourinhã Formation provide additional information on the Late Jurassic dinosaur fauna of the Iberian Peninsula and the biogeographical and stratigraphic distribution of Abelisauridae. Based on both morphological and cladistic analysis using a new data matrix of 140 characters on teeth, two isolated teeth have been successfully identified as belonging to an Abelisauridae, one to megalosaurid Torvosaurus tanneri, and one as a close relative of the enigmatic coelurosaur Richardoestesia gilmorei, spreading the already high diversity of predatory dinosaurs living in the Kimmeridgian – Tithonian of Southern Europe. If these referrals are correct, theropods from the Upper Jurassic of Europe are now represented by Ceratosauridae, Abelisauridae, Megalosauridae, Allosauroidea, Tyrannosauroidea, Compsognathidae, Deinonychosauria and Avialae corresponding to a mixture of Laurasian and Gondwanian elements.

Although materials of Torvosaurus tanneri and a close relative of Richardoestesia have already been identified in the Upper Jurassic of Portugal, an abelisaurid is here reported for the first time in the Lourinhã Formation and therefore represents the first record of Abelisauridae in the Late Jurassic of Laurasia and the second oldest record in the globe, revealing a first radiation of this clade in Europe back to the Jurassic.

As previously noted by Smith et al. (2005) and more recently by Han et al. (2011), this study also shows that morphometrical data, combined with numerous anatomical characters on teeth, proves to be useful in order to clarify the phylogenetical position of isolated theropod teeth. Although many dentition-based characters are homoplastic, several theropod clades such as Ceratosauridae, Abelisauridae, Spinosauridae, Compsognathidae and Tyrannosauroidea have distinctive teeth, characterized by a combination of features that were not taken into consideration previously. This provides tools for the identification of isolated theropod teeth, often more common than bones, therefore allowing expanding our knowledge about the geography and chronology of theropod taxa, as demonstrated in this case for abelisaurids. This is particularly encouraging for future research on theropod dentition and, thereby, additional information regarding the size and shape of crown, carinae, denticles and enamel texture and microstructure remain to be collected on teeth of many theropod taxa. Moreover, the dentition of a large number of theropod dinosaurs is usually briefly described, sometimes even avoided, and detailed descriptions of premaxilla, maxilla and dentary teeth of many well-preserved theropods such as Ceratosaurus, Torvosaurus, Allosaurus and Sinraptor still need to be done and would greatly facilitate the assignment of isolated teeth to specific clades or taxa.