A new small deinonychosaur ( Dinosauria : Theropoda ) from the Late Cretaceous of Patagonia , Argentina

Here we report on a new small deinonychosaurian theropod, Pamparaptor micros gen. et sp. nov., from the Late Cretaceous of Patagonia, Argentina. Pamparaptor micros exhibits a pedal structure previously unknown among South American deinonychosaurians. The new material provides new evidence about the morphology and taxonomic diversity of Patagonian deinonychosaurs. Pamparaptor is the smaller non-avialae Patagonian deinonychosaur, probably with about 0.50-0.70 meters, long. The pedal construction resembles, that of Troodontid or basal Dromaeosaurids. Nevertheless, up to now, we considered Pamparaptor a peculiar Patagonian Dromaeosaurid with troodontid-like pes.

Here we report on a new deinonychosaurian, Pamparaptor micros gen. nov sp. nov, collected from the Baal quarry, at the north coast of Barreales Lake, Neuquén, Argentina (Fig. 1). The specimen comes from the Portezuelo Formation, (Turonian-Coniacian), Neuquén Group, and it is represented by pedal elements. The new form is gracile and small, with about 0.50-0.70 meters, and its metatarsal construction is highly derived, resembling troodontids (Xu and Wang 2000). We offer here a brief description of this notable discovery.  of Barreales Lake, at 90 kilometers NW of Neuquén city, Neuquén province, Patagonia, Argentina. Up to now, the record of dinosaurs collected in the Portezuelo Formation includes the giant lognkosaurian titanosaur Futalognkosaurus dukei , the dromaeosaurids Unenlagia comahuensis (Novas and Puerta 1997), Unenlagia paynemili (Calvo et al. 2004), and Neuquenraptor argentinus (Novas and Pol 2005), the alvarezsaurid Patagonykus puertai (Novas 1997), the allosauroid Megaraptor namuhnuaiquii (Novas 1998), as well as abelisaurids remains , Juárez Valieri et al. 2008), undescribed theropods (Porfiri et al. 2005, and fragmentary remains of a possible neornithine bird .

Diagnosis:
Pamparaptor micros distinguishes from other deinonychosaurians in the following combination of features: slender metatarsus construction with metatarsals IV strongly compressed transversely on its distal half, acquiring a blade-like shape in caudal view; metatarsals IV and III subequal in length; proximal half of metatarsal III narrow and with subparalel margins along its length; metatarsal III lacking a distal ginglymus; metatarsal II distally overlapping metatarsal III; proximally, metatarsal II is approximately twice the wide of either metatarsals III and IV; phalanx 2.II longer than phalanx 2.I; in cranial view, distal end of metatarsal II with a small sulcus medially directed.

Description and comparison:
The articulated left foot is well preserved (MUCPv-1163). The metatarsal III is 9.3 centimeters long, thus suggesting that the whole lenght of the animal was approximately 0.50-0.70 meters. The first and fifth metatarsals are absent. Metatarsal II is shorter than metatarsals III and IV. The second metatarsal is less compressed mediolaterally than metatarsals III and IV. Metatarsal II is 3.5 mm thick for most of its 82.1 mm length. The distal articular surface of metatarsal II present a well-developed gynglimus, which is different to the troodontids Saurornithoides mongoliensis (IVPP V 10597), Tochisaurus nemegtensis (after Kurzanov 1987) and Troodon formosus (after Wilson and Currie 1985). The troodontid Borogovia gracilicrus (alter Olsmolska 1987) presents a poorly-developed gynglimus, which is less-developed than in Pamparaptor. The third and fourth metatarsals are the longest elements in the metatarsus. The third metatarsal expend for all front of metatarsus and is not squeezed out by the second and fourth metatarsals throughout most of the midshaft region as in the troodontids Troodon (Wilson and Currie 1985) and Tochisaurus (Kurzanov and Osmolska 1991). This morphology is like Saurornithoides mongoliensis (Currie and Peng 1993). The distal end of the third metatarsal supports the largest phalanx of the foot. In posterior view, the distal end of metatarsal III is covered by metatarsals II and IV. A significant synapomorphy in troodontids is that the fourth metatarsal are the strongest element in the metatarsus. In Pamparaptor, the fourth metatarsals are very think in the distal end (1mm), but in the proximal end are more robust (3.5 mm). The distal end and probably the middle shatf are mediolaterally compressed in this specimen. The first phalanx of the second digit is short (16.6 mm) and relatively robust. It is followed by the distinctive but longer (18.1 mm, maximum length) II-2. The ungual is strongly curved. The tip of the ungual for digit II was not found, but the base is quite deep, indicating a robust element. Phalanx III-l is 26.5 in length, and is the longest phalanx preserved. The next phalanx, III-2, is a minor element (length =17.4 mm). The third and fourth phalanges are unknown. The fourth digit is the most incomplete, with only the distal end of IV-1? and proximal end of IV-2? being preserved in articulation.

COMPARATIONS WITH THE UNENLAGIINAE Neuquenraptor
Here, we intensify the comparations with Neuquenraptor argentinus because it is the most complete pes into the Unenlagiins form. Pamparaptor micros shares with Neuquenraptor argentinus the following characteristics: metatarsal III strongly compressed and proximally pinched between metatarsals II and IV (present also in basals dromaeosaurids; see Novas and Pol 2005); metatarsal IV with a posterolateral flange (present also in basals dromaeosaurids; see Novas and Pol 2005); extensor sulcus on the proximal half of metatarsal II; metatarsal II with lateral expansion over the caudal surface of metatarsal III (originally considered an autapomorphy of N. argentinus by Novas and Pol 2005). Pamparaptor micros gen. et sp. nov. differs from Neuquenraptor in the proportions of the metatarsals and phalanges, especially those of digit II. In Pamparaptor, the metatarsals are considerably more slender and transversely compressed than in Neuquenraptor. Phalanx 1.III is more elongated than phalanx 2.III (Fig. 4). Curiously, metatarsals III and IV are subequal in length (Fig. 5), a morphology that is observed in troodontids and basal dromaeosaurids (Xu and Wang 2000), but not in Neuquenraptor and Rahonavis, in which these bones are unequal, and metatarsal III is longer than IV. In troodontids (e.g., Troodon formosus), metatarsal IV is usually more robust (Fig. 6).
In cranial view, the distal end of metatarsal II of Pamparaptor presents a sulcus directed medially, different to from Neuquenraptor in which the metatarsal is absent. The distal end of metatarsal III lacks of a ginglymoid articulation. In Pamparaptor micros, the expansion of the lateral crest of metatarsal IV begins proximally. The phalanges of digit II are subequal in length. In Neuquenraptor, pedal phalanx 2.I is bigger than 2.II, but in Pamparaptor, it is the reverse. The ungual phalanx of pedal digit II of Pamparaptor is proporcionally diferent than in Neuquenraptor. In Neuquenraptor and other derived Dromaeosaurids (Velociraptor, Dromaeosaurus, Deinonychus) the claw is lower and more acute than in Pamparaptor.

COMPARISONS WITH TROODONTIDS
The metatarsus of Pamparaptor shares with troodontids the small size and the slender structure. Moreover, metatarsals III and IV are subequal in length. However, proximally, the mediolateral widths of metatarsals IV and II are similar. In troodontids the metatarsal IV is the most robust. In Pamparaptor micros, the fourth digit is incomplete, which does not allow us to know which one is the longest finger.
On the other hand, the raptorial second digit is present in troodontids and dromaeosaurids. The size of this phalanx of digit II ranges between dromaeosaurids and troodontids. However, Russell and Dong (1993) described 26 characters that distinguising troodonts from dromaeosaurs. The characters 23 at 26 are of metatar- sion of articular surface; (26) Pedal phalanx II-1 longer than II-2 (this character is inverse in Pamparaptor); distal articulation of II-2 short (relative to the condition in dromaeosaurs; this character is absent in Pamparaptor); and II-3 shorter than II-1 and not strongly recurved (relative to the condition in dromaeosaurs; this character is absent in Pamparaptor).

DISCUSSION AND CONCLUSIONS
Originally, specimen MUCPv-1163 was referred to as Neuquenraptor by Porfiri et al. (2007) based on the similarities with this Patagonian form that was also documented in the same beds. In this context, specimen MUCPv-1163 was interpreted as a subadult stage of Neuquenraptor. However, complete technical preparation of the materials allowed confirming the absence of a displacement of metatarsal IV; therefore, there is a clear difference between MUCPv-1163 and Neuquenraptor, not related to their ontogenetic stage.
Available information suggests the presence of an endemic group of Deinonychosauria during Cenomanian times (Makovicky et al. 2005). Unfortunately, the record of deinonychosaurians from the Portezuelo Formation, which is the unit with the most complete record of theropods from South America, continues to be fragmentary, making difficult to solve the phylogenetic position of this clade.
The discovery of Pamparaptor supports a new lineage of Gondwanan deinonychosaurs, which is different from those present in Laurasia but with problematic interpretation of relationships. Nevertheless, the phylogenetic resolution of the Gondwanan dromaeosaurids is still problematic.
Neuquenraptor was considered as a junior synonym of Unenlagia (Makovicky et al. 2005, Turner et al. 2007) on the base that both came from the same stratigraphic and geographic provenances. However, based in the pedal phalanx 2.II of Unenlagia paynemili (Calvo et al. 2004), we consider that the pedal morphology of digit II of Neuquenraptor and Unenlagia is different. In Unenlagia paynemili, phalanx 1.II is longer than phalanx 2.II. as that present in Rahonavis and other dromaeosaurids. In Neuquenraptor, phalanx 1.II and 2.II are subequal. In Pamparaptor, phalanx 2.II is greater than phalanx 1.II. (Fig. 7). If Unenlagia paynemili's phalanx belonged to the same specimen, this would demostrate that at least Unenlagia paynemili and Neuquenraptor argentinus are different species; therefore, the Unenlagiinae clade is invalid and the basal politomy is unresolved.
The troodontids and basal dromaeosaurids pes presents several similarities (Xu and Wang 2000). Moreover, this convergence is present in the cranial morphology. The Gondwana raptors (unenlagiines?) Buitrerap- tor and Austroraptor, present low and elongated skulls as that of Byronosaurus. In this case, Pamparaptor presents metatarsal III at the same level of metatarsal IV, as it is seen in troodontids; however, metatarsal IV, of P. micros is less robust than that of troodontids.
Metatarsal IV is a very important tip in the comprehension of relationships about Pamparaptor, because if metatarsal IV is similar in size or more robust than metatarsal II add to the metatarsal III and IV are subequals in long we could be in presence of the firth troodontid dinosaur from South America. The characters present in Pamparaptor show a peculiar pedal construction with several characters that resemble the ones of troodontids. Nevertheless, the few evidence available and several characters checked and compared with troodontids demonstrate that Pamparaptor is a particular basal dromaeosaurid with a troodontid-like pes.

ACKNOWLEDGMENTS
We thank the technician Diego Rosales that found the specimen and the CePaLB team for the preparation of the material.