Noqvitates PUBLISHED by the AMERICAN MUSEUM of NATURAL HISTORY CENTRAL PARK WEST at 79TH STREET, NEW YORK, N.Y

Home , Avisauridae, Avisaurus, Concornis, Enantiornis, Lecho Formation, Neuquenornis

AMERICAN MUSEUM Noqvitates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 3083, 27 pp., 13 figures, 4 tables December 27, 1993

Enantiornithine (Aves) Tarsometatarsi from the Cretaceous Lecho Formation of Northwestern Argentina

LUIS M. CHIAPPE1

ABSTRACT Enantiornithine tarsometatarsi from the Maas- metatarsal II forms a sharp edge (convergent with trichtian Lecho Formation at the locality of El those of L. bretincola). Brete (northwestern Argentina) are described to- The relationships among these three species, and gether with their associated material. Three new their interrelationships with respect to other Late species, namely Yungavolucris brevipedalis, Lec- Cretaceous enantiornithine taxa, are explored. A tavis bretincola, and Soroavisaurus australis, are character analysis based on tarsometatarsal fea- recognized. Y. brevipedalis is distinguished by hav- tures is presented. This analysis supports the hy- ing a short and broad tarsometatarsus with a pul- pothesis that Soroavisaurus australis is the sister leylike trochlea of metatarsal II, and equally long group of a clade formed by Avisaurus archibaldi metatarsals III and IV with the distal end of the and a new form from the Two Medicine Forma- former laterally curved. The long and slender tar- tion of Montana. This clade is in turn the sister sometatarsus of L. bretincola is characterized, group of Neuquenornis volans. These four taxa among other features, for bearing a hypotarsus de- compose the monophyletic taxon Avisauridae. In veloped primarily over the metatarsal II. In turn, the present analysis, the relationships among Lec- S. australis exhibits a long and narrow fenestra tavis bretincola, Yungavolucris brevipedalis, and between the proximal halves ofmetatarsals III and Avisauridae remain unresolved. IV, and the plantar surface ofthe proximal half of RESUMEN Se describen los tarsometatarsos, y material aso- nuevas especies: Yungavolucris brevipedalis, Lec- ciado, de Enantiornithes del Maastrichtiano de la - tavis bretincola, y Soroavisaurus australis. Y. bre- Formacion Lecho en la localidad de El Brete (no- vipedalis se distingue por poseer un tarsometatarso roeste de Argentina). Se reconocen y nombran tres corto y ancho, la tr6clea del metatarso II en forma ' Frick Research Fellow, Department of Vertebrate Paleontology, American Museum of Natural History.

Copyright C American Museum of Natural History 1993 ISSN 0003-0082 / Price $3.60 2 AMERICAN MUSEUM NOVITATES NO. 3083 de polea, los metatarsos III y IV de longitud equi- otros Enantiornithes del Cretacico tardio. Se pre- valente, y el extremo distal del metatarso III cur- senta un analisis de caracteres basado en caracteres vado lateralmente. El delgado y largo tarsometa- del tarsometatarso. Este an'alisis sustenta la hi- tarso de L. bretincola se caracteriza, entre otros potesis que Soroavisaurus australis es el grupo her- rasgos, por poseer un hipotarso desarrollado prin- mano de un clado formado por Avisaurus archi- cipalmente sobre el metatarso II. A su vez, S. aus- baldi y una nueva forma de la Formacion Two tralis exhibe una larga y angosta fenestra entre las Medicine de Montana. Este clado es a su vez el mitades proximales de los metatarso III y IV, y la grupo hermano de Neuquenornis volans. Estos superficie plantar de la mitad proximal del me- cuatro taxones comprenden el taxon monofiletico tatarso II formando un borde filoso (convergente Avisauridae. El presente an'alisis no resuelve las con aquel de L. bretincola). relaciones entre Lectavis bretincola, Yungavolucris Se exploran las relaciones filogeneticas entre es- brevipedalis y Avisauridae. tas tres especies, y sus relaciones con respecto a

INTRODUCTION The Enantiornithes is a Cretaceous group limb elements is known (contra Walker, of volant birds originally recognized by 1981). Walker (1981) on the basis of the large as- In this paper the enantiornithine tarso- semblage of bones from the Maastrichtian metatarsi from El Brete, and the material as- Lecho Formation at the locality of El Brete sociated with them, are described. Three new (southern Province of Salta, northwestern species are recognized and their phylogenetic Argentina) (fig. 1). In this paper Walker il- relationships analyzed. The rest ofthe El Brete lustrated several bones from El Brete, in- collection will be described elsewhere. cluding the three types oftarsometatarsi here described, and briefly remarked on the pe- INSTITUTIONAL ABBREVIATIONS culiar anatomy of the group. After the erec- tion of the AMNH American Museum of Natural History Enantiornithes, several authors GI Geological Institute, Mongolian Academy of assigned previously described forms (Martin, Sciences, Ulaan Baatar 1983) as well as new taxa (Molnar, 1986; Nes- LH Universidad Aut6noma, Madrid sov and Jarkov, 1989; Chiappe, 1991a) to MACN Museo Argentino de Ciencias Naturales, Bue- this clade. To date, enantiornithine birds are nos Aires recorded in the Lower Cretaceous of Austra- MOR Museum of the Rockies, Bozeman lia (Molnar, 1986) and Spain (Sanz et al., in MUCPv Museo de Ciencias Naturales, Universidad press), and the Upper Cretaceous of South Nacional del Comahue, Neuquen America (Walker, 1981; Chiappe, 1991a, PU Princeton University, Princeton 1991b), North America (Martin, 1983; PVL Fundaci6n-Instituto Miguel Lillo, Tucumatn Chiappe, 1992a; Varricchio and in QM Queensland Museum, South Brisbane Chiappe, UCMP Museum of Paleontology, University of Cal- press), and Asia (Martin, 1983; Nessov, 1984; ifornia, Berkeley Nessov and Jarkov, 1989). Furthermore, a YPM Yale Peabody Museum, New Haven. putative member of this clade was recently discovered in the Lower Cretaceous ofChina (Zhou et al., 1992). LOCALITY AND GEOLOGICAL SETTING The discovery of the Enantiornithes The material described here was found in strongly influenced subsequent interpretation a small quarry (approximately 8 m wide) of of early avian evolution. Nevertheless, their the middle section of the Lecho Formation most significant evidence, the large assem- in the Estancia El Brete, in the southern tip blage of bones from El Brete, still remains ofthe Argentine province ofSalta (fig. 1). The mostly unstudied. Interpretation of the El Estancia El Brete is located within the dry, Brete enantiornithines is complex due to the forested, hilly landscape of the southern part poor association among the different speci- of the phytogeographic province of the Yun- mens. Few bones were found in articulation, ga. The first excavations ofthis quarry started and no association between hind and fore- in 1975, when a crew ofthe Fundacion-Insti- l1993 CHIAPPE: CRETACEOUS TARSOMETATARSI 3

65O21' 65'20'W

£~~0.5kmN

EX Subgroup Santa Barbara Subgroup Balbuena lEL BR EN E; | Subgroup Pirgua

Fig. 1. Map indicating the locations of the different subgroups of the Salta Group in the lands of the Estancia El Brete and the quarry (black circle) of the Lecho Formation in where the specimens were found. The label "El Brete" shows the farm houses of the Estancia. tuto Miguel Lillo (Tucuman, Argentina) dis- coastal plain, with abundant vegetation and covered the fossiliferous locality, which was ponds. The age of the Lecho Formation is reported by Bonaparte et al. (1977). generally regarded as Maastrichtian on the The Lecho Formation is part ofthe Upper basis ofits fauna (Bonaparte et al., 1977) and Cretaceous Balbuena Subgroup (Salta Group), stratigraphic relationships (Go6mez Omil et a near-border stratigraphical unit of the An- al., 1989). dean sedimentary basin (Bonaparte et al., 1977; Bonaparte and Powell, 1980). Recent MATERIALS AND METHODS stratigraphical studies defined the Lecho For- Anatomical nomenclature follows Baumel mation as those clastic lithofacies ofthe Bal- et al. (1979), using the English equivalents of buena Subgroup deposited in subaerial, flu- the Latin terminology. In regard to the tax- vial, and eolian environments (Gomez Omil onomic names here applied, the term Aves et al., 1989). Lithologically, the Lecho For- is used in reference to a group including all mation consists of mainly reddish, fine to taxa descended from the most recent com- medium, with an occasional coarse, sand- mon ancestor ofArchaeopteryx lithographica stone (Bonaparte et al., 1977). and modem birds. It is relevant to mention, The avian assemblage from El Brete comes however, that this application has been re- from fine-grained sandstones. The vertebrate cently challenged by new theoretical asser- fauna associated with the avian remains is tions restricting the name of widely known composed of several specimens of the ar- taxa to the crown group (De Queiroz and mored titanosaurid Saltasaurus loricatus, the Gauthier, 1990, 1992). Under this new in- small nonavian theropod Noasaurus leali, and terpretation, the taxon name Avialae (Gau- indeterminate nonavian theropod teeth thier, 1986) replaces Aves, and the latter sub- (Bonaparte and Powell, 1980). stitutes the taxon name Neomithes of the Bonaparte et al. (1977) suggested that the classical avian taxonomy. paleoenvironment of the Lecho Formation The three species erected in this paper are in the area ofEl Brete was a fluvial-lacustrine based on incomplete material, which is still 4 AMERICAN MUSEUM NOVITATES NO. 3083 sufficiently diagnostic. It should be noted that group for analyzing the enantiomithine in- these are not the first enantiomithine species terrelationships. M. olecranus (see Perle et named from El Brete. Walker (1981) applied al., 1993b for nomenclatural correction) has the name Enantiornis leali to the specimen been recently interpreted as the sister group PVL-4035 that includes a proximal end of of all known birds except A. lithographica the humerus articulated to the coracoid and (Perle et al., 1993a). The outgroup material scapula. The status ofthis species name, how- studied includes specimens MACN-N-03, ever, is dubious because Walker (1981) did MACN-N-10, MACN-N-Il, and MUCPv- not describe this specimen and the species 48 ofP. deferrariisi (Chiappe, 1992b), and GI name was merely listed in the legend to a N107/6 and GI N100/99 of M. olecranus table. Because of this, and the additional (Perle et al., 1993a). problems due to the lack of association be- Derived characters exclusive to a single in- tween forelimb and hind limb elements, this group taxa were omitted from the analysis in author believes that the tarsometatarsi de- order to avoid their influence on the consis- scribed in this paper must be allocated in tency index (Carpenter, 1988; Wiley et al., their own species instead of waiting for new 1991). The data matrix presented in table 4 material with association of limb elements. was processed using the implicit enumeration The enantiomithine material studied in- (ie) command of Hennig86 program (Farris, cludes all known tarsometatarsi (and mate- 1988). rial associated with them) from El Brete (PVL- 402 1-1, PVL-4040, PVL-4048, PVL-4052, SYSTEMATIC PALEONTOLOGY PVL-4053, PVL-4268, PVL-4690, PVL- 4692); the enantiomithine tibiotarsi PVL- Aves Linnaeus, 1758 4030, PVL-4032-1, PVL-4033, PVL-4036, PVL-4695, PVL-4696 from El Brete, and a Enantiomithes Walker, 1981 cast ofthe holotype (QM F12992) ofNanan- new tius eos (Molnar, 1986); the holotype (LH- Yungavolucris brevipedalis, species 2814) of Concornis lacustris (Sanz and Bus- DiAGNosIs: Enantiomithine bird with short calioni, 1992), recently reinterpreted as an and broad tarsometatarsus possessing the fol- enantiornithine bird (Sanz et al., in press); lowing autapomorphies: distal end much the holotype (MUCPv- 142) ofNeuquenornis broader than proximal end; broad, dorso- volans (Chiappe and Calvo, in press); a cast plantarly compressed, and pulleylike trochlea (MACN- 18685) ofthe holotype and the para- of metatarsal II; distal end of metatarsal III type PU- 17324 (now housed at the YPM) of laterally curved; metatarsal IV equal in length Avisaurus archibaldi (Brett-Surman and Paul, to metatarsal III; prominent dorsal ridge be- 1985), whose enantiomithine kinship was re- tween distal halves of metatarsals II and III. cently established (Chiappe, 1992a); and the ETYMOLOGY: Yungavolucris from Latin holotype (MOR-553E) of a new avisaurid volucris, meaning "bird," and Yunga refer- species from the Two Medicine Formation ring to the phytogeographic region in which (Varricchio and Chiappe, in press). El Brete is located; brevipedalis from Latin Polarization of character states used in the brevis, "short," and pedalis referring to the character analysis was established by using foot. Patagopteryx deferrariisi (Alvarenga and HOLOTYPE: PVL-4053 (fig. 2); nearly com- Bonaparte, 1992), and Mononykus olecranus plete right tarsometatarsus, lacking the cra- (Perle et al., 1993a). The Jurassic bird Ar- niolateral border ofthe lateral cotyla and the chaeopteryx lithographica was not selected cranial portion of the trochlea of metatarsal because its two-dimensional preservation IV. obscures several features. The sister-group REFERRED SPECIMENS: Incomplete right relationship between Enantiomithes and a tarsometatarsi PVL-4040 and PVL-4692 (fig. clade formed by P. deferrariisi plus the Or- 3), incomplete left tarsometatarsus PVL-4052 nithurae has been recently established by this (fig. 4), and specimen PVL-4268 represented author (Chiappe, 1992b, in press). Under this by the distal trochleae of right metatarsals II framework, P. deferrariisi is an adequate out- and III. 1 993 CHIAPPE: CRETACEOUS TARSOMETATARSI 5

TABLE 1 Measurements (in mm) of Yungavolucris brevipedalis, new species PVL specimens 4040 4052 4053 4268 4692 Total length, tarsometatarsus 41.6 51.9 41.9 - 42.8 Total length, metatarsal II 37.8 41.5 37.0 - 38.4 Maximum width, proximal end of tarsometatarsus 16.2 15.3 14.0 - 15.0 Maximum width, trochlea metatarsals II and III 19.6 20.4 19.4 20.0 19.7 Maximum width, trochlea metatarsal II 11.8 10.9 10.4 12.1 11.0 Maximum width, trochlea metatarsal III 7.2 6.8 6.2 6.3 6.5

LOCALITY AND HORIZON: 26OO2'07 "S tarsal II is very broad. The trochlea of this 65019'57"W. Estancia El Brete, Department metatarsal is ample, dorsoplantarly com- of Candelaria, Province of Salta, Argentina pressed and pulleylike (figs. 2F, 3C). Both (fig. 1). Lecho Formation, Upper Cretaceous rims of the trochlea are bulbous and are sep- (Maastrichtian). arated from each other by a broad central DESCRIPTION: The tarsometatarsus of Yun- furrow. Just proximal to the medial rim of gavolucris brevipedalis is short and stout. It the trochlea, on the medial margin ofthe dor- is dorsoplantarly compressed, being generally sal face, there is a dorsomedial projection (fig. convex dorsally and flat plantarly (fig. 2). Its 3A). The degree of development and orien- distal end is much broader than its proximal tation ofthis projection varies among the dif- end. Metatarsals II, III, and IV are firmly ferent specimens (cf. figs. 2A, 3A). attached to each other, although they are fused Metatarsals III and IV are bent laterally only in their proximal portion (figs. 2A, B, over their distal third, and they are much 3A). longer than metatarsal II (figs. 2A, B, 3A). In In the proximal end, the medial cotyla is the dorsal face ofthe metatarsal III,just distal well developed and suboval; the lateral cotyla to the proximal end, there is a slender and is less defined and its articular surface is elongated ridge for muscle attachment (fig. slightly convex medially (figs. 2E, 3B, 4). Both 2A). Distally, in spite of the lateral bending an intercondylar eminence and hypotarsus of the shaft, the trochlea of metatarsal III is are absent. in a vertical position (figs. 2A, B, F, 3A, C). Metatarsal II is the most robust ofthe three The medial rim of this trochlea is larger and metatarsals, but not the longest. It is straight more bulbous than the lateral (fig. 3A, B), throughout its length (figs. 2A, B, 3A). Dor- and deep collateral ligamental fossae exca- sally, in its middle portion it exhibits a very vate both sides. pronounced ovate-shaped tubercle, with its Metatarsal IV is very small with respect to long axis oriented proximodistally (figs. 2A, metatarsals III and II (figs. 2A, B, 3A). It is C, 3A). This tubercle was tentatively inter- more weakly attached to metatarsal III than preted as the area of attachment of the M. this one is to metatarsal II. The trochlea is tibialis cranialis (Brett-Surman and Paul, missing in all available specimens except in 1985; Chiappe, 1992a). The distal halfofthis the holotype where only the plantar half is metatarsal is fairly compressed dorsoplan- preserved. It is simple and laterally com- tarly, especially in its medial portion. In the pressed (fig. 2A, B, D, F). lateral portion of the distal half, its contact REMARKS: Comparisons between the tar- with metatarsal III forms a prominent ridge sometatarsi of Yungavolucris brevipedalis and (figs. 2A, C, D, 3A). The distal end of meta- those of the remaining enantiornithine taxa 6 AMERICAN MUSEUM NOVITATES NO. 3083

Fig. 2. Holotype specimen (PVL-4053) of Yungavolucris brevipedalis, n. sp. Right tarsometatarsus in dorsal (A), plantar (B), medial (C), lateral (D), proximal (E), and distal (F) views. dp dorsomedial projection of the distal end of metatarsal II, dr distal ridge between the contact of metatarsals II and III, tu tubercle for the attachment of the M. tibialis cranialis, lc lateral cotyla, mc medial cotyla, rm elongate proximal ridge for muscle attachment, II-IV metatarsals TI-IV. and the outgroup forms indicate several au- tarsometatarsus in Yungavolucris brevipeda- tapomorphies diagnosing the former taxon. lis is unique to this species among the En- The remarkable difference between the antiornithes (cf. figs. 2A, 7A, 12A). In So- width of the proximal and distal ends of the roavisaurus australis, Avisaurus archibaldi, 1993 CHIAPPE: CRETACEOUS TARSOMETATARSI 7 A B

II II IV III III IV D C E

IV III 11

I --MM J o m cms Fig. 2. Continued. Neuquenornis volans, Concornis lacustris, and end. This condition is unknown in Lectavis the Two Medicine Formation form, the distal bretincola. In the outgroups, the difference end is only slightly broader than the proximal between the width ofthe proximal and distal 8 AMERICAN MUSEUM NOVITATES NO. 3083

Fig. 3. Right tarsometatarsus (PVL-4692) of Yungavolucris brevipedalis, n. sp. in dorsal (A), proximal (B), and distal (C) views, dp dorsomedial projection of the distal end of metatarsal II, tu tubercle for the attachment of the M. tibialis cranialis, lc lateral cotyla, mc medial cotyla, rm elongate proximal ridge for muscle attachment, II-IV metatarsals II-IV. ends ofmetatarsals II-IV is also considerably feature in the outgroups indicates that it is smaller than in Y. brevipedalis. The signifi- another autapomorphic character of Y. brevi- cant difference in the width of the two ends pedalis. of the tarsometatarsus is autapomorphic for In contrast to Soroavisaurus australis (figs. Y. brevipedalis. 7A, B, 8), Avisaurus archibaldi (fig. 12), Con- The tarsometatarsus of Yungavolucris cornis lacustris, and the Two Medicine For- brevipedalis differs from all other known en- mation form, metatarsal IV of Yungavolucris antiomithine tarsometatarsi (the condition is brevipedalis is as long as, not shorter than, unknown in Lectavis bretincola) in that the metatarsal III (fig. 2A, B). The condition re- distal end of metatarsal III curves laterally mains unknown in Neuquenornis volans and (figs. 2A, B, 3A), instead of being straight. A Lectavis bretincola in which the distal end of straight metatarsal III is also present in the this metatarsal is missing. Metatarsal IV is outgroup species Patagopteryx deferrariisi and shorter than metatarsal III in the outgroup Mononykus olecranus. This feature is inter- forms. The subequal length ofmetatarsals III preted as an autapomorphy of Y. brevipe- and IV is also regarded as an autapomorphy dalis. of Y. brevipedalis. The presence in Yungavolucris brevipedalis The tarsometatarsus of Yungavolucris of a prominent ridge on the distal half of the brevipedalis also differs from the tarsometa- dorsal surface ofthe tarsometatarsus between tarsus ofSoroavisaurus australis (figs. 7F, 1 1), the union of the metatarsals II and III is ab- Avisaurus archibaldi (fig. 11), Neuquenornis sent in Lectavis bretincola (fig. 6A), Soroavi- volans, Concornis lacustris, and the Two saurus australis (figs. 7A, 8A, 9), Neuquenor- Medicine Formation form in having a re- nis volans (fig. 9), Avisaurus archibaldi (fig. markably broad, dorsoplantarly compressed, 1 2A), Concornis lacustris, and the Two Med- and pulleylike trochlea of metatarsal II (figs. icine Formation form. The absence of this 2F, 3C). In Lectavis bretincola the condition 1993 CHIAPPE: CRETACEOUS TARSOMETATARSI 9 A B

mc Ic

IV III 11 III

cms Fig. 3. Continued. is unknown. This peculiar morphology is also suboval articular surface of proximal end of absent in the outgroups, being here inter- tibiotarsus; distal tibiotarsal condyles strong- preted as another autapomorphy of Y. brevi- ly projected cranially; plantar surface ofprox- pedalis. imal half of metatarsal II forming a prominent edge; subcircular knob on proximal end of metatarsal II; hypotarsus mostly devel- Lectavis bretincola, new species oped over metatarsal II. DIAGNOSIS: Enantiomithine bird with long ETYMOLOGY: Lectavis from Latin lectus, and slender tibiotarsus and tarsometatarsus equivalent to the the Spanish word lecho possessing the following autapomorphies: ("bed"), in reference to the Lecho Formation,

mc Ic

cms Fig. 4. Left tarsometatarsus (PVL-4052) of Yungavolucris brevipedalis, n. sp. in proximal view. Ic lateral cotyla, mc medial cotyla. 10 AMERICAN MUSEUM NOVITATES NO. 3083

TABLE 2 Measurements (in mm) of Lectavis bretincola, new species Specimen PVL-4021-1 Maximum preserved length, tarsometatarsus 88.0 Maximum width, proximal end of tarsometatarsus 15.0 Maximum depth, proximal end of tarsometatarsus 10.7 Total length, tibiotarsus 156.0 Maximum depth, mid-shaft of tibiotarsus 4.8 Maximum width, distal end of tibiotarsus 14.3 Maximum depth, lateral condyle of tibiotarsus 12.7 Maximum depth, medial condyle of tibiotarsus 14.3

and avis, "bird"; bretincola from the com- II-IV are strongly attached to each other over bination of "Brete" and the Latin incola, their length, and they are completely fused which means "an inhabitant." proximally (fig. 6A). The fusion between HOLOTYPE: PVL-4021-1 (figs. 5, 6). Left metatarsals II and III starts more distally (ap- tibiotarsus and incomplete tarsometatarsus, proximately 24.7 mm from the proximal end) missing the distal end of the latter and most than does the fusion between metatarsals III of metatarsal IV. and IV (approximately 17.3 mm from the LOCALITY AND HORIZON: 26002'07 "S proximal end). Dorsally, the tarsometatarsus 65°19'57"W. Estancia El Brete, Department of L. bretincola is in general flat, while in of Candelaria, Province of Salta, Argentina plantar view it is deeply excavated (fig. 6B, (fig. 1). Lecho Formation, Upper Cretaceous C). This excavation is especially pronounced (Maastrichtian). in the proximal portion, where metatarsal II DESCRIPTION: The tibiotarsus of Lectavis forms a strong ridge that medially borders a bretincola is poorly preserved. It is long, slen- depressed area formed by metatarsals III and der, and craniocaudally compressed (fig. 5). IV (fig. 6B, C). The morphology of the proximal end differs In dorsal view metatarsal III is the broad- with respect to the other El Brete enantior- est, narrowing smoothly distally. Its proximal nithine tibiae (e. g., PVL-4032-1, PVL-4033, portion is dorsoplantarly thin while distally PVL-4036, PVL-4696), Concornis lacustris it becomes thicker. Metatarsal II is the more (Sanz and Buscalioni, 1992), and the Austra- conspicuous of the three. Its main axis is in lian enantiornithine Nanantius eos (Molnar, the dorsoplantar plane, and plantarly it forms 1986) in that in proximal view it is not cir- a prominent and robust ridge (fig. 6B, C, D), cular but ellipsoidal, with a lateromedial ma- which attenuates distally. In the most prox- jor axis. The shaft is compressed craniocau- imal portion, this ridge is continuous with a dally (fig. 5). Distally, the condyles of the triangular-shaped hypotarsus (fig. 6B). In tibiotarsus of L. bretincola are strongly pro- contrast to the typical modem avian condi- jected cranially, much more than in any other tion, the hypotarsus of Lectavis bretincola is enantiomithine tibiotarsi from El Brete (e.g., mostly developed on the plantar surface of PVL-4030, PVL-4033, PVL-4695) and in N. metatarsal II. The medial surface of this eos. As in the remaining enantiomithines, the metatarsal is basically flat, with only a slightly medial condyle is inflated, and considerably depressed central area in its most proximal larger than the lateral one (fig. 5A). The in- portion (fig. 6D). Dorsally, in the most prox- tercondylar notch separating these condyles imal portion ofmetatarsal II, there is a prom- is less pronounced and broader than in the inent round tubercle. Above this tubercle, a remaining El Brete tibiotarsi and in N. eos. nearly crescent-shaped fossa is located be- Well-developed epicondylar fossae are pres- tween metatarsals II and III (fig. 6A). Meta- ent on the sides of the condyles. tarsal IV is badly preserved. It is significantly Lectavis bretincola possesses a long and smaller than the remaining ones (fig. 6A, B, slender tarsometatarsus (fig. 6). Metatarsals C). Its proximal portion is remarkably thin. 1 993 CHIAPPE: CRETACEOUS TARSOMETATARSI I1I A B

mc Ic mc cms Fig. 5. Holotype specimen (PVL-402 1-1) ofLectavis bretincola, n. sp. Left tibiotarsus in cranial (A), and medial (B), views. Ic lateral condyle, mc medial condyle.

On the basis of the groove left for its contact major axes are oriented in a mediodorsal- on the lateral border ofmetatarsal III, as with lateroplantar direction. The medial cotyla is the latter, metatarsal IV became thicker to- slightly larger than the lateral cotyla, and its ward the distal end. dorsal border is more elevated as well. The Both cotylae in the proximal articular sur- plantar border of the medial cotyla, above face are ovate in outline (fig. 6E), and their the hypotarsus, is projected proximally (fig. 12 AMERICAN MUSEUM NOVITATES NO. 3083

Fig. 6. Holotype specimen (PVL-402 1-1) ofLectavis bretincola, n. sp. Left tarsometatarsus in dorsal (A), plantar (B), lateral (C), medial (D), and proximal (E) views. tu tubercle for the attachment of the M. tibialis cranialis, lc lateral cotyla, mc medial cotyla, II-IV metatarsals II-IV. 1993 CHIAPPE: CRETACEOUS TARSOMETATARSI 13 A B C D

I11

11 III E

cms Ic

mc Fig. 6. Continued. 14 AMERICAN MUSEUM NOVITATES NO. 3083

6E). As was pointed out by Walker (1981), is distinct from the tarsometatarsi of Yun- the hypotarsus has a subrectangular shape in gavolucris brevipedalis (fig. 2B, C), Neuquen- proximal view, and lacks any crest or sulcus ornis volans, Avisaurus archibaldi (fig. 12B), (fig. 6B, E). and the Two Medicine form in that the prox- The distal end of the tarsometatarsus of imal half forms a robust and laterally com- the single specimen of Lectavis bretincola is pressed medial edge that projects plantarly missing. With the available material it is not (fig. 6A, C, D). A medial ridgelike plantar possible to estimate, even roughly, the ap- projection is also present in the tarsometa- proximate length of this tarsometatarsus. tarsus ofSoroavisaurus australis (figs. 7B, D, REMARKS: Character-state comparisons 8B). However, this condition differs from the among Lectavis bretincola, the remaining en- condition present in L. bretincola in that the antiornithines, and the outgroup forms reveal ridge is less robust and more compressed lat- several autapomorphic features ofthe former erally, and in that its most prominent area is species. not located in the proximal end but more In the tibiotarsus, the suboval proximal end distally (cf. figs. 6D, 7D). Furthermore, in L. and the strong cranial projection ofthe distal bretincola the proximal region of this ridge condyles differ from those of the remaining receives the contribution of metatarsal III, enantiornithines and the outgroups, and are whereas it is exclusively formed by metatar- here considered autapomorphies of Lectavis sal II in S. australis. This prominent plantar bretincola. ridge is completely absent in the outgroup In the tarsometatarsus, Lectavis bretincola forms. The important differences between the differs from all other enantiornithine taxa in plantar ridges of the tarsometatarsus of L. having a hypotarsus (fig. 6B). Contrary to the bretincola and S. australis make a hypothesis statements ofBrett-Surman and Paul (1985), of homology of these structures unlikely. the tarsometatarsus of Soroavisaurus aus- Hence, the medial plantar ridge ofL. bretin- tralis (figs. 7B, E, 8B) and Avisaurus archi- cola is regarded as an autapomorphy. baldi (fig. 12B) lacks a hypotarsus (Chiappe, 1992a), a condition also shared with Yun- AVISAURIDAE BRETT-SURMAN AND PAUL, 1985 gavolucris brevipedalis (figs. 2B, E, 3B, 4), the Two Medicine Formation form, and the out- DiAGNOSIS: Enantiomithine birds with a group forms. The presence of a hypotarsus strong transverse convexity ofthe dorsal sur- in L. bretincola is here regarded as an aut- face of the mid-shaft of metatarsal III, a dis- apomorphy of this taxon. As was mentioned tinct plantar projection of the medial rim of above, the hypotarsus ofL. bretincola differs the trochlea of metatarsal III, and a laterally from the typical modem avian condition in compressed J-shaped metatarsal I (see that it appears to be formed mainly over Chiappe, 1992a). metatarsal II, instead of capping mostly the DEFINITION: Phylogenetically defined (see proximal end ofmetatarsal III (Baumel et al., De Queiroz and Gauthier, 1990, 1992) as the 1979). This indicates that the development common ancestor ofNeuquenornis volans and of a plantar structure on the proximal end of Avisaurus archibaldi plus all its descendants. the tarsometatarsus, a hypotarsus, arose in- INCLUDED SPECIES: Avisaurus archibaldi dependently in modem birds and certain en- (Brett-Surman and Paul, 1985), Neuquenor- antiomithines. nis volans (Chiappe and Calvo, in press), So- Lectavis bretincola differs from other en- roavisaurus australis, n. sp. (this paper), un- antiomithines in that the tubercle on meta- named new species from the Two Medicine tarsal II is circular instead ofovate as in Yun- Formation (Varricchio and Chiappe, in press). gavolucris brevipedalis (figs. 2A, 3A), Soroavisaurus australis (figs. 7A, 8A), Avi- saurus archibaldi (fig. 12A), and the Two Soroavisaurus australis, new species Medicine form. The round morphology of DiAGNoSIS: Enantiomithine avisaurid with this tubercle in L. bretincola is interpreted as the following autapomorphies: plantar sur- an autapomorphy. face ofproximal halfofmetatarsal II forming The tarsometatarsus ofLectavis bretincola a sharp edge (convergent with Lectavis bre- 1 993 CHIAPPE: CRETACEOUS TARSOMETATARSI 15 tincola); long and narrow fenestra open be- TABLE 3 tween the proximal halves of metatarsals III Measurements (in mm) of Soroavisaurus australis, and IV; dorsally projected edge on the troch- new species lea for metatarsal IV. PVL- PVL- ETYMOLOGY: Soroavisaurus from Latin so- 4690 4048 ror, meaning "sister," and Avisaurus (Brett- Surman and Paul, 1985), referring to the sis- Total length, tarsometatarsus 46.9 51.5 these two Maximum width, proximal end ter-group relationship inferred for of tarsometatarsus 10.0 11.3 taxa; australis refers to the ocurrence of this Maximum depth, proximal end species in the Southern Hemisphere. of tarsometatarsus 6.3 6.1 HOLOTYPE: PVL-4690 (fig. 7), left tarso- Total length, metatarsal II 44.2 48.9 metatarsus. Specimen previously referred to Total length, metatarsal IV 43.3 48.4 as Avisaurus sp. (Brett-Surman and Paul, Maximum width, mid-shaft 1985; Chiappe, 1992a; Chiappe and Calvo, of tarsometatarsus 9.9 9.6 in press). Maximum width, distal end REFERRED SPECIMENS: PVL-4048, left tar- of tarsometatarsus 14.6 15.2 sometatarsus, metatarsal I with proximal and Maximum width, trochlea metatarsal II 5.8 5.2 distal phalanges of digit I, and four inter- Maximum width, trochlea mediate phalanges. This specimen was also metatarsal III 4.1 4.2 formerly identified as Avisaurus sp. (Brett- Maximum width, trochlea Surman and Paul, 1985; Chiappe, 1992a; metatarsal IV 4.0 4.3 Chiappe and Calvo, in press). Maximum length, metatarsal I - 12.2 LOCALITY AND HORIZON: 26°02'07' S 65019'57"W. Estancia El Brete, Department of Candelaria, Province of Salta, Argentina (fig. 1). Lecho Formation, Upper Cretaceous (Maastrichtian). tarsal II is more laterally compressed. The DESCRIPTION: The tarsometatarsus of So- middle portion of its plantar surface forms a roavisaurus australis is gracile (figs. 7, 8). prominent edge that limits the medial border Metatarsals II-IV are fused only proximally of the abovementioned plantar depression (figs. 7A, B, 8), and distally these metatarsals (figs. 7B, D, 8B). In Avisaurus archibaldi, the are not as closely apposed as in Yungavolucris plantar border of this metatarsal is slightly brevipedalis or Lectavis bretincola. The prox- projected plantarly, but it does not form this imal end is transversely much broader than prominent ridge (fig. 12B). In the dorsal sur- cranioplantarly. The plane of the proximal face of the proximal third an oval tubercle is articular surface is inclined dorsally (fig. 7C, developed. This tubercle is not as prominent D). In proximal view, the articular surface is as in A. archibaldi, and is located more prox- kidney-shaped, with the concave part facing imally than in the latter species (figs. 7A, 8A). plantarly (fig. 7E). In this respect, S. australis In medial view, the shaft of this metatarsal differs from Avisaurus archibaldi, in which is slightly sigmoid, as best exhibited by the the proximal surface is more oval. The me- holotype specimen. Distally, metatarsal II is dial cotyla is subcircular and of approxi- more widely separated from metatarsal III mately the same size as the lateral one. The than the latter is from metatarsal IV (figs. 7A, articular surface of the latter was a smooth 8A). Metatarsal IV is weak and dorsoplan- convexity of the medial portion (fig. 7E). tarly compressed, especially in its distal half. Metatarsal III is the largest. It is straight In its proximal half, it is extremely slender and dorsally broader than the other two (figs. and more bar-shaped, defining the lateral 7A, 8A). The cranial surface of this metatar- border of an elongate fenestra separating it sal is convex, especially in the middle of the from metatarsal III (figs. 7A, B, 8). This fe- shaft. Plantarly, this metatarsal constitutes nestra is absent in all other enantiornithines. the floor of a depression defined by the plan- Distally, the three trochleae are well de- tar projection of the proximal two-thirds of veloped. As in all avisaurids, the trochlea for metatarsals II and IV (figs. 7B, 8B). Meta- metatarsal III possesses a remarkable plantar AMERICAN MUSEUM NOVITATES NO. 3083

Fig. 7. Holotype specimen (PVL-4690) of Soroavisaurus australis, n. sp. Left tarsometatarsus in dorsal (A), plantar (B), lateral (C), medial (D), proximal (E), and distal (F) views. fe fenestra between metatarsals III and IV, tu tubercle for the attachment of the M. tibialis cranialis, Ic lateral cotyla, mc medial cotyla, II-IV metatarsals II-IV. 1 993 CHIAPPE: CRETACEOUS TARSOMETATARSI 17 A B

III III C D

E mc

c IC

t u I 4 F

I1 III IVt

L . cms Fig. 7. Continued. 18 AMERICAN MUSEUM NOVITATES NO. 3083

Fig. 8. Left tarsometatarsus (PVL-4048) of Soroavisaurus australis, n. sp. in dorsal (A), and plantar (B) views. fe fenestra between metatarsals III and IV, tu tubercle for the attachment of the M. tibialis cranialis, II-IV metatarsals II-IV. projection of the internal rim (Chiappe, trochlea ofmetatarsal II ofS. australis differs 1992a). Collateral fossae are present on both from that ofA. archibaldi in that it has a stout sides of this trochlea. The trochlea for meta- central edge developed on the dorsal surface tarsal II is the broadest, and plantarly it ex- that is totally absent in A. archibaldi (fig. 1 1). hibits lateral and medial prominences sepa- The tarsometatarsus of PVL-4048 is as- rated by a wide depression (figs. 7B, 8B, 11). sociated with metatarsal I and the hallux, and The trochlea for metatarsal IV of Soroavi- four intermediate phalanges. Unfortunately, saurus australis clearly differs from that of this author was not able to find any record Avisaurus archibaldi (fig. 11). In S. australis of the articulation between these bones; they this trochlea has a subtriangular outline in were disarticulated, although with a single distal view, with a medial projection that number and within a single box. Comparison nearly encloses distally a tiny fenestra be- with the articulated foot ofNeuquenornis vo- tween metatarsals III and VI (a structure sim- lans (fig. 9) suggests that they all belong to ilar to the distal vascular foramen of Orni- the same specimen and were probably dis- thurae) (figs. 7A, B, 8B). Conversely, in A. articulated during preparation. Metatarsal I archibaldi this trochlea is crescentic in distal of Soroavisaurus australis is J-shaped, as in view, with a concave medial side. The small N. volans (fig. 9). It is strongly compressed fenestra between metatarsals III and IV in A. laterally and tapers proximally. As in N. vo- archibaldi is also enclosed distally by a me- lans, although metatarsal I was not reversed dial projection of the latter trochlea, but the (Fig. 9) the hallux was reverted. The hallux projection is located much further dorsally is formed by two large phalanges (fig. 10). than in S. australis (fig. 1 1). Furthermore, the The proximal phalanx is dorsoplantarly com- 1 993 CHIAPPE: CRETACEOUS TARSOMETATARSI 19 A B

III

cms Fig. 8. Continued. pressed. On its distal head, the articular facet differs from the tarsometatarsus of the re- of this phalanx is large and extensively de- maining enantiornithine taxa in the posses- veloped on the plantar surface (fig. 10). This sion of a long and narrow fenestra between suggests that the main movement ofthe distal the proximal halves ofmetatarsals III and IV phalanx was in a plantar direction. Well-de- (figs. 7A, B, 8). This fenestra is also absent veloped collateral fossae are present on both in outgroup taxa. The presence of a fenestra sides of the distal head. The distal phalanx between metatarsals II and III is here inter- is sharp, laterally compressed, and strongly preted as an autapomorphy of S. australis. arched (fig. 10). Plantarly, it possesses a small, Another autapomorphic character of So- round flexor tubercle at its base. roavisaurus autralis is the presence of a stout REMARKS: The two known specimens of and thick dorsal trochlear projection ofmeta- Soroavisaurus australis were previously re- tarsal IV (figs. 7F, 11). This projection is not ferred to Avisaurus sp. (Brett-Surman and developed in either Avisaurus archibaldi (fig. Paul, 1985; Chiappe, 1992a; Chiappe and 1 1), Concornis lacustris, or the new form from Calvo, in press). The result of the present the Two Medicine Formation. In Yungavolu- cladistic analysis (see below), however, sup- cris brevipedalis the dorsal portion of the ports their allocation to a different genus. trochlea for the metatarsal IV is missing. Soroavisaurus australis exhibits several au- However, considering the remarkable reduc- tapomorphic features. The tarsometatarsus tion of this trochlea (fig. 2), the presence of 20 AMERICAN MUSEUM NOVITATES NO. 3083

III..-..A

III III Soroavisaurus australis Neuquenornis volans Fig. 9. Avisaurid metatarsals I-IV and metatarsal I in dorsal and medial view. (A) Soroavisaurus australis (based on PVL-4048), (B) Neuquenornis volans (based on MUCPv-142). Scale bars 2 mm. a stout dorsal ridge is unlikely. This condition CHARACTER ANALYSIS is also uncertain in both Neuquenornis volans and Lectavis bretincola. A dorsal, thick, cen- The interrelationships ofEnantiomithes are in very difficult to evaluate because the El Brete tral ridge is absent the trochlea of meta- specimens are mostly disarticulated. There is tarsal IV in the outgroups. no information about the association be- As was previously discussed (see the Re- tween fore- and hind-limb elements, or their marks under Lectavis bretincola), the pres- association with the pelvis or axial skeleton. ence of a prominent, sharp, and laterally This problem is not overcome by the re- compressed ridge on the plantar surface of maining enantiomithine forms recovered in metatarsal II of Soroavisaurus australis (figs. different parts ofthe world, which are mostly 7B, D, 8B) is considered an autapomorphy. fragmentary or still unpublished. For this rea- son, this paper presents a tentative approach to the relationships of Late Cretaceous en- TABLE 4 antiomithines on the basis of the tarsometa- Data Matrix of 10 Characters (0 primitive, 1 de- tarsal morphology alone. rived, ? uncertain) of the Tarsometatarsus. A cladistic analysis ofthe relationships between the new taxa Yungavolucris brevipe- Mononykus olecranus 0000000000 dalis, Lectavis bretincola, Soroavisaurus aus- Patagopteryx deferrariisi 0010000000 tralis, and Avisaurus archibaldi (Brett-Surman Yungavolucris brevipedalis 1001??0001 and Paul, 1985), Neuquenornis volans Lectavis bretincola 1?1 1??00?? (Chiappe and Calvo, in press), and a new Avisaurus archibaldi 1111?11111 form from the Two Medicine Formation of Soroavisaurus australis o101 11 Montana (Varricchio and Chiappe, in press) Neuquenornis volans 1 1?1??101 was performed. Ten binary characters (table Two Medicine form 1 10?1 11111 4) were analyzed cladistically and their dis- 1 993 CHIAPPE: CRETACEOUS TARSOMETATARSI 21

a M p cm

Fig. 10. Digit I of Soroavisaurus australis (PVL-4048) in medial view.

tribution among the abovementioned taxa 2. Medial rim ofthe trochlea ofmetatarsal and the outgroups is discussed below. III with a strong plantar projection: absent 1. Metatarsal IV: large (0) or considerably (0) or present (1). The medial rim of the smaller (1). Metatarsal IV is considerably trochlea ofmetatarsal III is projected strongly smaller than metatarsals II and III in all en- plantarly in Soroavisaurus australis (figs. 7F, antiornithine taxa considered in the analysis 11), Avisaurus archibaldi (fig. 11), Neuque- (figs. 2, 3, 6-8, 12; see also Walker, 1981; nornis volans (Chiappe and Calvo, in press), Chiappe, 1991a, 1992a). Variable sizes of and the Two Medicine Formation form (Va- metatarsal IV are present among the different rricchio and Chiappe, in press). On the other taxa (cf. figs. 2A, 6A, 7A, 1 2A), although these hand, this strong projection is absent in Yun- differences are not clear enough for scoring gavolucris brevipedalis (figs. 2F, 3C) and the distinct character states. This enantiomithine outgroups. condition is absent in both Patagopteryx de- 3. Plantar surface of tarsometatarsus ex- ferrariisi and Mononykus olecranus. cavated: absent (0) or present (1). In Lectavis

A B

I v III I Soroavisaurus australis Avisaurus archibaldi Fig. 1 1. Comparisons between the distal end ofthe tarsometatarsus ofSoroavisaurus australis (PVL- 4690) and Avisaurus archibaldi (cast MACN-18685). II-IV metatarsals II-IV. 22 AMERICAN MUSEUM NOVITATES NO. 3083

Fig. 12. Cast (MACN- 18685) ofthe right tarsometatarsus (holotype specimen) ofAvisaurus archibaldi (Brett-Surman and Paul, 1985) in dorsal (A) and plantar (B) views. tu tubercle for the attachment of the M. tibialis cranialis, II-IV metatarsals II-IV. bretincola (fig. 6B, C), Soroavisaurus australis sal II: absent (0) or present (1). The dorsal (figs. 7B, 8B), Avisaurus archibaldi (fig. 1 2B), faces of metatarsal II of Yungavolucris brev- and Neuquenornis volans (Chiappe and Cal- ipedalis, Lectavis bretincola, Soroavisaurus vo, in press) the plantar face of the tarso- australis, Avisaurus archibaldi, and the Two metatarsus is excavated to variable degrees Medicine Formation form exhibit a pro- exhibiting a proximodistal fossa bounded by nounced tubercle (figs. 2A, 3A, 6A, 7A, 8A, metatarsals II and IV, giving a crescentic cross 12A), previously interpreted as the area of section to the tarsometatarsus. Within the re- attachment ofthe M. tibialis cranialis (Brett- maining enantiornithine taxa considered here, Surman and Paul, 1985). Variable positions this fossa is absent in Yungavolucris brevi- of this tubercle occur among the abovemen- pedalis (fig. 2B) and the new form from the tioned forms. The presence or absence ofthis Two Medicine Formation (Varricchio and tubercle is uncertain in Neuquenornis volans, Chiappe, in press). Within the outgroup, this in which the proximal half of the shaft of fossa is absent in Mononykus olecranus, but metatarsal II is missing (Chiappe and Calvo, the tarsometatarsus of Patagopteryx defer- in press). A tubercle on the dorsal surface of rariisi is excavated plantarly. metatarsal II is absent in the outgroups. 4. Tubercle on the dorsal face of metatar- 5. J-shaped metatarsal I: absent (0) or 1993 CHIAPPE: CRETACEOUS TARSOMETATARSI 23 A B

IV II II IV III III

cms Fig. 12. Continued. present (1). Metatarsal I of Soroavisaurus IV is also concave medially and asymmet- australis and Neuquenornis volans is J-shaped rical, although to a lesser extent than in A. in medial view (fig. 9), and extremely laterally archibaldi (Varricchio and Chiappe, in press). compressed. Unfortunately, metatarsal I is This trochlear morphology is absent in So- missing in Yungavolucris brevipedalis, Lec- roavisaurus australis (Fig. 1 1), while the con- tavis bretincola, Avisaurus archibaldi, and the dition is uncertain in Yungavolucris brevipe- Two Medicine enantiomithine. This meta- dalis, Lectavis bretincola, and Neuquenornis tarsal morphology is also absent in the out- volans. This condition is also absent in the groups, in which metatarsal I is straight or outgroup taxa. slightly curved as in Patagopteryx deferrari- 7. Proximal articular surface dorsally in- isi, and Mononykus olecranus. clined: absent (0) or present (1). The articular 6. Trochlea ofmetatarsal IV medially con- surface of the proximal end ofSoroavisaurus cave: absent (0) or present (1). In Avisaurus australis (fig. 7C, D), Avisaurus archibaldi (fig. archibaldi the trochlea for metatarsal IV is 12A; see also Brett-Surman and Paul, 1985; remarkably concave medially, being highly fig. 3C, E), and the Two Medicine Formation asymmetrical (figs. 1 1, 1 2B). In the Two form (Varricchio and Chiappe, in press) slants Medicine form, the trochlea for metatarsal toward the dorsal border such that this plane 24 AMERICAN MUSEUM NOVITATES NO. 3083

]3Kr AVISAURIDAE ENANTIORNITHES

B / 4

- AVISAURIDAE

.ENANTIORNITHES

ENANTIORNITHES

Fig. 13. Fundamental cladograms (length 12, consistency index 0.83, retention index 0.87) resulting from the analysis of the data matrix by using the implicit enumeration option of Hennig86 program. A, node 1: characters 1, 4; node 2: character 10*; node 3: characters 2, 5*, 8; node 4: characters 7*, 9; node 5: character 6. B, node 1: characters 1, 4, 10; node 2: character 3*; node 3: characters 2*, 5*, 8; node 4: characters 7*, 9; node 5: character 6. C, node 1: characters 1, 4, 10; node 2: characters 2, 5*, 8; node 3: characters 7*, 9; node 4: character 6. Asterisked characters have an ambiguous optimization. forms a sharp angle relative to the dorsal plane 8. Strong transverse convexity of the dor- of the tarsometatarsus. This character is un- sal surface of the mid-shaft of metatarsal III: certain in Neuquenornis volans. Otherwise, absent (0) or present (1). The dorsal surface the articular surface of the proximal end is of the mid-shaft of metatarsal III is strongly not inclined in Yungavolucris brevipedalis (fig. convex transversely in Soroavisaurus aus- 2C, D), Lectavis bretincola (fig. 6C, D), or in tralis (figs. 7A, C, 8A), Avisaurus archibaldi any of the outgroup forms. (fig. 12A; see also Brett-Surman and Paul, 1 993 CHIAPPE: CRETACEOUS TARSOMETATARSI 25

1985; fig. 3H), Neuquenornis volans (Chiappe are successive sister taxa of Avisauridae, al- and Calvo, in press), and the Two Medicine though their sequence is reversed. The re- Formation enantiornithine (Varricchio and maining tree (fig. 1 3C), identical to the strict Chiappe, in press). Conversely, this condi- consensus, depicts these three taxa clustered tion is absent in Yungavolucris brevipedalis in an unresolved tricotomy. (figs. 2A, 3A) and Lectavis bretincola (fig. 6A). The Enantiomithes, the clade composed of This condition is also absent in the outgroup all ingroup taxa, is unambiguously diagnosed forms in which the dorsal surface ofthe mid- by a small metatarsal IV (character 1), a tu- shaft is instead slightly convex or nearly flat. bercle in the dorsal face ofmetatarsal II (char- 9. Distal end of metatarsal II strongly acter 4), and a trochlea ofmetatarsal II broad- curved medially: absent (0) or present (1). In er than the trochleae of metatarsals III and Soroavisaurus australis (figs. 7A, B, 8), Avi- IV (character 10). In the fundamental clado- saurus archibaldi (fig. 12), and the Two Med- gram of figure 1 3A, the latter character sup- icine enantiornithine (Varricchio and ports ambiguously the sister-group relation- Chiappe, in press), the distal end of metatar- ship between Yungavolucris brevipedalis and sal II is strongly twisted medially. This con- Avisauridae. On the other hand, the sister- dition is absent in Yungavolucris brevipedalis group relationship of the latter taxon and (figs. 2A, B, 3A) and Neuquenornis volans (fig. Lectavis bretincola, resulting in the funda- 9). In Lectavis bretincola the condition re- mental cladogram offigure 1 3B, is supported mains unknown. The distal end ofmetatarsal ambiguously by the excavated plantar surface II is not curved medially in Patagopteryx de- of the tarsometatarsus (character 3). The ab- ferrariisi and Mononykus olecranus. sence of the latter character is optimized as 10. Trochlea ofmetatarsal II broader than autapomorphic of both Y. brevipedalis and the trochleae ofmetatarsal III and IV: absent the Two Medicine Formation form in the (0) or present (1). In Yungavolucris brevipe- remaining trees (fig. 1 3A, C). dalis (figs. 2A, B, F, 3A, C), Soroavisaurus Neuquenornis volans, soroavisaurus aus- australis (figs. 7A, B, F, 11), Avisaurus ar- tralis, Avisaurus archibaldi, and the Two chibaldi (figs. 11, 1 2B), Neuquenornis volans Medicine enantiomithine are clustered in the (Chiappe and Calvo, in press), and the Two monophyletic Avisauridae in all resultant Medicine Formation form (Varricchio and cladograms (fig. 13). The monophyly of the Chiappe, in press) the trochlea of metatarsal Avisauridae is supported unambiguously by II is the broadest trochlea of the tarsometa- the presence ofa strong plantar projection of tarsus. This condition is uncertain in Lectavis the medial rim of the trochlea of metatarsal bretincola. Otherwise, the trochlea of meta- III (character 2), and the transverse convexity tarsal II is not broader than the trochlea of ofthe dorsal surface ofthe mid-shaft ofmeta- metatarsal III in Patagopteryx deferrariisi and tarsal III (character 8) in the cladograms of Mononykus olecranus. figure 1 3A and C. Character 2 has an ambiguous optimization in the remaining cladogram (fig. 13B) in which, pending determi- PHYLOGENETIC DISCUSSION nation ofthe condition in Lectavis bretincola, it may diagnose a more inclusive clade. In Three most parsimonious cladograms addition, the presence of a J-shaped meta- (length 12, consistency index 0.83, retention tarsal I (character 5) is an ambiguous syna- index 0.87) resulted from the cladistic anal- pomorphy ofAvisauridae in all the resultant ysis of tarsometatarsal characters (fig. 13). cladograms (fig. 13). Pending determination These three fundamental cladograms differ ofthe condition in L. bretincola and Y. brevi- in depicting the relationships among Lectavis pedalis, the latter character may represent a bretincola, Yungavolucris brevipedalis, and the more inclusive synapomorphy. Avisauridae, a group composed of Neuque- Within the Avisauridae, Neuquenornis vo- nornis volans, Soroavisaurus australis, Avi- lans is the sister group ofthe clade composed saurus archibaldi, and the Two Medicine by Soroavisaurus australis, Avisaurus archi- Formation form. In two of these cladograms baldi, and the Two Medicine Formation form (fig. 1 3A, B), L. bretincola and Y. brevipedalis (fig. 13). In all resultant cladograms, the latter 26 AMERICAN MUSEUM NOVITATES NO. 3083 three taxa are united by one unambiguous ter-group relationship ofthe only two known synapomorphy: the strong medial curvature North American avisaurids (Avisaurus ar- of the distal end of metatarsal II (character chibaldi and the Two Medicine Formation 9). In addition, this clade is supported am- form). It finally corroborates previous claims biguously by another synapomorphy: the about the avian affinity of Avisaurus archi- possession ofa dorsally inclined proximal ar- baldi (Martin 1991; Chiappe, 1 992a), a taxon ticular surface (character 7). Pending deter- that was originally regarded as a nonavian mination in N. volans, this character may be theropod (Brett-Surman and Paul, 1985; a synapomorphy ofAvisauridae. S. australis, Norman, 1990). in turn, is the sister taxon ofthe clade formed by A. archibaldi and the unnamed form from ACKNOWLEDGMENTS the Two Medicine Formation (fig. 13). The latter clade is unambiguously diagnosed by I am especially grateful to Jaime Powell the possession ofa medially concave trochlea (Instituto-Fundacion "Miguel Lillo," Uni- of metatarsal IV (character 5). versidad Nacional de Tucuman) for letting me study the specimens here described. I also CONCLUSIONS thank Jose Bonaparte (Museo Argentino de Ciencias Naturales), Jorge Calvo (Museo de Yungavolucris brevipedalis, Lectavis bre- Ciencias Naturales, Universidad Nacional del tincola, and Soroavisaurus australis are three Comahue), Mark Norell (AMNH), John Os- diagnosable taxa possessing synapomorphies trom (Yale Peabody Museum), Altangarel that cluster them in the monophyletic En- Perle (Mongolian Museum of Natural His- antiornithes. The recognition of these three tory), Jose Luis Sanz (Universidad Autono- new enantiornithine species, in addition to a ma de Madrid), and David Varricchio (Mu- worldwide distribution (see above), empha- seum of the Rockies) for making available sizes the important role played by this group for study material under their custody. I am in the Late Cretaceous avifauna. grateful to James Clark (AMNH) and David The proposed phylogenetic hypothesis is Varricchio who made very useful comments the first attempt to understand the interre- on the manuscript, and to Mark Norell, Storrs lationships of the Late Cretaceous enantior- Olson (Smithsonian Institution), and Larry nithines. This hypothesis clearly indicates the Witmer (New York College of Osteopathic nonmonophyletic status of the enantiorni- Medicine) for their critical reviews. Photo- thine assemblage of El Brete, with some el- graphs were skillfully made by Frank Ippolito ements (e.g., Soroavisaurus australis) sharing (AMNH). This research was supported by a a most recent common ancestor with North Frick Research Fellowship of the American American taxa. It also supports the mono- Museum of Natural History. phyly ofthe Avisauridae, establishing the sis- REFERENCES Alvarenga, H. M. F., and J. F. Bonaparte Brete, northwestern Argentina (Saurop- 1992. A new flightless land bird from the Cre- oda-Coelurosauria-Carnosauria-Aves). taceous ofPatagonia. In K. E. Campbell Mem. Soc. Geol. France (n. ser.) 139: (ed.), Papers in avian paleontology, 19-28. honoring Pierce Brodkorb, pp. 51-64. Bonaparte, J. F., J. A. Salfitty, G. Bossi, and J. E. Los Angeles: Natural History Museum Powell of Los Angeles County, Science Series 1977. Hallazgos de dinosaurios y aves creta- 36. cicas en la Formaci6n Lecho de El Brete Baumel, J. J, A. S. King, A. M. Lucas, J. E. Brea- (Salta), pr6ximo al limite con Tucuman. zile, and H. E. Evans Acta Geol. Lilloana 14: 5-17. 1979. Nomina Anatomica Avium. London: Brett-Surman, M. K., and G. Paul Academic Press, 637 pp. 1985. A new family of bird-like dinosaurs Bonaparte, J. F., and J. E. Powell linking Laurasia and Gondwanaland. J. 1980. A continental assemblage of tetrapods Vertebr. Paleontol. 5(2): 133-138. from the Upper Cretaceous beds of El Carpenter, J. M. 1993 CHIAPPE: CRETACEOUS TARSOMETATARSI 27

1988. Choosing among multiple equally par- 1991. Mesozoic birds and the origin of birds. simonious cladograms. Cladistics 4: In H-P. Schultze and L. Trueb (eds.), 291-296. Origins of the higher groups of tetra- Chiappe, L. M. pods, pp. 485-540. Ithaca: Comstock. 1991a. Cretaceous avian remains from Pata- Molnar, R. E. gonia shed new light on the early radi- 1986. An enantiornithine bird from the Lower ation ofbirds. Alcheringa 15(3-4): 333- Cretaceous of Queensland, Australia. 338. Nature 322: 736-738. 1991 b. Cretaceous birds of Latin-America. Nessov, L. A. Cretaceous Res. 12: 55-63. 1984. Pterosaurs and birds of the Late Cre- 1992a. Enantiornithine tarsometatarsi and the taceous of Central Asia. Paleontol. Zh. avian affinity of the Late Cretaceous 1: 47-57 [In Russian]. Avisauridae. J. Vertebr. Paleontol. 12(3): Nessov, L. A., and A. A. Jarkov 344-350. 1989. New Cretaceous-Paleogene birds of the 1992b. Osteologia y sistematica de Patagopte- USSR and some remarks on the origin ryx deferrariisi Alvarenga y Bonaparte and evolution of the Class Aves. Proc. (Aves) del Cretacico de Patagonia. Filo- Zool. Inst., Leningrad, 197: 78-97 [In genia e historia biogeogrifica de las aves Russian]. Cretacicas de America del Sur. Ph.D. Norman, D. B. diss., Universidad de Buenos Aires, 1990. Problematic Theropoda: "Coeluro- Buenos Aires, 429 pp. saurs." In D. B. Weishampel, P. Dod- In press. Phylogenetic position of the Creta- son, and H. Osm6lska (eds.), The Di- ceous birds of Argentina: Enantior- nosauria, pp. 280-305. Berkeley: Univ. nithes and Patagopteryx deferrariisi. In California Press. D. S. Peters (ed.), Proc. 3rd Symp. Soc. Perle, A., M. A. Norell, L. M. Chiappe, and J. M. Avian Paleontology and Evolution. Clark Frankfurt. 1993a. Flightless bird from the Cretaceous of Chiappe, L. M., and J. 0. Calvo Mongolia. Nature 362:623-626. In press. Neuquenornis volans, a new Upper 1993b. Correction. Flightless bird from the Cretaceous bird (Enantiornithes: Avi- Cretaceous of Mongolia. Nature 363: sauridae) from Patagonia, Argentina. J. 188. Vertebr. Paleontol. Sanz, J. L., and A. Buscalioni De Queiroz, K., and J. Gauthier 1992. A new bird from the Early Cretaceous 1990. Phylogeny as a central principle in tax- of Las Hoyas, Spain, and the Early ra- onomy: phylogenetic definitions of tax- diation of birds. Paleontology 35(4): on names. Syst. Zool. 39: 307-322. 829-845. 1992. Phylogenetic taxonomy. Annu. Rev. Sanz, J. L., J. F. Bonaparte, L. M. Chiappe, and Ecol. Syst. 23: 449-480. A. Buscalioni Farris, J. In press. La posici6n filogenetica de Concornis 1988. Hennig86 references. Documentation lacustris del Cretacico inferior de Las for version 1.5. Privately published. Hoyas, Espania. X Jorn. Arg. Paleontol. Gauthier, J. A. Vert., Ameghiniana. 1986. Saurischian monophyly and the origin Varricchio, D. J., and L. M. Chiappe of birds. In K. Padian (ed.), The origin In press. A new bird from the Cretaceous Two ofbirds and the evolution offlight. Mem. Medicine Formation of Montana. J. Calif. Acad. Sci. 8: 1-55. Vertebr. Paleontol. Gomez Omil, R. J., A. Boll, and R. M. Hernandez Walker, C. A. 1989. Cuenca Cretacico-Terciaria del No- 1981. New subclass of birds from the Creta- roeste Argentino (Grupo Salta). In G. ceous of South America. Nature 292: A. Chebli and L. A. Spalletti (eds.), 51-53. Cuencas sedimentarias argentinas. Ser. Wiley, E. O., D. Siegel-Causey, D. R. Brooks, and Correl. Geol. 6: 43-64. V. A. Funk Martin, L. D. 1991. The compleat cladist. Univ. Kansas 1983. The origin and early radiation of birds. Mus. Nat. Hist., Spec. Publ. 19: 1-158. In A. H. Bush and G. A. Clark, Jr. (eds.), Zhou, Z., F. Jin, and J. Zhang Perspectives in ornithology, pp. 291- 1992. Preliminary report on a Mesozoic bird 338. Cambridge: Cambridge University from Liaoning, China. Chinese Sci. Bull. Press. 37(16): 1365-1368. Recent issues of the Novitates may be purchased from the Museum. Lists of back issues of the Novitates, Bulletin, and Anthropological Papers published during the last five years are available free of charge. Address orders to: American Museum of Natural History Library, Department D, Central Park West at 79th St., New York, N.Y. 10024.

0) This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper).