americanmuseum noqvitates - american museum of natural history

AMERICAN MUSEUM

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

Enantiornithine (Aves) Tarsometatarsi from theCretaceous Lecho Formation of

Northwestern Argentina

LUIS M. CHIAPPE1

ABSTRACT

Enantiornithine tarsometatarsi from the Maas-trichtian Lecho Formation at the locality of ElBrete (northwestern Argentina) are described to-gether with their associated material. Three newspecies, namely Yungavolucris brevipedalis, Lec-tavis bretincola, and Soroavisaurus australis, arerecognized. Y. brevipedalis is distinguished by hav-ing a short and broad tarsometatarsus with a pul-leylike trochlea of metatarsal II, and equally longmetatarsals III and IV with the distal end of theformer laterally curved. The long and slender tar-sometatarsus of L. bretincola is characterized,among other features, for bearing a hypotarsus de-veloped primarily over the metatarsal II. In turn,S. australis exhibits a long and narrow fenestrabetween the proximal halves ofmetatarsals III andIV, and the plantar surface of the proximal half of

metatarsal II forms a sharp edge (convergent withthose of L. bretincola).The relationships among these three species, and

their interrelationships with respect to other LateCretaceous enantiornithine taxa, are explored. Acharacter analysis based on tarsometatarsal fea-tures is presented. This analysis supports the hy-pothesis that Soroavisaurus australis is the sistergroup of a clade formed by Avisaurus archibaldiand a new form from the Two Medicine Forma-tion of Montana. This clade is in turn the sistergroup of Neuquenornis volans. These four taxacompose the monophyletic taxon Avisauridae. Inthe present analysis, the relationships among Lec-tavis bretincola, Yungavolucris brevipedalis, andAvisauridae remain unresolved.

RESUMEN

Se describen los tarsometatarsos, y material aso-ciado, de Enantiornithes del Maastrichtiano de la -

Formacion Lecho en la localidad de El Brete (no-roeste de Argentina). Se reconocen y nombran tres

nuevas especies: Yungavolucris brevipedalis, Lec-tavis bretincola, y Soroavisaurus australis. Y. bre-vipedalis se distingue por poseer un tarsometatarsocorto 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

AMERICAN MUSEUM NOVITATES

de polea, los metatarsos III y IV de longitud equi-valente, y el extremo distal del metatarso III cur-vado lateralmente. El delgado y largo tarsometa-tarso de L. bretincola se caracteriza, entre otrosrasgos, por poseer un hipotarso desarrollado prin-cipalmente sobre el metatarso II. A su vez, S. aus-tralis exhibe una larga y angosta fenestra entre lasmitades proximales de los metatarso III y IV, y lasuperficie plantar de la mitad proximal del me-tatarso II formando un borde filoso (convergentecon aquel de L. bretincola).

Se exploran las relaciones filogeneticas entre es-tas tres especies, y sus relaciones con respecto a

otros Enantiornithes del Cretacico tardio. Se pre-senta un analisis de caracteres basado en caracteresdel tarsometatarso. Este an'alisis sustenta la hi-potesis que Soroavisaurus australis es el grupo her-mano de un clado formado por Avisaurus archi-baldi y una nueva forma de la Formacion TwoMedicine de Montana. Este clado es a su vez elgrupo hermano de Neuquenornis volans. Estoscuatro taxones comprenden el taxon monofileticoAvisauridae. El presente an'alisis no resuelve lasrelaciones entre Lectavis bretincola, Yungavolucrisbrevipedalis y Avisauridae.

INTRODUCTION

The Enantiornithes is a Cretaceous groupof volant birds originally recognized byWalker (1981) on the basis of the large as-semblage of bones from the MaastrichtianLecho Formation at the locality of El Brete(southern Province of Salta, northwesternArgentina) (fig. 1). In this paper Walker il-lustrated several bones from El Brete, in-cluding the three types oftarsometatarsi heredescribed, and briefly remarked on the pe-culiar anatomy of the group. After the erec-tion of the Enantiornithes, several authorsassigned previously described forms (Martin,1983) as well as new taxa (Molnar, 1986; Nes-sov and Jarkov, 1989; Chiappe, 1991a) tothis clade. To date, enantiornithine birds arerecorded in the Lower Cretaceous of Austra-lia (Molnar, 1986) and Spain (Sanz et al., inpress), and the Upper Cretaceous of SouthAmerica (Walker, 1981; Chiappe, 1991a,1991b), North America (Martin, 1983;Chiappe, 1992a; Varricchio and Chiappe, inpress), and Asia (Martin, 1983; Nessov, 1984;Nessov and Jarkov, 1989). Furthermore, aputative member of this clade was recentlydiscovered in the Lower Cretaceous ofChina(Zhou et al., 1992).The discovery of the Enantiornithes

strongly influenced subsequent interpretationof early avian evolution. Nevertheless, theirmost significant evidence, the large assem-blage of bones from El Brete, still remainsmostly unstudied. Interpretation of the ElBrete enantiornithines is complex due to thepoor association among the different speci-mens. Few bones were found in articulation,and no association between hind and fore-

limb elements is known (contra Walker,1981).In this paper the enantiornithine tarso-

metatarsi from El Brete, and the material as-sociated with them, are described. Three newspecies are recognized and their phylogeneticrelationships analyzed. The rest ofthe El Bretecollection will be described elsewhere.

INSTITUTIONAL ABBREVIATIONSAMNH American Museum of Natural HistoryGI Geological Institute, Mongolian Academy of

Sciences, Ulaan BaatarLH Universidad Aut6noma, MadridMACN Museo Argentino de Ciencias Naturales, Bue-

nos AiresMOR Museum of the Rockies, BozemanMUCPv Museo de Ciencias Naturales, Universidad

Nacional del Comahue, NeuquenPU Princeton University, PrincetonPVL Fundaci6n-Instituto Miguel Lillo, TucumatnQM Queensland Museum, South BrisbaneUCMP Museum of Paleontology, University of Cal-

ifornia, BerkeleyYPM Yale Peabody Museum, New Haven.

LOCALITY AND GEOLOGICAL SETTINGThe material described here was found in

a small quarry (approximately 8 m wide) ofthe middle section of the Lecho Formationin the Estancia El Brete, in the southern tipofthe Argentine province ofSalta (fig. 1). TheEstancia El Brete is located within the dry,forested, hilly landscape of the southern partof the phytogeographic province of the Yun-ga. The first excavations ofthis quarry startedin 1975, when a crew ofthe Fundacion-Insti-

2 NO. 3083

CHIAPPE: CRETACEOUS TARSOMETATARSI

65O21' 65'20'W

£~~0.5kmN

EX Subgroup Santa BarbaraSubgroup 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 theEstancia El Brete and the quarry (black circle) of the Lecho Formation in where the specimens werefound. The label "El Brete" shows the farm houses of the Estancia.

tuto Miguel Lillo (Tucuman, Argentina) dis-covered the fossiliferous locality, which wasreported by Bonaparte et al. (1977).The Lecho Formation is part ofthe Upper

Cretaceous Balbuena Subgroup (Salta Group),a near-border stratigraphical unit of the An-dean sedimentary basin (Bonaparte et al.,1977; Bonaparte and Powell, 1980). Recentstratigraphical studies defined the Lecho For-mation as those clastic lithofacies of the Bal-buena Subgroup deposited in subaerial, flu-vial, and eolian environments (Gomez Omilet al., 1989). Lithologically, the Lecho For-mation consists of mainly reddish, fine tomedium, with an occasional coarse, sand-stone (Bonaparte et al., 1977).The avian assemblage from El Brete comes

from fine-grained sandstones. The vertebratefauna associated with the avian remains iscomposed of several specimens of the ar-mored titanosaurid Saltasaurus loricatus, thesmall nonavian theropod Noasaurus leali, andindeterminate nonavian theropod teeth(Bonaparte and Powell, 1980).

Bonaparte et al. (1977) suggested that thepaleoenvironment of the Lecho Formationin the area ofEl Brete was a fluvial-lacustrine

coastal plain, with abundant vegetation andponds. The age of the Lecho Formation isgenerally regarded as Maastrichtian on thebasis of its fauna (Bonaparte et al., 1977) andstratigraphic relationships (Go6mez Omil etal., 1989).

MATERIALS AND METHODS

Anatomical nomenclature follows Baumelet al. (1979), using the English equivalents ofthe Latin terminology. In regard to the tax-onomic names here applied, the term Avesis used in reference to a group including alltaxa descended from the most recent com-mon ancestor ofArchaeopteryx lithographicaand modem birds. It is relevant to mention,however, that this application has been re-cently challenged by new theoretical asser-tions restricting the name of widely knowntaxa to the crown group (De Queiroz andGauthier, 1990, 1992). Under this new in-terpretation, the taxon name Avialae (Gau-thier, 1986) replaces Aves, and the latter sub-stitutes the taxon name Neomithes of theclassical avian taxonomy.The three species erected in this paper are

based on incomplete material, which is still

l1993 3


sufficiently diagnostic. It should be noted thatthese are not the first enantiomithine speciesnamed from El Brete. Walker (1981) appliedthe name Enantiornis leali to the specimenPVL-4035 that includes a proximal end ofthe humerus articulated to the coracoid andscapula. The status ofthis species name, how-ever, is dubious because Walker (1981) didnot describe this specimen and the speciesname was merely listed in the legend to atable. Because of this, and the additionalproblems due to the lack of association be-tween forelimb and hind limb elements, thisauthor believes that the tarsometatarsi de-scribed in this paper must be allocated intheir own species instead of waiting for newmaterial with association of limb elements.The enantiomithine material studied in-

cludes all known tarsometatarsi (and mate-rial associated with them) from El Brete (PVL-402 1-1, PVL-4040, PVL-4048, PVL-4052,PVL-4053, PVL-4268, PVL-4690, PVL-4692); the enantiomithine tibiotarsi PVL-4030, PVL-4032-1, PVL-4033, PVL-4036,PVL-4695, PVL-4696 from El Brete, and acast ofthe holotype (QM F12992) ofNanan-tius eos (Molnar, 1986); the holotype (LH-2814) of Concornis lacustris (Sanz and Bus-calioni, 1992), recently reinterpreted as anenantiornithine bird (Sanz et al., in press);the holotype (MUCPv- 142) ofNeuquenornisvolans (Chiappe and Calvo, in press); a cast(MACN- 18685) ofthe holotype and the para-type PU- 17324 (now housed at the YPM) ofAvisaurus archibaldi (Brett-Surman and Paul,1985), whose enantiomithine kinship was re-cently established (Chiappe, 1992a); and theholotype (MOR-553E) of a new avisauridspecies from the Two Medicine Formation(Varricchio and Chiappe, in press).Polarization of character states used in the

character analysis was established by usingPatagopteryx deferrariisi (Alvarenga andBonaparte, 1992), and Mononykus olecranus(Perle et al., 1993a). The Jurassic bird Ar-chaeopteryx lithographica was not selectedbecause its two-dimensional preservationobscures several features. The sister-grouprelationship between Enantiomithes and aclade formed by P. deferrariisi plus the Or-nithurae has been recently established by thisauthor (Chiappe, 1992b, in press). Under thisframework, P. deferrariisi is an adequate out-

group for analyzing the enantiomithine in-terrelationships. M. olecranus (see Perle etal., 1993b for nomenclatural correction) hasbeen recently interpreted as the sister groupof all known birds except A. lithographica(Perle et al., 1993a). The outgroup materialstudied includes specimens MACN-N-03,MACN-N-10, MACN-N-Il, and MUCPv-48 ofP. deferrariisi (Chiappe, 1992b), and GIN107/6 and GI N100/99 of M. olecranus(Perle et al., 1993a).Derived characters exclusive to a single in-

group taxa were omitted from the analysis inorder to avoid their influence on the consis-tency index (Carpenter, 1988; Wiley et al.,1991). The data matrix presented in table 4was processed using the implicit enumeration(ie) command of Hennig86 program (Farris,1988).

SYSTEMATIC PALEONTOLOGY

Aves Linnaeus, 1758

Enantiomithes Walker, 1981

Yungavolucris brevipedalis, new speciesDiAGNosIs: Enantiomithine bird with short

and broad tarsometatarsus possessing the fol-lowing autapomorphies: distal end muchbroader than proximal end; broad, dorso-plantarly compressed, and pulleylike trochleaof metatarsal II; distal end of metatarsal IIIlaterally curved; metatarsal IV equal in lengthto metatarsal III; prominent dorsal ridge be-tween distal halves of metatarsals II and III.ETYMOLOGY: Yungavolucris from Latin

volucris, meaning "bird," and Yunga refer-ring to the phytogeographic region in whichEl Brete is located; brevipedalis from Latinbrevis, "short," and pedalis referring to thefoot.HOLOTYPE: PVL-4053 (fig. 2); nearly com-

plete right tarsometatarsus, lacking the cra-niolateral border of the lateral cotyla and thecranial portion of the trochlea of metatarsalIV.REFERRED SPECIMENS: Incomplete right

tarsometatarsi PVL-4040 and PVL-4692 (fig.3), incomplete left tarsometatarsus PVL-4052(fig. 4), and specimen PVL-4268 representedby the distal trochleae of right metatarsals IIand III.

NO. 30834


TABLE 1Measurements (in mm) of Yungavolucris brevipedalis, new species

PVL specimens4040 4052 4053 4268 4692

Total length, tarsometatarsus 41.6 51.9 41.9 - 42.8Total length, metatarsal II 37.8 41.5 37.0 - 38.4Maximum width, proximal endof tarsometatarsus 16.2 15.3 14.0 - 15.0

Maximum width, trochleametatarsals II and III 19.6 20.4 19.4 20.0 19.7

Maximum width, trochleametatarsal II 11.8 10.9 10.4 12.1 11.0

Maximum width, trochleametatarsal III 7.2 6.8 6.2 6.3 6.5

LOCALITY AND HORIZON: 26OO2'07 "S65019'57"W. Estancia El Brete, Departmentof Candelaria, Province of Salta, Argentina(fig. 1). Lecho Formation, Upper Cretaceous(Maastrichtian).

DESCRIPTION: The tarsometatarsus of Yun-gavolucris brevipedalis is short and stout. Itis dorsoplantarly compressed, being generallyconvex dorsally and flat plantarly (fig. 2). Itsdistal end is much broader than its proximalend. Metatarsals II, III, and IV are firmlyattached to each other, although they are fusedonly in their proximal portion (figs. 2A, B,3A).

In the proximal end, the medial cotyla iswell developed and suboval; the lateral cotylais less defined and its articular surface isslightly convex medially (figs. 2E, 3B, 4). Bothan intercondylar eminence and hypotarsusare absent.

Metatarsal II is the most robust ofthe threemetatarsals, but not the longest. It is straightthroughout its length (figs. 2A, B, 3A). Dor-sally, in its middle portion it exhibits a verypronounced ovate-shaped tubercle, with itslong axis oriented proximodistally (figs. 2A,C, 3A). This tubercle was tentatively inter-preted as the area of attachment of the M.tibialis cranialis (Brett-Surman and Paul,1985; Chiappe, 1992a). The distal halfof thismetatarsal is fairly compressed dorsoplan-tarly, especially in its medial portion. In thelateral portion of the distal half, its contactwith metatarsal III forms a prominent ridge(figs. 2A, C, D, 3A). The distal end of meta-

tarsal II is very broad. The trochlea of thismetatarsal is ample, dorsoplantarly com-pressed and pulleylike (figs. 2F, 3C). Bothrims of the trochlea are bulbous and are sep-arated from each other by a broad centralfurrow. Just proximal to the medial rim ofthe trochlea, on the medial margin ofthe dor-sal face, there is a dorsomedial projection (fig.3A). The degree of development and orien-tation ofthis projection varies among the dif-ferent specimens (cf. figs. 2A, 3A).

Metatarsals III and IV are bent laterallyover their distal third, and they are muchlonger than metatarsal II (figs. 2A, B, 3A). Inthe dorsal face ofthe metatarsal III, just distalto the proximal end, there is a slender andelongated ridge for muscle attachment (fig.2A). Distally, in spite of the lateral bendingof the shaft, the trochlea of metatarsal III isin a vertical position (figs. 2A, B, F, 3A, C).The medial rim of this trochlea is larger andmore bulbous than the lateral (fig. 3A, B),and deep collateral ligamental fossae exca-vate both sides.

Metatarsal IV is very small with respect tometatarsals III and II (figs. 2A, B, 3A). It ismore weakly attached to metatarsal III thanthis one is to metatarsal II. The trochlea ismissing in all available specimens except inthe holotype where only the plantar half ispreserved. It is simple and laterally com-pressed (fig. 2A, B, D, F).REMARKS: Comparisons between the tar-

sometatarsi of Yungavolucris brevipedalis andthose of the remaining enantiornithine taxa

1 993 5


Fig. 2. Holotype specimen (PVL-4053) of Yungavolucris brevipedalis, n. sp. Right tarsometatarsusin dorsal (A), plantar (B), medial (C), lateral (D), proximal (E), and distal (F) views. dp dorsomedialprojection of the distal end of metatarsal II, dr distal ridge between the contact of metatarsals II andIII, tu tubercle for the attachment of the M. tibialis cranialis, lc lateral cotyla, mc medial cotyla, rmelongate proximal ridge for muscle attachment, II-IV metatarsals TI-IV.

and the outgroup forms indicate several au-tapomorphies diagnosing the former taxon.The remarkable difference between the

width of the proximal and distal ends of the

tarsometatarsus in Yungavolucris brevipeda-lis is unique to this species among the En-antiornithes (cf. figs. 2A, 7A, 12A). In So-roavisaurus australis, Avisaurus archibaldi,

6 NO. 3083


A

rm

IV III

C

IIII

D

B

III IV

E

mc

F

IV III 11

I

--MM J o m

cmsFig. 2. Continued.

Neuquenornis volans, Concornis lacustris, andthe Two Medicine Formation form, the distalend is only slightly broader than the proximal

end. This condition is unknown in Lectavisbretincola. In the outgroups, the differencebetween the width of the proximal and distal

71993


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 forthe attachment of the M. tibialis cranialis, lc lateral cotyla, mc medial cotyla, rm elongate proximalridge for muscle attachment, II-IV metatarsals II-IV.

ends ofmetatarsals II-IV is also considerablysmaller than in Y. brevipedalis. The signifi-cant difference in the width of the two endsof the tarsometatarsus is autapomorphic forY. brevipedalis.The tarsometatarsus of Yungavolucris

brevipedalis differs from all other known en-antiomithine tarsometatarsi (the condition isunknown in Lectavis bretincola) in that thedistal end of metatarsal III curves laterally(figs. 2A, B, 3A), instead of being straight. Astraight metatarsal III is also present in theoutgroup species Patagopteryx deferrariisi andMononykus olecranus. This feature is inter-preted as an autapomorphy of Y. brevipe-dalis.The presence in Yungavolucris brevipedalis

of a prominent ridge on the distal half of thedorsal surface ofthe tarsometatarsus betweenthe union of the metatarsals II and III is ab-sent in Lectavis bretincola (fig. 6A), Soroavi-saurus australis (figs. 7A, 8A, 9), Neuquenor-nis volans (fig. 9), Avisaurus archibaldi (fig.1 2A), Concornis lacustris, and the Two Med-icine Formation form. The absence of this

feature in the outgroups indicates that it isanother autapomorphic character of Y. brevi-pedalis.

In contrast to Soroavisaurus australis (figs.7A, B, 8), Avisaurus archibaldi (fig. 12), Con-cornis lacustris, and the Two Medicine For-mation form, metatarsal IV of Yungavolucrisbrevipedalis is as long as, not shorter than,metatarsal III (fig. 2A, B). The condition re-mains unknown in Neuquenornis volans andLectavis bretincola in which the distal end ofthis metatarsal is missing. Metatarsal IV isshorter than metatarsal III in the outgroupforms. The subequal length ofmetatarsals IIIand IV is also regarded as an autapomorphyof Y. brevipedalis.The tarsometatarsus of Yungavolucris

brevipedalis also differs from the tarsometa-tarsus ofSoroavisaurus australis (figs. 7F, 1 1),Avisaurus archibaldi (fig. 11), Neuquenornisvolans, Concornis lacustris, and the TwoMedicine Formation form in having a re-markably broad, dorsoplantarly compressed,and pulleylike trochlea of metatarsal II (figs.2F, 3C). In Lectavis bretincola the condition

8 NO. 3083


A


is unknown. This peculiar morphology is alsoabsent in the outgroups, being here inter-preted as another autapomorphy of Y. brevi-pedalis.

Lectavis bretincola, new species

DIAGNOSIS: Enantiomithine bird with longand slender tibiotarsus and tarsometatarsuspossessing the following autapomorphies:

suboval articular surface of proximal end oftibiotarsus; distal tibiotarsal condyles strong-ly projected cranially; plantar surface ofprox-imal half of metatarsal II forming a promi-nent edge; subcircular knob on proximal endof metatarsal II; hypotarsus mostly devel-oped over metatarsal II.

ETYMOLOGY: Lectavis from Latin lectus,equivalent to the the Spanish word lecho("bed"), in reference to the Lecho Formation,

mc Ic

Fig. 4. Left tarsometatarsus (PVL-4052) of Yungavolucris brevipedalis,lateral cotyla, mc medial cotyla.

cmsn. sp. in proximal view. Ic

B

Icmc

IV

III

C

III 11

91993


TABLE 2Measurements (in mm) of Lectavis bretincola, new species

Specimen PVL-4021-1Maximum preserved length, tarsometatarsus 88.0Maximum width, proximal end of tarsometatarsus 15.0Maximum depth, proximal end of tarsometatarsus 10.7Total length, tibiotarsus 156.0Maximum depth, mid-shaft of tibiotarsus 4.8Maximum width, distal end of tibiotarsus 14.3Maximum depth, lateral condyle of tibiotarsus 12.7Maximum depth, medial condyle of tibiotarsus 14.3

and avis, "bird"; bretincola from the com-bination of "Brete" and the Latin incola,which means "an inhabitant."HOLOTYPE: PVL-4021-1 (figs. 5, 6). Left

tibiotarsus and incomplete tarsometatarsus,missing the distal end of the latter and mostof metatarsal IV.LOCALITY AND HORIZON: 26002'07 "S

65°19'57"W. Estancia El Brete, Departmentof Candelaria, Province of Salta, Argentina(fig. 1). Lecho Formation, Upper Cretaceous(Maastrichtian).

DESCRIPTION: The tibiotarsus of Lectavisbretincola is poorly preserved. It is long, slen-der, and craniocaudally compressed (fig. 5).The morphology of the proximal end differswith respect to the other El Brete enantior-nithine tibiae (e. g., PVL-4032-1, PVL-4033,PVL-4036, PVL-4696), Concornis lacustris(Sanz and Buscalioni, 1992), and the Austra-lian enantiornithine Nanantius eos (Molnar,1986) in that in proximal view it is not cir-cular but ellipsoidal, with a lateromedial ma-jor axis. The shaft is compressed craniocau-dally (fig. 5). Distally, the condyles of thetibiotarsus of L. bretincola are strongly pro-jected cranially, much more than in any otherenantiomithine tibiotarsi from El Brete (e.g.,PVL-4030, PVL-4033, PVL-4695) and in N.eos. As in the remaining enantiomithines, themedial condyle is inflated, and considerablylarger than the lateral one (fig. 5A). The in-tercondylar notch separating these condylesis less pronounced and broader than in theremaining El Brete tibiotarsi and in N. eos.Well-developed epicondylar fossae are pres-ent on the sides of the condyles.Lectavis bretincola possesses a long and

slender tarsometatarsus (fig. 6). Metatarsals

II-IV are strongly attached to each other overtheir length, and they are completely fusedproximally (fig. 6A). The fusion betweenmetatarsals II and III starts more distally (ap-proximately 24.7 mm from the proximal end)than does the fusion between metatarsals IIIand IV (approximately 17.3 mm from theproximal end). Dorsally, the tarsometatarsusof L. bretincola is in general flat, while inplantar view it is deeply excavated (fig. 6B,C). This excavation is especially pronouncedin the proximal portion, where metatarsal IIforms a strong ridge that medially borders adepressed area formed by metatarsals III andIV (fig. 6B, C).

In dorsal view metatarsal III is the broad-est, narrowing smoothly distally. Its proximalportion is dorsoplantarly thin while distallyit becomes thicker. Metatarsal II is the moreconspicuous of the three. Its main axis is inthe dorsoplantar plane, and plantarly it formsa prominent and robust ridge (fig. 6B, C, D),which attenuates distally. In the most prox-imal portion, this ridge is continuous with atriangular-shaped hypotarsus (fig. 6B). Incontrast to the typical modem avian condi-tion, the hypotarsus of Lectavis bretincola ismostly developed on the plantar surface ofmetatarsal II. The medial surface of thismetatarsal is basically flat, with only a slightlydepressed central area in its most proximalportion (fig. 6D). Dorsally, in the most prox-imal portion ofmetatarsal II, there is a prom-inent round tubercle. Above this tubercle, anearly crescent-shaped fossa is located be-tween metatarsals II and III (fig. 6A). Meta-tarsal IV is badly preserved. It is significantlysmaller than the remaining ones (fig. 6A, B,C). Its proximal portion is remarkably thin.

10 NO. 3083


A

mc

B

Ic mc

cmsFig. 5. Holotype specimen (PVL-402 1-1 ) of Lectavis 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 contacton the lateral border ofmetatarsal III, as withthe latter, metatarsal IV became thicker to-ward the distal end.Both cotylae in the proximal articular sur-

face are ovate in outline (fig. 6E), and their

major axes are oriented in a mediodorsal-lateroplantar direction. The medial cotyla isslightly larger than the lateral cotyla, and itsdorsal border is more elevated as well. Theplantar border of the medial cotyla, abovethe hypotarsus, is projected proximally (fig.

1 993 I1I


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 theM. tibialis cranialis, lc lateral cotyla, mc medial cotyla, II-IV metatarsals II-IV.

NO. 308312


B

11 IIIE

cms

mcFig. 6. Continued.

A C D

tu

I11

Ic

1993 13


6E). As was pointed out by Walker (1981),the hypotarsus has a subrectangular shape inproximal view, and lacks any crest or sulcus(fig. 6B, E).The distal end of the tarsometatarsus of

the single specimen of Lectavis bretincola ismissing. With the available material it is notpossible to estimate, even roughly, the ap-proximate length of this tarsometatarsus.REMARKS: Character-state comparisons

among Lectavis bretincola, the remaining en-antiornithines, and the outgroup forms revealseveral autapomorphic features ofthe formerspecies.

In the tibiotarsus, the suboval proximal endand the strong cranial projection ofthe distalcondyles differ from those of the remainingenantiornithines and the outgroups, and arehere considered autapomorphies of Lectavisbretincola.

In the tarsometatarsus, Lectavis bretincoladiffers from all other enantiornithine taxa inhaving a hypotarsus (fig. 6B). Contrary to thestatements ofBrett-Surman and Paul (1985),the tarsometatarsus of Soroavisaurus aus-tralis (figs. 7B, E, 8B) and Avisaurus archi-baldi (fig. 12B) lacks a hypotarsus (Chiappe,1992a), a condition also shared with Yun-gavolucris brevipedalis (figs. 2B, E, 3B, 4), theTwo Medicine Formation form, and the out-group forms. The presence of a hypotarsusin L. bretincola is here regarded as an aut-apomorphy of this taxon. As was mentionedabove, the hypotarsus of L. bretincola differsfrom the typical modem avian condition inthat it appears to be formed mainly overmetatarsal II, instead of capping mostly theproximal end ofmetatarsal III (Baumel et al.,1979). This indicates that the developmentof a plantar structure on the proximal end ofthe tarsometatarsus, a hypotarsus, arose in-dependently in modem birds and certain en-antiomithines.

Lectavis bretincola differs from other en-antiomithines in that the tubercle on meta-tarsal II is circular instead ofovate as in Yun-gavolucris brevipedalis (figs. 2A, 3A),Soroavisaurus australis (figs. 7A, 8A), Avi-saurus archibaldi (fig. 12A), and the TwoMedicine form. The round morphology ofthis tubercle in L. bretincola is interpreted asan autapomorphy.The tarsometatarsus of Lectavis bretincola

is distinct from the tarsometatarsi of Yun-gavolucris brevipedalis (fig. 2B, C), Neuquen-ornis volans, Avisaurus archibaldi (fig. 12B),and the Two Medicine form in that the prox-imal half forms a robust and laterally com-pressed medial edge that projects plantarly(fig. 6A, C, D). A medial ridgelike plantarprojection is also present in the tarsometa-tarsus of Soroavisaurus australis (figs. 7B, D,8B). However, this condition differs from thecondition present in L. bretincola in that theridge is less robust and more compressed lat-erally, and in that its most prominent area isnot located in the proximal end but moredistally (cf. figs. 6D, 7D). Furthermore, in L.bretincola the proximal region of this ridgereceives the contribution of metatarsal III,whereas it is exclusively formed by metatar-sal II in S. australis. This prominent plantarridge is completely absent in the outgroupforms. The important differences between theplantar ridges of the tarsometatarsus of L.bretincola and S. australis make a hypothesisof homology of these structures unlikely.Hence, the medial plantar ridge of L. bretin-cola is regarded as an autapomorphy.

AVISAURIDAE BRETT-SURMAN AND PAUL, 1985

DiAGNOSIS: Enantiomithine birds with astrong transverse convexity ofthe dorsal sur-face of the mid-shaft of metatarsal III, a dis-tinct plantar projection of the medial rim ofthe trochlea of metatarsal III, and a laterallycompressed J-shaped metatarsal I (seeChiappe, 1992a).

DEFINITION: Phylogenetically defined (seeDe Queiroz and Gauthier, 1990, 1992) as thecommon ancestor ofNeuquenornis volans andAvisaurus archibaldi plus all its descendants.INCLUDED SPECIES: Avisaurus archibaldi

(Brett-Surman and Paul, 1985), Neuquenor-nis volans (Chiappe and Calvo, in press), So-roavisaurus australis, n. sp. (this paper), un-named new species from the Two MedicineFormation (Varricchio and Chiappe, in press).

Soroavisaurus australis, new speciesDiAGNoSIS: Enantiomithine avisaurid with

the following autapomorphies: plantar sur-face ofproximal halfofmetatarsal II forminga sharp edge (convergent with Lectavis bre-

14 NO. 3083


tincola); long and narrow fenestra open be-tween the proximal halves of metatarsals IIIand IV; dorsally projected edge on the troch-lea for metatarsal IV.ETYMOLOGY: Soroavisaurus from Latin so-

ror, meaning "sister," and Avisaurus (Brett-Surman and Paul, 1985), referring to the sis-ter-group relationship inferred for these twotaxa; australis refers to the ocurrence of thisspecies in the Southern Hemisphere.HOLOTYPE: PVL-4690 (fig. 7), left tarso-

metatarsus. Specimen previously referred toas Avisaurus sp. (Brett-Surman and Paul,1985; Chiappe, 1992a; Chiappe and Calvo,in press).REFERRED SPECIMENS: PVL-4048, left tar-

sometatarsus, metatarsal I with proximal anddistal phalanges of digit I, and four inter-mediate phalanges. This specimen was alsoformerly identified as Avisaurus sp. (Brett-Surman and Paul, 1985; Chiappe, 1992a;Chiappe and Calvo, in press).LOCALITY AND HORIZON: 26°02'07' S

65019'57"W. Estancia El Brete, Departmentof Candelaria, Province of Salta, Argentina(fig. 1). Lecho Formation, Upper Cretaceous(Maastrichtian).

DESCRIPTION: The tarsometatarsus of So-roavisaurus australis is gracile (figs. 7, 8).Metatarsals II-IV are fused only proximally(figs. 7A, B, 8), and distally these metatarsalsare not as closely apposed as in Yungavolucrisbrevipedalis or Lectavis bretincola. The prox-imal end is transversely much broader thancranioplantarly. The plane of the proximalarticular surface is inclined dorsally (fig. 7C,D). In proximal view, the articular surface iskidney-shaped, with the concave part facingplantarly (fig. 7E). In this respect, S. australisdiffers from Avisaurus archibaldi, in whichthe proximal surface is more oval. The me-dial cotyla is subcircular and of approxi-mately the same size as the lateral one. Thearticular surface of the latter was a smoothconvexity of the medial portion (fig. 7E).

Metatarsal III is the largest. It is straightand dorsally broader than the other two (figs.7A, 8A). The cranial surface of this metatar-sal is convex, especially in the middle of theshaft. Plantarly, this metatarsal constitutesthe floor of a depression defined by the plan-tar projection of the proximal two-thirds ofmetatarsals II and IV (figs. 7B, 8B). Meta-

TABLE 3Measurements (in mm) of Soroavisaurus australis,

new species

PVL- PVL-4690 4048

Total length, tarsometatarsus 46.9 51.5Maximum width, proximal endof tarsometatarsus 10.0 11.3

Maximum depth, proximal endof tarsometatarsus 6.3 6.1

Total length, metatarsal II 44.2 48.9Total length, metatarsal IV 43.3 48.4Maximum width, mid-shaftof tarsometatarsus 9.9 9.6

Maximum width, distal endof tarsometatarsus 14.6 15.2

Maximum width, trochleametatarsal II 5.8 5.2

Maximum width, trochleametatarsal III 4.1 4.2

Maximum width, trochleametatarsal IV 4.0 4.3

Maximum length, metatarsal I - 12.2

tarsal II is more laterally compressed. Themiddle portion of its plantar surface forms aprominent edge that limits the medial borderof the abovementioned plantar depression(figs. 7B, D, 8B). In Avisaurus archibaldi, theplantar border of this metatarsal is slightlyprojected plantarly, but it does not form thisprominent ridge (fig. 12B). In the dorsal sur-face of the proximal third an oval tubercle isdeveloped. This tubercle is not as prominentas in A. archibaldi, and is located more prox-imally than in the latter species (figs. 7A, 8A).In medial view, the shaft of this metatarsalis slightly sigmoid, as best exhibited by theholotype specimen. Distally, metatarsal II ismore widely separated from metatarsal IIIthan the latter is from metatarsal IV (figs. 7A,8A). Metatarsal IV is weak and dorsoplan-tarly compressed, especially in its distal half.In its proximal half, it is extremely slenderand more bar-shaped, defining the lateralborder of an elongate fenestra separating itfrom metatarsal III (figs. 7A, B, 8). This fe-nestra is absent in all other enantiornithines.

Distally, the three trochleae are well de-veloped. As in all avisaurids, the trochlea formetatarsal III possesses a remarkable plantar

151 993


Fig. 7. Holotype specimen (PVL-4690) of Soroavisaurus australis, n. sp. Left tarsometatarsus indorsal (A), plantar (B), lateral (C), medial (D), proximal (E), and distal (F) views. fe fenestra betweenmetatarsals III and IV, tu tubercle for the attachment of the M. tibialis cranialis, Ic lateral cotyla, mcmedial cotyla, II-IV metatarsals II-IV.

NO. 3083


B

tu

III

Emc

c IC

F

I1 III IVt

L .cms

Fig. 7. Continued.

A

IIID

tu

C

I

4

171 993


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. tibialiscranialis, II-IV metatarsals II-IV.

projection of the internal rim (Chiappe,1992a). Collateral fossae are present on bothsides of this trochlea. The trochlea for meta-tarsal II is the broadest, and plantarly it ex-hibits lateral and medial prominences sepa-rated by a wide depression (figs. 7B, 8B, 11).The trochlea for metatarsal IV of Soroavi-saurus australis clearly differs from that ofAvisaurus archibaldi (fig. 11). In S. australisthis trochlea has a subtriangular outline indistal view, with a medial projection thatnearly encloses distally a tiny fenestra be-tween metatarsals III and VI (a structure sim-ilar to the distal vascular foramen of Orni-thurae) (figs. 7A, B, 8B). Conversely, in A.archibaldi this trochlea is crescentic in distalview, with a concave medial side. The smallfenestra between metatarsals III and IV in A.archibaldi is also enclosed distally by a me-dial projection of the latter trochlea, but theprojection is located much further dorsallythan in S. australis (fig. 1 1). Furthermore, the

trochlea ofmetatarsal II ofS. australis differsfrom that ofA. archibaldi in that it has a stoutcentral edge developed on the dorsal surfacethat is totally absent in A. archibaldi (fig. 1 1).The tarsometatarsus of PVL-4048 is as-

sociated with metatarsal I and the hallux, andfour intermediate phalanges. Unfortunately,this author was not able to find any recordof the articulation between these bones; theywere disarticulated, although with a singlenumber and within a single box. Comparisonwith the articulated foot ofNeuquenornis vo-lans (fig. 9) suggests that they all belong tothe same specimen and were probably dis-articulated during preparation. Metatarsal Iof Soroavisaurus australis is J-shaped, as inN. volans (fig. 9). It is strongly compressedlaterally and tapers proximally. As in N. vo-lans, although metatarsal I was not reversed(Fig. 9) the hallux was reverted. The halluxis formed by two large phalanges (fig. 10).The proximal phalanx is dorsoplantarly com-

18 NO. 3083


A

tu

II

III


pressed. On its distal head, the articular facetof this phalanx is large and extensively de-veloped on the plantar surface (fig. 10). Thissuggests that the main movement ofthe distalphalanx was in a plantar direction. Well-de-veloped collateral fossae are present on bothsides of the distal head. The distal phalanxis sharp, laterally compressed, and stronglyarched (fig. 10). Plantarly, it possesses a small,round flexor tubercle at its base.REMARKS: The two known specimens of

Soroavisaurus australis were previously re-

ferred to Avisaurus sp. (Brett-Surman andPaul, 1985; Chiappe, 1992a; Chiappe andCalvo, in press). The result of the presentcladistic analysis (see below), however, sup-ports their allocation to a different genus.

Soroavisaurus australis exhibits several au-tapomorphic features. The tarsometatarsus

differs from the tarsometatarsus of the re-maining enantiornithine taxa in the posses-sion of a long and narrow fenestra betweenthe proximal halves ofmetatarsals III and IV(figs. 7A, B, 8). This fenestra is also absentin outgroup taxa. The presence of a fenestrabetween metatarsals II and III is here inter-preted as an autapomorphy of S. australis.Another autapomorphic character of So-

roavisaurus autralis is the presence of a stoutand thick dorsal trochlear projection ofmeta-tarsal IV (figs. 7F, 11). This projection is notdeveloped in either Avisaurus archibaldi (fig.1 1), Concornis lacustris, or the new form fromthe Two Medicine Formation. In Yungavolu-cris brevipedalis the dorsal portion of thetrochlea for the metatarsal IV is missing.However, considering the remarkable reduc-tion of this trochlea (fig. 2), the presence of

B

1 993 19


III

III..-..A

III

Soroavisaurus australis Neuquenornis volansFig. 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 conditionis also uncertain in both Neuquenornis volansand Lectavis bretincola. A dorsal, thick, cen-tral ridge is absent in the trochlea of meta-tarsal IV in the outgroups.As was previously discussed (see the Re-

marks under Lectavis bretincola), the pres-ence of a prominent, sharp, and laterallycompressed ridge on the plantar surface ofmetatarsal II of Soroavisaurus australis (figs.7B, D, 8B) is considered an autapomorphy.

TABLE 4Data Matrix of 10 Characters (0 primitive, 1 de-

rived, ? uncertain) of the Tarsometatarsus.

Mononykus olecranus 0000000000Patagopteryx deferrariisi 0010000000Yungavolucris brevipedalis 1001??0001Lectavis bretincola 1?1 1??00??Avisaurus archibaldi 1111?11111Soroavisaurus australis o101 1 1Neuquenornis volans 1 1?1??101Two Medicine form 1 10?1 11111

CHARACTER ANALYSIS

The interrelationships ofEnantiomithes arevery difficult to evaluate because the El Bretespecimens are mostly disarticulated. There isno information about the association be-tween fore- and hind-limb elements, or theirassociation with the pelvis or axial skeleton.This problem is not overcome by the re-maining enantiomithine forms recovered indifferent parts of the world, which are mostlyfragmentary or still unpublished. For this rea-son, this paper presents a tentative approachto the relationships of Late Cretaceous en-antiomithines on the basis of the tarsometa-tarsal morphology alone.A cladistic analysis ofthe relationships be-

tween the new taxa Yungavolucris brevipe-dalis, Lectavis bretincola, Soroavisaurus aus-tralis, and Avisaurus archibaldi (Brett-Surmanand Paul, 1985), Neuquenornis volans(Chiappe and Calvo, in press), and a newform from the Two Medicine Formation ofMontana (Varricchio and Chiappe, in press)was performed. Ten binary characters (table4) were analyzed cladistically and their dis-

20 NO. 3083


aM p

cm

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

tribution among the abovementioned taxaand the outgroups is discussed below.

1. Metatarsal IV: large (0) or considerablysmaller (1). Metatarsal IV is considerablysmaller than metatarsals II and III in all en-antiornithine taxa considered in the analysis(figs. 2, 3, 6-8, 12; see also Walker, 1981;Chiappe, 1991a, 1992a). Variable sizes ofmetatarsal IV are present among the differenttaxa (cf. figs. 2A, 6A, 7A, 1 2A), although thesedifferences are not clear enough for scoringdistinct character states. This enantiomithinecondition is absent in both Patagopteryx de-ferrariisi and Mononykus olecranus.

A

I v III I

Soroavisaurus australis

2. Medial rim ofthe trochlea ofmetatarsalIII with a strong plantar projection: absent(0) or present (1). The medial rim of thetrochlea ofmetatarsal III is projected stronglyplantarly in Soroavisaurus australis (figs. 7F,11), Avisaurus archibaldi (fig. 11), Neuque-nornis volans (Chiappe and Calvo, in press),and the Two Medicine Formation form (Va-rricchio and Chiappe, in press). On the otherhand, this strong projection is absent in Yun-gavolucris brevipedalis (figs. 2F, 3C) and theoutgroups.

3. Plantar surface of tarsometatarsus ex-cavated: absent (0) or present (1). In Lectavis

B

Avisaurus archibaldiFig. 1 1. Comparisons between the distal end of the tarsometatarsus of Soroavisaurus australis (PVL-

4690) and Avisaurus archibaldi (cast MACN-18685). II-IV metatarsals II-IV.

1 993 21


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 ofthe M. tibialis cranialis, II-IV metatarsals II-IV.

bretincola (fig. 6B, C), Soroavisaurus australis(figs. 7B, 8B), Avisaurus archibaldi (fig. 1 2B),and Neuquenornis volans (Chiappe and Cal-vo, in press) the plantar face of the tarso-metatarsus is excavated to variable degreesexhibiting a proximodistal fossa bounded bymetatarsals II and IV, giving a crescentic crosssection to the tarsometatarsus. Within the re-maining enantiornithine taxa considered here,this fossa is absent in Yungavolucris brevi-pedalis (fig. 2B) and the new form from theTwo Medicine Formation (Varricchio andChiappe, in press). Within the outgroup, thisfossa is absent in Mononykus olecranus, butthe tarsometatarsus of Patagopteryx defer-rariisi is excavated plantarly.

4. Tubercle on the dorsal face of metatar-

sal II: absent (0) or present (1). The dorsalfaces of metatarsal II of Yungavolucris brev-ipedalis, Lectavis bretincola, Soroavisaurusaustralis, Avisaurus archibaldi, and the TwoMedicine Formation form exhibit a pro-nounced tubercle (figs. 2A, 3A, 6A, 7A, 8A,12A), previously interpreted as the area ofattachment of the M. tibialis cranialis (Brett-Surman and Paul, 1985). Variable positionsof this tubercle occur among the abovemen-tioned forms. The presence or absence ofthistubercle is uncertain in Neuquenornis volans,in which the proximal half of the shaft ofmetatarsal II is missing (Chiappe and Calvo,in press). A tubercle on the dorsal surface ofmetatarsal II is absent in the outgroups.

5. J-shaped metatarsal I: absent (0) or

22 NO. 3083


A

IV II II

B

III IIIIV


present (1). Metatarsal I of Soroavisaurusaustralis and Neuquenornis volans is J-shapedin medial view (fig. 9), and extremely laterallycompressed. Unfortunately, metatarsal I ismissing in Yungavolucris brevipedalis, Lec-tavis bretincola, Avisaurus archibaldi, and theTwo Medicine enantiomithine. This meta-tarsal morphology is also absent in the out-groups, in which metatarsal I is straight orslightly curved as in Patagopteryx deferrari-isi, and Mononykus olecranus.

6. Trochlea ofmetatarsal IV medially con-cave: absent (0) or present (1). In Avisaurusarchibaldi the trochlea for metatarsal IV isremarkably concave medially, being highlyasymmetrical (figs. 1 1, 1 2B). In the TwoMedicine form, the trochlea for metatarsal

IV is also concave medially and asymmet-rical, although to a lesser extent than in A.archibaldi (Varricchio and Chiappe, in press).This trochlear morphology is absent in So-roavisaurus australis (Fig. 1 1), while the con-dition is uncertain in Yungavolucris brevipe-dalis, Lectavis bretincola, and Neuquenornisvolans. This condition is also absent in theoutgroup taxa.

7. Proximal articular surface dorsally in-clined: absent (0) or present (1). The articularsurface of the proximal end of Soroavisaurusaustralis (fig. 7C, D), Avisaurus archibaldi (fig.12A; see also Brett-Surman and Paul, 1985;fig. 3C, E), and the Two Medicine Formationform (Varricchio and Chiappe, in press) slantstoward the dorsal border such that this plane

231993


A

]3Kr AVISAURIDAE

ENANTIORNITHES

B /4

- AVISAURIDAE

.ENANTIORNITHES

ENANTIORNITHES

Fig. 13. Fundamental cladograms (length 12, consistency index 0.83, retention index 0.87) resultingfrom 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 planeof the tarsometatarsus. This character is un-certain in Neuquenornis volans. Otherwise,the articular surface of the proximal end isnot inclined in Yungavolucris brevipedalis (fig.2C, D), Lectavis bretincola (fig. 6C, D), or inany of the outgroup forms.

8. Strong transverse convexity of the dor-sal surface of the mid-shaft of metatarsal III:absent (0) or present (1). The dorsal surfaceof the mid-shaft of metatarsal III is stronglyconvex transversely in Soroavisaurus aus-tralis (figs. 7A, C, 8A), Avisaurus archibaldi(fig. 12A; see also Brett-Surman and Paul,

24 NO. 3083


1985; fig. 3H), Neuquenornis volans (Chiappeand Calvo, in press), and the Two MedicineFormation enantiornithine (Varricchio andChiappe, in press). Conversely, this condi-tion is absent in Yungavolucris brevipedalis(figs. 2A, 3A) and Lectavis bretincola (fig. 6A).This condition is also absent in the outgroupforms in which the dorsal surface ofthe mid-shaft is instead slightly convex or nearly flat.

9. Distal end of metatarsal II stronglycurved medially: absent (0) or present (1). InSoroavisaurus australis (figs. 7A, B, 8), Avi-saurus archibaldi (fig. 12), and the Two Med-icine enantiornithine (Varricchio andChiappe, in press), the distal end of metatar-sal II is strongly twisted medially. This con-dition is absent in Yungavolucris brevipedalis(figs. 2A, B, 3A) and Neuquenornis volans (fig.9). In Lectavis bretincola the condition re-mains unknown. The distal end ofmetatarsalII is not curved medially in Patagopteryx de-ferrariisi and Mononykus olecranus.

10. Trochlea ofmetatarsal II broader thanthe trochleae of metatarsal III and IV: absent(0) or present (1). In Yungavolucris brevipe-dalis (figs. 2A, B, F, 3A, C), Soroavisaurusaustralis (figs. 7A, B, F, 11), Avisaurus ar-chibaldi (figs. 11, 1 2B), Neuquenornis volans(Chiappe and Calvo, in press), and the TwoMedicine Formation form (Varricchio andChiappe, in press) the trochlea of metatarsalII is the broadest trochlea of the tarsometa-tarsus. This condition is uncertain in Lectavisbretincola. Otherwise, the trochlea of meta-tarsal II is not broader than the trochlea ofmetatarsal III in Patagopteryx deferrariisi andMononykus olecranus.

PHYLOGENETIC DISCUSSION

Three most parsimonious cladograms(length 12, consistency index 0.83, retentionindex 0.87) resulted from the cladistic anal-ysis of tarsometatarsal characters (fig. 13).These three fundamental cladograms differin depicting the relationships among Lectavisbretincola, Yungavolucris brevipedalis, and theAvisauridae, a group composed of Neuque-nornis volans, Soroavisaurus australis, Avi-saurus archibaldi, and the Two MedicineFormation form. In two of these cladograms(fig. 1 3A, B), L. bretincola and Y. brevipedalis

are successive sister taxa of Avisauridae, al-though their sequence is reversed. The re-maining tree (fig. 1 3C), identical to the strictconsensus, depicts these three taxa clusteredin an unresolved tricotomy.The Enantiomithes, the clade composed of

all ingroup taxa, is unambiguously diagnosedby a small metatarsal IV (character 1), a tu-bercle in the dorsal face ofmetatarsal II (char-acter 4), and a trochlea ofmetatarsal II broad-er than the trochleae of metatarsals III andIV (character 10). In the fundamental clado-gram of figure 1 3A, the latter character sup-ports ambiguously the sister-group relation-ship between Yungavolucris brevipedalis andAvisauridae. On the other hand, the sister-group relationship of the latter taxon andLectavis bretincola, resulting in the funda-mental cladogram of figure 1 3B, is supportedambiguously by the excavated plantar surfaceof the tarsometatarsus (character 3). The ab-sence of the latter character is optimized asautapomorphic of both Y. brevipedalis andthe Two Medicine Formation form in theremaining trees (fig. 1 3A, C).Neuquenornis volans, soroavisaurus aus-

tralis, Avisaurus archibaldi, and the TwoMedicine enantiomithine are clustered in themonophyletic Avisauridae in all resultantcladograms (fig. 13). The monophyly of theAvisauridae is supported unambiguously bythe presence of a strong plantar projection ofthe medial rim of the trochlea of metatarsalIII (character 2), and the transverse convexityofthe dorsal surface ofthe mid-shaft ofmeta-tarsal III (character 8) in the cladograms offigure 1 3A and C. Character 2 has an ambig-uous optimization in the remaining clado-gram (fig. 13B) in which, pending determi-nation ofthe condition in Lectavis bretincola,it may diagnose a more inclusive clade. Inaddition, the presence of a J-shaped meta-tarsal I (character 5) is an ambiguous syna-pomorphy of Avisauridae in all the resultantcladograms (fig. 13). Pending determinationofthe condition in L. bretincola and Y. brevi-pedalis, the latter character may represent amore inclusive synapomorphy.Within the Avisauridae, Neuquenornis vo-

lans is the sister group ofthe clade composedby Soroavisaurus australis, Avisaurus archi-baldi, and the Two Medicine Formation form(fig. 13). In all resultant cladograms, the latter

251 993


three taxa are united by one unambiguoussynapomorphy: the strong medial curvatureof the distal end of metatarsal II (character9). In addition, this clade is supported am-biguously by another synapomorphy: thepossession ofa dorsally inclined proximal ar-ticular surface (character 7). Pending deter-mination in N. volans, this character may bea synapomorphy ofAvisauridae. S. australis,in turn, is the sister taxon ofthe clade formedby A. archibaldi and the unnamed form fromthe Two Medicine Formation (fig. 13). Thelatter clade is unambiguously diagnosed bythe possession ofa medially concave trochleaof metatarsal IV (character 5).

CONCLUSIONS

Yungavolucris brevipedalis, Lectavis bre-tincola, and Soroavisaurus australis are threediagnosable taxa possessing synapomorphiesthat cluster them in the monophyletic En-antiornithes. The recognition of these threenew enantiornithine species, in addition to aworldwide distribution (see above), empha-sizes the important role played by this groupin the Late Cretaceous avifauna.The proposed phylogenetic hypothesis is

the first attempt to understand the interre-lationships of the Late Cretaceous enantior-nithines. This hypothesis clearly indicates thenonmonophyletic status of the enantiorni-thine assemblage of El Brete, with some el-ements (e.g., Soroavisaurus australis) sharinga most recent common ancestor with NorthAmerican taxa. It also supports the mono-phyly ofthe Avisauridae, establishing the sis-

ter-group relationship ofthe only two knownNorth American avisaurids (Avisaurus ar-chibaldi and the Two Medicine Formationform). It finally corroborates previous claimsabout the avian affinity of Avisaurus archi-baldi (Martin 1991; Chiappe, 1 992a), a taxonthat was originally regarded as a nonaviantheropod (Brett-Surman and Paul, 1985;Norman, 1990).

ACKNOWLEDGMENTSI am especially grateful to Jaime Powell

(Instituto-Fundacion "Miguel Lillo," Uni-versidad Nacional de Tucuman) for lettingme study the specimens here described. I alsothank Jose Bonaparte (Museo Argentino deCiencias Naturales), Jorge Calvo (Museo deCiencias Naturales, Universidad Nacional delComahue), Mark Norell (AMNH), John Os-trom (Yale Peabody Museum), AltangarelPerle (Mongolian Museum of Natural His-tory), Jose Luis Sanz (Universidad Autono-ma de Madrid), and David Varricchio (Mu-seum of the Rockies) for making availablefor study material under their custody. I amgrateful to James Clark (AMNH) and DavidVarricchio who made very useful commentson the manuscript, and to Mark Norell, StorrsOlson (Smithsonian Institution), and LarryWitmer (New York College of OsteopathicMedicine) for their critical reviews. Photo-graphs were skillfully made by Frank Ippolito(AMNH). This research was supported by aFrick Research Fellowship of the AmericanMuseum of Natural History.

REFERENCESAlvarenga, H. M. F., and J. F. Bonaparte

1992. A new flightless land bird from the Cre-taceous ofPatagonia. In K. E. Campbell(ed.), Papers in avian paleontology,honoring Pierce Brodkorb, pp. 51-64.Los Angeles: Natural History Museumof Los Angeles County, Science Series36.

Baumel, J. J, A. S. King, A. M. Lucas, J. E. Brea-zile, and H. E. Evans1979. Nomina Anatomica Avium. London:

Academic Press, 637 pp.Bonaparte, J. F., and J. E. Powell

1980. A continental assemblage of tetrapodsfrom the Upper Cretaceous beds of El

Brete, northwestern Argentina (Saurop-oda-Coelurosauria-Carnosauria-Aves).Mem. Soc. Geol. France (n. ser.) 139:19-28.

Bonaparte, J. F., J. A. Salfitty, G. Bossi, and J. E.Powell1977. Hallazgos de dinosaurios y aves creta-

cicas en la Formaci6n Lecho de El Brete(Salta), pr6ximo al limite con Tucuman.Acta Geol. Lilloana 14: 5-17.

Brett-Surman, M. K., and G. Paul1985. A new family of bird-like dinosaurs

linking Laurasia and Gondwanaland. J.Vertebr. Paleontol. 5(2): 133-138.

Carpenter, J. M.

26 NO. 3083


1988. Choosing among multiple equally par-simonious cladograms. Cladistics 4:291-296.

Chiappe, L. M.1991a. Cretaceous avian remains from Pata-

gonia shed new light on the early radi-ation ofbirds. Alcheringa 15(3-4): 333-338.

1991 b. Cretaceous birds of Latin-America.Cretaceous Res. 12: 55-63.

1992a. Enantiornithine tarsometatarsi and theavian affinity of the Late CretaceousAvisauridae. J. Vertebr. Paleontol. 12(3):344-350.

1992b. Osteologia y sistematica de Patagopte-ryx deferrariisi Alvarenga y Bonaparte(Aves) del Cretacico de Patagonia. Filo-genia e historia biogeogrifica de las avesCretacicas de America del Sur. Ph.D.diss., Universidad de Buenos Aires,Buenos Aires, 429 pp.

In press. Phylogenetic position of the Creta-ceous birds of Argentina: Enantior-nithes and Patagopteryx deferrariisi. InD. S. Peters (ed.), Proc. 3rd Symp. Soc.Avian Paleontology and Evolution.Frankfurt.

Chiappe, L. M., and J. 0. CalvoIn press. Neuquenornis volans, a new Upper

Cretaceous bird (Enantiornithes: Avi-sauridae) from Patagonia, Argentina. J.Vertebr. Paleontol.

De Queiroz, K., and J. Gauthier1990. Phylogeny as a central principle in tax-

onomy: phylogenetic definitions of tax-on names. Syst. Zool. 39: 307-322.

1992. Phylogenetic taxonomy. Annu. Rev.Ecol. Syst. 23: 449-480.

Farris, J.1988. Hennig86 references. Documentation

for version 1.5. Privately published.Gauthier, J. A.

1986. Saurischian monophyly and the originof birds. In K. Padian (ed.), The originofbirds and the evolution offlight. Mem.Calif. Acad. Sci. 8: 1-55.

Gomez Omil, R. J., A. Boll, and R. M. Hernandez1989. Cuenca Cretacico-Terciaria del No-

roeste Argentino (Grupo Salta). In G.A. Chebli and L. A. Spalletti (eds.),Cuencas sedimentarias argentinas. Ser.Correl. Geol. 6: 43-64.

Martin, L. D.1983. The origin and early radiation of birds.

In A. H. Bush and G. A. Clark, Jr. (eds.),Perspectives in ornithology, pp. 291-338. Cambridge: Cambridge UniversityPress.

1991. Mesozoic birds and the origin of birds.In H-P. Schultze and L. Trueb (eds.),Origins of the higher groups of tetra-pods, pp. 485-540. Ithaca: Comstock.

Molnar, R. E.1986. An enantiornithine bird from the Lower

Cretaceous of Queensland, Australia.Nature 322: 736-738.

Nessov, L. A.1984. Pterosaurs and birds of the Late Cre-

taceous of Central Asia. Paleontol. Zh.1: 47-57 [In Russian].

Nessov, L. A., and A. A. Jarkov1989. New Cretaceous-Paleogene birds of the

USSR and some remarks on the originand evolution of the Class Aves. Proc.Zool. Inst., Leningrad, 197: 78-97 [InRussian].

Norman, D. B.1990. Problematic Theropoda: "Coeluro-

saurs." In D. B. Weishampel, P. Dod-son, and H. Osm6lska (eds.), The Di-nosauria, pp. 280-305. Berkeley: Univ.California Press.

Perle, A., M. A. Norell, L. M. Chiappe, and J. M.Clark1993a. Flightless bird from the Cretaceous of

Mongolia. Nature 362:623-626.1993b. Correction. Flightless bird from the

Cretaceous of Mongolia. Nature 363:188.

Sanz, J. L., and A. Buscalioni1992. A new bird from the Early Cretaceous

of Las Hoyas, Spain, and the Early ra-diation of birds. Paleontology 35(4):829-845.

Sanz, J. L., J. F. Bonaparte, L. M. Chiappe, andA. BuscalioniIn press. La posici6n filogenetica de Concornis

lacustris del Cretacico inferior de LasHoyas, Espania. X Jorn. Arg. Paleontol.Vert., Ameghiniana.

Varricchio, D. J., and L. M. ChiappeIn press. A new bird from the Cretaceous Two

Medicine Formation of Montana. J.Vertebr. Paleontol.

Walker, C. A.1981. New subclass of birds from the Creta-

ceous of South America. Nature 292:51-53.

Wiley, E. O., D. Siegel-Causey, D. R. Brooks, andV. A. Funk1991. The compleat cladist. Univ. Kansas

Mus. Nat. Hist., Spec. Publ. 19: 1-158.Zhou, Z., F. Jin, and J. Zhang

1992. Preliminary report on a Mesozoic birdfrom Liaoning, China. Chinese Sci. Bull.37(16): 1365-1368.

1993 27

Recent issues of the Novitates may be purchased from the Museum. Lists of back issues of theNovitates, Bulletin, and Anthropological Papers published during the last five years are availablefree 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).

americanmuseum noqvitates - american museum of natural history

Documents