Introduction

The stork family (Ciconiidae) is a well-defined group of waterbirds, traditionally divided into three tribes: the Mycteriini, the Ciconiini, and the Leptoptilini (Kahl 1971, 1972, 1979). They were already differentiated by the early Tertiary with the first record occuring in the early Oligocene of the Fayum series in Egypt (Olson 1985; Feduccia 1996). However, most fossil storks are known from the Tertiary of Europe, Asia, the Americas, and Africa, and are based on isolated and fragmentary remains that preclude our understanding on the phylogeny and relationships within the Ciconiidae (Feduccia 1996).

The oldest South American fossil stork, Ciconiopsis antarctica Ameghino, 1899, was described from the early Oligocene of Santa Cruz (Argentina), but its ordinal assignment has been questioned (Olson 1985; Agnolin 2004). More recently, isolated fragments of tarsometatarsi undoubtly referable to the Mycteriini were reported from the late Miocene in Entre Ríos Province, Argentina (Noriega 1995; Noriega and Agnolin 2006). Recently, Louchait et al. (2005) described new Tertiary fossil Leptoptilini from Africa and revised previous records of other extinct Old World storks.

This contribution describes a new species of a large stork referable to the genus Leptoptilos, recovered from the late Miocene of Argentina. Fossils assignable to Leptoptilos have hitherto been unknown from Tertiary deposits of South America. Therefore, this finding constitutes the first record for both the genus and the tribe on this continent.

Methods

Comparisons were made with the extant species from the MHNT, LAC-MNHN, and CICYTTP collections: Mycteria americana (tribe Mycteriini), Ciconia maguari (tribe Ciconiini), and five species of the tribe Leptoptilini including Ephippiorhynchus senegalensis, Jabiru mycteria, Leptoptilos crumeniferus, L. dubius, and L. javanicus.

Fig. 1.

A. Location of the Peninsula Valdés area with the fossiliferous site at Punta Buenos Aires, Chubut Province, Patagonia; Argentina. B. Stratigraphic section of the Puerto Madryn Formation. The bone symbol indicates the fossiliferous level.

Measurements were taken following Becker (1986, 1987), Ono (1980), Noriega (2002), and Louchart et al. (2005). Anatomical nomenclature follows Baumel and Witmer (1993).

Geologic setting and paleoenvironment

Study of the Neogene marine ingression in Patagonia, informally known as “Patagoniense-Entrerriense”, began with Darwin's travel in 1846, and was followed by Ameghino (1889, 1898, and 1906). These invertebrate-rich sediments of marine origin were later studied by von Ihering (1907), Rovereto (1913, 1921), and Windhausen (1931). Frenguelli (1926) and Feruglio (1949) determined these beds to be Miocene-Pliocene in age. Haller (1978, 1981) extended the analysis to northern Patagonia, proposing two stratigraphic units —the Gaiman and Puerto Madryn formations—to include the deposits laid down by the transgression of the “Patagoniense-Entrerriense” or “Patagoniano” sea. The same author assigned a late Miocene age to these levels, and interpreted them as belonging to a temperate and shallow seawater environment. Radiometric dating of these levels at the locality of Punta Craker yielded an age of 9.41 Ma (Zinsmeister et al. 1981). Studies on the distribution of facies and lateral correlations were made by Scasso and Del Río (1987), who proposed a unique sedimentary and regressive cycle for the “Patagoniense—Entrerriense” units. The sedimentation of the Puerto Madryn Formation, the fossil bearing unit, was characterized by a transition from off-shore to literal environment (Scasso and Del Río 1987).

The fossil specimen described herein was recovered at the locality of Punta Buenos Aires at Península Valdés, Chubut Province, Argentina (Fig. 1A). The complete sequence of the Puerto Madryn Formation reaches a thickness of up to 50 m; however, the base of the formation does not emerge at the fossil locality (Fig. 1B). The base of the profile has green and gray mudstone and limestone units, corresponding to Facies 4a of Scasso and Del Río' s profile (1987). The fossil stork comes from this mudstone deposit.

Systematic paleontology

Fig. 2.

The marabou stork Leptoptilos patagonicus sp. nov., specimen MEF-1363 (holotype) from the Puerto Madryn Formation (late Miocene), Punta Buenos Aires, Peninsula Valdés, Chubut Province, Patagonia, Argentina. A. Right humerus in anconal (A₁), palmar (A₂), and distal (A₃) views. B. Sternum in lateral aspect. C, D. Right (C) and left (D) ulnae in palmar and anconal views. E, F. Left (E) and right (F) radii, in palmar views. G. Left carpometacarpus in internal (G₁) and external (G₂) views.

Description

Humerus.—The humerus is larger than in Ephippiorhynchus senegalensis and Jabiru mycteria, but shorter than in L. dubius and L. crumeniferus (Table 1; Fig. 2A). The width of distal end of shaft at the level of the tuberculum supracondylare dorsale, just proximal to the epiphysis, is greater than in most of the storks compared (Table 1). However, the distal end width is proportionally smaller in comparison to other measurements of the bone. The distal epiphysis is moderately flexed anteriad as in Jabiru. The tuberculum supracondylare dorsale is well developed proximally, angling moderately relative to shaft as in Ciconia, and not projected laterally as observed in the remaining genera compared. The epicondylus ventralis is less protrudent medio-distally than in Jabiru, Ephippiorhynchus, Leptoptilos, and Mycteria; similar to those of Ciconia. The scars for M. pronator profundus and M. flexor carpi ulnaris are larger and deeper than those of the species compared, giving a pronounced excavation to the area of the ventral side distal to the epicondylus ventralis and undercutting the latter more markedly. The processus flexorius is more strongly projected distally than in all genera compared. The tuberculum supracondylar ventrale is similar to that of Ephippiorhynchus and Leptoptilos, i.e., elongated vertically and flattened, whereas it is more transverse and rounded in the remaining species compared. The impression of M. brachialis is large and deep, being well excavated medially and shallower laterally, with its proximal end forming a marked ridge along the lateral edge of the shaft as in Jabiru. The fossa olecrani is broad and deeper than in the comparative species.

Table 1.

Measurements (in mm) of the humerus and the ulna of Leptoptilini. Humerus: greatest preserved length measured from the proximal margin of insertion of M. latissimus dorsi caudalis through the midpoint of the condylus medialis (L.h); transverse width at midshaft (W-s.h); depth at midshaft (D-s.h); transverse width of distal end from the tuberculum supracondylare dorsale to the epicondylus ventralis (W-d.h). Ulna: greatest length measured from the olecranon through the condylus ventralis ulnaris (L.u); width of proximal end through cotyles (W-p.u); width of midshaft (W-s.u); depth of midshaft (D-s.u); width of distal end through condyles (W-d.u).

Table 2.

Measurements (in mm) of the radius and the carpometacarpus of Leptoptilini. Radius: greatest length from cotyla humeralis to facies articularis radiocarpalis (L.r; estimated by summing both right and left partial radii of MEF-1363); width of proximal end from facies articularis ulnaris to head (W-p.r); width of distal end (W-d.r). Carpometacarpus: greatest length from the most proximal portion of the trochlea carpalis through facies articularis digitalis minor (L.c); transverse width of proximal end from the ventral rim of the trochlea carpalis through processus extensorius (Cc); depth of proximal end from ventral to dorsal rims of trochlea carpalis (D-p.c); length of the os metacarpale alulare I from processus extensorius to processus alularis (L-McI); depth at midshaft of os metacarpale majus (D-s.c); transverse width of midshaft of os metacarpale majus (W-s.c); greatest depth of distal end, measured across the dorsal edge of facies articularis digitalis major (D-d.c); transverse width of distal end from edge of facies articularis digitalis major through facies articularis digitalis minor (W-d.c).

Ulna.—The ulna is slightly more robust and longer than in the species used in comparison with the exception of L. dubius and L. crumeniferus (Table 1; Fig. 2C, D). The cotyla ventralis is shallow and subelliptical as in Jabiru, Ephippiorhynchus, and Leptoptilos; whereas it is deeper and more rounded in Ciconia and Mycteria. The impressio brachialis is less deep than in Jabiru, with its bordering bony ridges less marked, as in Leptoptilos, Ephippiorhynchus, Ciconia, and Mycteria. The impression of M. scapulotriceps is more distal and elongated. The condylus dorsalis is proportionally greater, more protruding medially than in the living species compared.

Radius.—The cotyla humeralis is quadrangular as in Jabiru, Leptoptilos, and Ephippiorhynchus, whereas it is more subelliptical in Ciconia. The facies articularis ulnaris is prominent. The tuberculum bicipitalis radialis is more distally extended than in Jabiru, similar to those of Leptoptilos, Ephippiorhynchus, and Ciconia. The pneumatic foramina at proximal and distal ends are absent, as in Ephippiorhynchus and Ciconia, unlike the conspicuous pneumatization shown by Jabiru and Leptoptilos. The facies articularis radiocarpalis is more elongated transversally and the sulcus tendinosus is larger than in the species compared, making distal end wider. The tuberculum aponeurosis ventralis is less bulbous than in Jabiru and Leptoptilos, as in Ephippiorhynchus and Ciconia (Fig. 2E, F).

Carpometacarpus.—The internal rim of the trochlea carpalis is less rounded and less protruding posteriorly, merging with the shaft more distally than in all the species compared. The proximal and distal metacarpal symphyses are similar to those of Leptoptilos, proportionally larger than in the remaining species compared (Fig. 2G; Table 2).

Sternum.—The size and robustness of the sternum is slightly greater than those of Jabiru mycteria. The sulcus articularis coracoideus is wider than those of Ephippiorhynchus and Ciconia, similar in width to that of Jabiru and Leptoptilos. The ventral and dorsal lips of the sulcus articularis coracoideus are wider than in all the species compared, more similar to those of Jabiru and Leptoptilos. The pila coracoidea is wider than in Ephippiorhynchus and Ciconia, with its medial edge more rounded as in Jabiru and not clear-cut as in the former (Fig. 2B).

Pelvis.—The foramen ilioischiadicum is bigger and less elliptical than in Ephippiorhynchus and Jabiru, its shape being considerably more similar to those of Leptoptilos and Ciconia. The ala ischii is more expanded ventrally with its lower border curved, not straight as in the species compared (Fig. 3C).

Tibiotarsus.—The shape of the facies articularis medialis is quadrangular, similar to that of Ephippiorhynchus and Leptoptilos, with its medial and posterior borders forming a close to 90° angle; the junction of these borders is rounded in Jabiru, Mycteria, and Ciconia. The crista fibularis is proportionately shorter and more spread outwards than in Ephippiorhynchus and Jabiru. The condylus medialis and the condylus lateralis are less projected anteriorly, but their cranial surfaces are more developed proximally than in all the species compared. The distal width through condyles is similar to that measured in L. dubius, larger than those of Ephippiorhynchus, Jabiru, and Mycteria (Tables 3, 4), with the condylus medialis and the condylus lateralis more or less parallel and aligned to the respective borders of the shaft; the width of the distal end of tibiotarsus is proportionately broader in Ciconia than in L. patagonicus due to its more pronounced mediolateral expansion. The pons supratendineus is covered by a matrix which hides its morphology (Fig. 3A).

Fig. 3.

The marabou stork Leptoptilos patagonicus sp. nov., specimen MEF-1363 (holotype) from the Puerto Madryn Formation (late Miocene), Punta Buenos Aires, Peninsula Valdés, Chubut Province, Patagonia, Argentina. A. Right tibiotarsus in posterior (A₁) and anterior (A₂) views. Details of the distal end of right tibiotarsus in medial (A₃), anterior (A₄), and lateral (A₅) views. B. Tip of mandible in lateral (B₁) and dorsal (B₂) views. C. Pelvis in lateral view.

Skull fragments.—The tip of the mandibular symphysis is robust, with the edges of the tomial shelf sharp and high. The fragmentary state of the ramus mandibulae and the os articulare make it difficult to recognize morphological features on both of them (Fig. 3B).

Discussion and conclusions

The traditional systematic arrangement divides the Ciconiidae in the tribes Mycteriini, Ciconiini, and Leptoptilini on the basis of external morphology (Kahl 1972, 1979). The tribe Leptoptilini comprises three genera (Ephippiorhynchus, Jabiru, and Leptoptilos), with six living species distributed in the Neotropics, Africa, the Oriental Region, SE of Asia, New Guinea, and Australia. These storks are distinctive by their large size and massive bills. The genus Jabiru is sometimes included in Ephippiorhynchus due to its resemblance in feeding and display behaviour (Wood 1984), but seems to be morphologically intermediate between this genus and Leptoptilos (Elliott 1992).

Fig. 4.

Tibiotarsus ratios in the living Leptoptilini, fossil Leptoptilos, and L. patagonicus sp. nov., specimen MEF-1363 (holotype) from the Puerto Madryn Formation (late Miocene), Punta Buenos Aires, Peninsula Valdés, Chubut Province, Patagonia, Argentina. A. Minimum shaft width to total length ratio. B. Distal depth to distal width ratio. Modified from Louchart et al. (2005).

The classification derived from DNA hybridization studies recognizes only the Ciconiinae under the family level (Sibley and Monroe 1990). A more recent cladistic approach, based on molecular evidence, suggests that the Leptoptilini is probably a paraphyletic group which comprises basal species of storks (Slikas 1997). Osteological characters are not commonly used in ornithology to discriminate groups within the family level and, indeed, there are only a few discrete characters that are useful in separating the tribes of Ciconiidae (e.g., Cheneval 1984; Haarhoff 1988; Olson 1991; Louchart et al. 2005). Thus, diagnoses combining external and molecular characters with those of the skeleton will be necessary in the future for sound systematic revision of the subordinated natural groups of storks. Because this phylogenetic task largely exceeds the goal of our contribution, we adopt provisionally the traditional classification scheme, referring the specimen to the Leptoptilini due to its phenetic similarities with the genera compared in this tribe in overall morphology, robustness, size, and inter-segment proportions of limb bones (Louchart et al. 2005). The similarities of the specimen MEF-1363 with Leptoptilini are clearly observed in the sternum, humerus, ulna, radius, tibiotarsus, and by having the limb bones larger and more robust than in Mycteriini and Ciconiini (Louchart et al. 2005).

Table 3.

Measurements (in mm) of the pelvis and the tibiotarsus of Leptoptilini. Pelvis: length from the anterior end of ala preacetabularis ilii to the posterior end of ala postacetabularis ilii (Tl); least width of ala preacetabularis ilii measured from the constriction of its lateral free edge to crista spinosa synsacri (Pf); width measured from the lateral edge of antitrochanter to crista spinosa synsacri (Ph); length from the anterior end of ala preacetabularis ilii to the anterior end of pubis, just ventral to foramen acetabuli (Pt); length of foramen ilioischiadicum at its long axis (Lif); tibiotarsus: total length from facies articularis at proximal end to distal portion of condyles (L); depth at midshaft (D-s.t); minimal width of shaft medio-laterally (Mw); width of distal articular end medio-laterally (Dw); depth of distal articular end cranio-caudally (Dd); greatest depth of condylus lateralis (D-l.con).

We adopt the criteria of Louchart et al. (2005) for generic identification of the specimen MEF-1363 within the Leptoptilini. The proportions of the tibiotarsus in Leptoptilos are similar to those in Jabiru, being the tibiotarsus more robust in the former than in Ephippiorhynchus (Table 4; Fig. 4A). The carpometacarpus in Leptoptilos and Ephippiorhynchus is more slender than in Jabiru. The same conditions of the tibiotarsus and carpometacarpus, similar to Leptoptilos, are also observed on the specimen MEF-1363. Moreover, the proportions of the distal end of the tibiotarsus are typical in Leptoptilos in comparison to Ephippiorhynchus and Jabiru, the ratio of its depth to its width being less in the former than in the later two (Miller et al. 1997; Louchart et al. 2005). The distal end of the tibiotarsus MEF-1363 exhibits a ratio similar to that of Leptoptilos (Table 4; Fig. 4B). Consequently, we establish that the fossil specimen herein described belongs to Leptoptilos.

Table 4.

Measurements (in mm) of the tibiotarsus of extinct and living Leptoptilini, modified from Louchart et al. (2005: 555; Table 2). Abbreviations: L, total length; dw, width of distal articular end medio-laterally; dd, depth of distal articular end cranio-caudally; mw, minimal width of shaft medio-laterally. Most values are single measurements; means are in bold, with the sample sizes (n) at right.

A comparison with the truly giant fossil L. falconeri and the largest individuals of extant species of Leptoptilos shows that L. patagonicus was a large stork. The tibiotarsus of L. patagonicus is shorter in length than those referred to L. falconeri (Table 4). The largest living L. crumeniferus and L. dubius have tibiotarsi of similar length or slightly shorter than that of L. patagonicus (Table 4), but the inter-segment proportions between wing and leg bones are quite different among them; wing elements of L. patagonicus are considerably smaller than those of the former.

Finally, as noted in the diagnosis and description, MEF-1363 exhibits a combination of morphological characters which, together with the presence of short wings relative to the legs, merit the recognition of a new species.

Leptoptilini have a rich temporal and spatial distribution in the Paleartic, Oriental, Ethiopian, and Australasian regions (Louchart et al. 2005: 561; Table 5), with Jabiru mycteria as the only Neotropical species with the fossil record restricted to the Late Pleistocene of Peru (Campbell 1979). Fossils assignable to the Leptoptilini have previously been unknown from the Tertiary of South America. Therefore, Leptoptilos patagonicus constitutes not only the first record for the genus in the Neogene of South America, but it also reveals the presence of an old lineage of the tribe in this continent. The paleobiogeographic significance of this find cannot be fully determined until the phylogenetic relationships of L. patagonicus to the remaining extinct and living storks are elucidated. This will also provide tools to link the natural history of this Patagonian marabou stork with those of Eurasia, Africa or Australasia.