The oldest fossil record of the extant penguin genus Spheniscus—a new species from the Miocene of Peru


The oldest fossil record of the extant penguin genus Spheniscus—a new species from the Miocene of Peru

Ursula B. Göhlich [ursula.goehlich@nhm−], Université Claude Bernard – Lyon 1, UMR P.E.P.S., Bâtiment Géode, 2 rue Dubois, F−69622 Villeurbanne Cedex, France; present address: Naturhistorisches Museum Wien, Geolo− gisch−Paläontologische Abteilung, Burgring 7, A−1010 Wien, Austria.

Described here is a partial postcranial skeleton and additional disarticulated but associated bones of the new fossil penguin Spheniscus muizoni sp. nov. from the latest middle/earliest late Miocene (11–13 Ma) locality of Cerro la Bruja in the Pisco Formation, Peru. This fossil species can be attributed to the extant genus Spheniscus by postcranial morphology and is the oldest known record of this genus. Spheniscus muizoni sp. nov. is about the size of the extant Jackass and Magellanic penguins (Spheniscus demersus and Spheniscus magellanicus). Beside Spheniscus urbinai and Spheniscus megaramphus it is the third species of Spheniscus represented in the Pisco Formation. This study contains morphological comparisons with Tertiary penguins of South America and with most of the extant penguin species.

Key words : Spheniscidae, Pisco Formation, Miocene, Peru.

Acta Palaeontologica Polonica 52 (2): 285–298.


The fossil record of penguins in South America comes from both Atlantic and Pacific coasts and is restricted to findings in Argentina, Chile, and Peru; their stratigraphic distribution ranges from the late Eocene?–early Oligocene (Acosta Hospi− taleche 2005) to the late Pliocene (Emslie and Correa 2003). Most of the described fossil species come from the early Mio− cene of Patagonia, Argentina (see Ameghino 1891, 1895, 1899, 1901, 1905, 1920; Moreno andMercerat 1891; Simpson 1972, 1975a). Recently, the South American fossil penguins have been taxonomically revised (e.g., Acosta Hospitaleche 2003, 2005; Acosta Hospitaleche and Canto 2005; Acosta Hospitaleche and Tambussi 2004a, b; Acosta Hospitaleche et al. 2005) and concluded in a reduction of 35 formerly named penguin species to about 14 different taxa,which are exclusive to South America (Acosta Hospitaleche and Tambussi 2004a, b). The following taxa are known fromArgentina: Arthrodytes andrewsi (Ameghino, 1901) and Paraptenodytes robustus (Ameghino, 1895) (both from the upper Eocene–lower Oligo− cene San Julian Formation, Patagonia), Eretiscus tonnii (Sim− pson, 1981), Palaeospheniscus bergi Moreno and Mercerat, 1891, Ps. patagonicus Moreno and Mercerat, 1891, Ps. bilo− culata Simpson, 1970 (all from the lower Miocene Gaiman Formation, Patagonia), Paraptenodytes antarcticus (Moreno and Mercerat, 1891) (from the lower Miocene Monte Léon Formation and the lower upper Miocene Puerto Madryn For− mation, both Patagonia) and Madrynornis mirandus Acosta Hospitaleche, Tambussi,Donato, and Cozzuol, 2007 (this vol− ume; from the lower upper Miocene Puerto Madryn Forma− tion, Patagonia). The Chilean fossil penguin taxa include: Paraptenodytes robustus, Pa. antarcticus, Palaeospheniscus sp., Pygoscelis grandis Walsh and Suárez, 2006, Py. cal− dernensis Acosta Hospitaleche, Chávez, and Fritis, 2006, Spheniscus sp. (all from the upper Miocene–lower Pliocene, Bahia Inglesa Formation), and Spheniscus chilensis Emslie and Correa, 2003 (late Pliocene, Mejillones Formation). The following taxa are known from Peru: a Spheniscidae indet. (late Eocene–early Oligocene, Otuma Formation), Palaeo− spheniscus sp. (lower middleMiocene, Chilcatay Formation), Spheniscus urbinai Stucchi, 2002 and S. megaramphus Stuc− chi, Urbina, and Giraldo, 2003 (both from the upper Mio− cene–lower Pliocene of the Pisco Formation).

Thus, findings of fossil representatives of the extant pen− guin genus Spheniscus are restricted to Peru and Chile and have been known from the late Miocene to late Pliocene hith− erto. For the Pisco Formation of Peru, Noriega and Tambussi (1989) and Cheneval (1993) already mentioned probable new taxa of Spheniscidae, but did not name them; recently they have been described by Stucchi (2002) and Stucchi et al. (2003) as Spheniscus urbinai from different sites of late Mio− cene to early Pliocene age and as S. megaramphus from the upper Miocene of the Montemar locality. In Chile, fossils of Spheniscus are described as cf. Spheniscus by Walsh and Hume (2001) from the middle Miocene–lower Pliocene Bahía Inglesa Formation and as S. chilensis Emslie and Cor− rea, 2003 from the late Pliocene (Mejillones Formation) of Cuenca del Tiburón, Península de Mejillones.

Herein a new fossil species of Spheniscus is described again from the Pisco Formation in Peru, but coming from older deposits at the Cerro la Bruja locality of latest mid− dle/earliest late Miocene (11–13 Ma) age.

Geological setting

The locality of Cerro la Bruja, Department of Ica, is situated in the marine Pisco Formation (Fig. 1). The age of the Pisco Formation ranges from the middle Miocene to early Pliocene (Muizon and DeVries 1985). The deposit of Cerro la Bruja is dated on biostratigraphic data to be of latest middle Miocene to earliest late Miocene age (13–11 Ma; Muizon 1988), and therefore represents the oldest deposit within the Pisco For− mation.

The Pisco Formation corresponds to a marine transgres− sion along the southern coast of Peru during the Neogene and is known for its abundant marine vertebrate fauna. Its geol− ogy and palaeoecology in the Sacaco area were studied by Muizon and DeVries (1985). The marine bird fauna is one of the richest known from the Neogene of South America. Be− side Cerro la Bruja other fossil penguin bearing localities in the Pisco Formation are: El Jahuay (ELJ), ca. 9 Ma, late Mio− cene; Aguada de Lomas (AGL), ca. 7–8 Ma (Muizon et al. 2003), late Miocene; Montemar (MTM), ca. 6 Ma, late Mio− cene; Sacaco Sud (SAS), ca. 5 Ma, early Pliocene; Sacaco (SAO), ca. 3.5 Ma, early Pliocene (datings from Muizon 1984, 1988; Muizon and DeVries 1985). For additional ver− tebrate faunas see Conclusions.

Material and methods

The osteological terminology used here follows Baumel et al. (1993) but terms are translated into English. Measure− ments were taken after von den Driesch (1976) and Stephan (1979). For metrical comparisons, using published data, it has to be considered that different authors (e.g., Simpson 1946, 1975b, 1979a, b), had slightly deviating ways of mea− suring the greatest lengths of some limb bones; often they measured the functional lengths, taking recessed points or sulci between condyles (e.g., for femur, distal tibiotarsus, distal tarsometatarsus) as landmarks. However, the differ− ences in the values are never more than a few mm.

The penguin material of Cerro la Bruja described here co− mes from excavations of Christian de Muizon and belongs to MNHN, but is temporarily housed in the collections of UCBL. For morphological comparisons with the fossil Chil− ean species, Spheniscus chilensis, parts of the paratype mate− rial (UF 143295, 143299, 144112, 144134, 144148, 144149, 144154, 144168, 144169) were at disposal by loan. Morpho− logical comparisons with extant penguins are based on fol− lowing specimens provided by the collections of SAPM and UCBL: Spheniscus humboldti (SAPM 5 male, 12 female; UCBL 25588), S. demersus (SAPM 12 male, 13 female), S. magellanicus (SAPM 8 male, 9 female), Eudyptes chrysochome (junior synonym E. crestatus), (SAPM 6 male, 5 female), E. chrysolophus (SAPM 1 ?), Aptenodytes patagonicus (SAPM 6 male, 8 female), Pygoscelis papua (SAPM 12 male, 3 female), and Py. adeliae (UCBL XII 1978). Metric data of extant penguin species were acquired on the aforementioned specimens as well as additional material from the SAPM. Measurements of S. mendiculus are based on specimens at the LACM and were provided by Kenneth Campbell.

Systematic palaeontology

Sphenisciformes Sharpe, 1891

Spheniscidae Bonaparte, 1831

Genus Spheniscus Moehring, 1758

Type species: Spheniscus demersus (Linnaeus, 1758).

Remarks.—Based on the following osteological features the postcranial penguin material of Cerro la Bruja, presented here, can be affiliated to Spheniscus on generic level. Humerus with humeral head barely swollen proximally (more swollen in Aptenodytes, Pygoscelis, and Eudyptes), proximal outline of humerus without proximal notch between dorsal tubercle and the head (notched in Aptenodytes, Pygo− scelis, and Eudyptes, slightly notched in Palaeospheniscus); bipartite pneumotricipital fossa with a deep cranial fossa (non−bipartite in Paraptenodytes and Arthrodytes); proximal border of pneumotricipital fossa in ventral view straight and almost horizontally (proximally concave in Aptenodytes and Eudyptes); caudal−most process of the ventral epicondyle hardly surpassing ventral margin of distal humerus shaft (in cranial/caudal view; clearly surpassing in Aptenodytes and Pygoscelis). Coracoid with probably closed supracoracoid fo ramen (open in Pygoscelis, Eudyptula, and Aptenodytes and probably in Palaeospheniscus). Femoral head and trochanter proximally at same level (trochanter higher in Eudyptes, Pygoscelis, and Aptenodytes). Tarsometatarsus quite elongate, with an elongation index (maximal length/proximal width) of 2.07, thus smaller than in Palaeospheniscus (elongation in dex: 2.2–2.4); extensor sulci long and deep (shallower in Eudyptes, Pygoscelis, and Aptenodytes); trochlea II and tro chlea IV of about same length distally (unlike in Palaeo spheniscus, Eudyptes, Pygoscelis, and Aptenodytes, where trochlea IV is somewhat shorter than trochlea II); the position and arrangement of the two proximal vascular foramina and of the very pronounced impressions of the extensor retinaculum corresponds best with Spheniscus; lateral intertrochlear inci sion deeper incised proximally than medial one.

Spheniscus muizoni sp. nov.

Figs. 2, 3, 4A, C, E, G, I, 5A.

Holotype: Partial postcranial skeleton MNHN PPI 147: subcomplete left and right coracoid (147a); cranial end of left and subcomplete right scapula (147b); subcomplete left and right humerus (147c); left com plete ulna (147d); proximal and distal end of right femur (147e); com plete right and proximal end of left tibiotarsus (147f); proximal end of left fibula (147g); right complete tarsometatarsus (147h), cranial por tion of sternum with articular sulcus for coracoid and fragment of the craniolateral process (147i); two fragmentary thoracic vertebrae (147j) from the caudal region; seven caudal vertebrae (147k); fragmentary synsacrum (147l);

Paratypes: Additional isolated bones from the type locality (MNHN PPI 148–153): distal fragmentary half of left coracoid, worn (148); right subcomplete coracoid (155); left complete ulna (149); left complete ra dius (150); right complete carpometacarpus (154); distal end of right fe mur (151); cranial end of pygostyle (152); rib fragment without ends (153).

Derivation of the name: Named after Christian de Muizon (MNHN) who collected the studied material and graciously placed it at my dis posal; and in recognition of his efforts and paleontological investiga tions on the vertebrate fauna of the Pisco Formation in Peru.

Type locality and horizon: Cerro la Bruja, Department of Ica, Peru; Pisco Formation, latest middle/earliest late Miocene, ca. 13–11 Ma (Muizon 1988).

Diagnosis.—Small−sized fossil Spheniscidae; similar in size to extant Spheniscus magellanicus (Forster, 1781) and S. demersus (Linnaeus, 1758), slightly smaller than S. hum boldti Meyen, 1834, but lager than S. mendiculus Sundevall, 1871. Distinctly smaller than the fossil species S. urbinai and S. megaramphus, but of similar size to S. chilensis (Tables 1, 2 and Fig. 6).

Carpometacarpus with distinct step between the proximal carpal trochlea and the extensor process of the alular metacar pal (step weaker or absent in extant species of Spheniscus), an almost complete fusion of proximal alular digit with the major metacarpal (fusion less advanced in S. chilensis), and an open spatium without ossified synchondroses (unlike S. chilensis with synchondroses).

Humerus with humeral head barely swollen proximally, proximal outline without proximal notch between dorsal tuberculum and the head, but proximal outline less steep (in caudal view) than in all compared extant species of Sphe niscus; proximal border of pneumotricipital fossa forms no ventrocaudally projecting lip, a feature shared with S. urbinai and S. chilensis, but unlike in extant species of Spheniscus and Pygoscelis; proximal border of pneumotricipital fossa in ven tral view straight and almost horizontally like in extant Sphe niscus species (unlike in S. chilensis, S. urbinai, Dege hendeyi, and extant species of Aptenodytes and Eudyptes where it is proximally concave); distinct concave indentation of bicipital crest between ventral tuberculum and shaft, but which is ab sent in S. chilensis; preaxial angle situated more proximally than in other penguin taxa, at about mid of shaft length; cau dal−most tip at the ventral epicondylus (process−like crest cau dally bordering the humerotricipital sulcus) slightly oriented upwards (ventroproximally); cranial−most tip at ventral epi condylus (process−like crest caudally bordering the scapulotri cipital sulcus) projecting ventrodistally somewhat longer and oriented slightly more distally than in S. chilensis.

Femoral head and trochanter at about the same level prox imally; medial condyle cranially very low and hardly pro jecting (unlike in all other compared extant and fossil pen guins), lateral condyle prominent and reaching further proxi mally than medial one.

Tibiotarsus characterized by its straight (not medially de flected) and relatively narrow (in cranial/caudal view) distal end; distinct and deep medial impression of the medial col lateral ligament situated mediocaudally on proximal end. Tarsometatarsus quite elongate and slender; proximal end relatively narrow; elongation index (2.07) higher than in S. urbinai, but smaller than in Palaeospheniscus; extensor sulci long and deep; extremely deep dorsal infracotylar fossa; proximal vascular foramina on different level (same level in Palaeospheniscus), medial proximal vascular fora men small and situated very proximomedially; hypotarsus with medial hypotarsal crest oriented medioplantarly and bent medially; medial hypotarsal crest divided in two crests, with lateral one being very weak, thin, and shorter; lateral hypotarsal crest blunt and low; lateral intertrochlear incision longer proximally than medial one; trochlea II and IV of about same length distally (unlike in Palaeospheniscus); lat eral condyle of trochlea III (in distal view) at same level dor sally as trochlea IV (plantarly recessed in extant Spheniscus and S. chilensis); lateral and medial condyle of trochlea III (in distal view) dorsally at about same (medial one dorsally more swollen in extant Spheniscus and S. urbinai); trochlea II oriented somewhat obliquely; lateral condyle of trochlea IV (in distal view) plantarly tapering.

Description and comparisons

Remarks.—Formeasurements of Spheniscus muizoni sp. nov. see Table 1. The material represents bones of at least two in dividuals. None of the bones provides any indication that the studied individuals of the new species are juveniles. No cra nial material is preserved; for this reason comparisons with S. megaramphus are impossible, which is only known by its skull and mandible. However, S. muizoni sp. nov. is dis tinctly smaller than both S. megaramphus and S. urbinai (Fig. 6). In the following, comparisons are priorly made with both extant and fossil species of Spheniscus, but also with ex tant taxa, with fossil South American taxa (Palaeosphe niscus, Paraptenodytes, Arthrodytes, and Eretiscus), and subordinate with the South African fossil taxa (Nucleornis, Dege, Inguza, and ?Palaeospheniscus huxleyorum), because the later are supposed to be closely related with Spheniscus by Olson (1985: 151).

Scapula (Fig. 2A).—Two fragmentary scapulae, a cranial end of a left scapula and a damaged cranial half of a right scapula, lacking most of the cranial end, are preserved (MNHN PPI 147b1−2). In cranial view, the cranial end is mediolaterally narrow; specifically the ventral part is slightly narrower than in S. humboldti and S. demersus, and more closely resembling S. magellanicus. The coracoid tubercle is elliptical and also mediolaterally narrow. The acromion is broken off in both specimens (which is of different length and direction within extant genera). Medially on the cranial end, a distinct but low crest runs from the acromion diago nally ventrocaudally to below the coracoid tubercle; such a distinct crest could not be observed in any of the compared extant penguin taxa. The glenoid process is mediodistally ob− late. The medial articular facet for the coracoid on the gle noid process is bordered medially and ventrally by a weak edge. The collum is quite mediolaterally flattened; the ven tral margin of the scapular blade carries a little projection about 20 mm caudally the cranial end.

Coracoid (Fig. 2F).—Fragmentary left and right coracoid of the partial skeleton (MNHN PPI 147a/1−2) and a distal frag mentary half of a worn left coracoid (MNHN PPI 148) are present. The left coracoid lacks the procoracoid process and the lateral angle of the distal end, the right one lacks the prox imal end. The coracoid corresponds well with the morphol ogy of the extant species of Spheniscus and of S. urbinai and S. chilensis. The acrocoracoid process is long and angled rectangular to the bone; its distal margin is almost straight in dorsal/ventral view. In proximal view, the ventral inflexion of the procoracoid process is a little less close than in the ex tant species of Spheniscus. The scapular cotyla is round and relatively large; the articular facet for the humerus is oval and flat and not well defined. Despite the procoracoid process is lacking in both specimens, the fracture at the right coracoid indicates a closed supracoracoid foramen as in all extant Spheniscus; its length can be estimated as about 6 mm. There is no osteological indication if there was a second foramen, larger and distally to the first one, as often developed in ex tant Spheniscus; however, its presence or absence can vary intraspecifically, following own observations. The distal end of the coracoid is dorsally strongly concave; the lateral pro cess is short and thin, and forms laterally a concave margin. The medial angle is thickened, proximodistally elongated and forms a pin−like process in proximal direction. Ventral to the medial angle, the surface of the coracoid forms a concav ity. The articular facet for the sternum is separated from the medial angle by a little indentation on the distal margin.

The coracoid of Spheniscus muizoni sp. nov. distingui shes from the fragmentary preserved coracoid of S. chilensis (Emslie and Correa 2003: fig. 2B) only by a somewhat shorter supracoracoid foramen. However, its length is vari able in extant Spheniscus species and therefore probably not diagnostic.

The closed supracoracoid foramen distinguishes S. mui zoni sp. nov. from Palaeospheniscus, in which it is open, as suggested by Simpson (1946: 49). In contrary to Ps. pata gonicus and Ps. bergi (figured in Moreno and Mercerat 1891: pl. 1: 23, 25), the coracoid of S. muizoni sp. nov. is smaller and more slender, especially at the level of the fora men.

Humerus (Figs. 2C, 4A).—Left and right humeri (MNHN PPI 147c1+2) are preserved; the right one is quite complete; only the cranial surface of the proximal end and the caudal surface of the shaft are damaged and the cranial edge of the ventral epicondylus is broken off; the left humerus lacks the ventral epicondylus and the caudal and ventral walls of the pneumo− tricipital fossa are broken off.As is typical for fossil and extant species of Spheniscus, S. muizoni sp. nov. has a weakly proxi− mally swollen humeral head, and also lacks the proximal notch between dorsal tubercle and humeral head (in caudal view; present in Palaeospheniscus); in both S. muizoni sp. nov. and S. urbinai the proximal outline from the head to the dorsal edge is less steep than in all compared extant species of Spheniscus. The pneumotricipital fossa is bipartite, with the cranial fossa being deep; this contrasts the genera Parapteno dytes and Arthrodytes, which fossa is non−bipartite (Simpson 1972: 18; Acosta Hospitaleche 2005: 407). The proximal bor der of the pneumotricipital fossa is in ventral view relatively straight and horizontal; this is also typical for the extant Spheniscus species, but contrasts the development of the fossil Spheniscus species, S. urbinai and S. chilensis, as well as Dege hendeyi and the extant penguins Eudyptes chrysocome, E. chrysolophus, and Aptenodytes patagonicus, where this bor der is concave proximally. Furthermore, in S. muizoni sp. nov. this margin, proximally bordering the pneumotricipital fossa, forms no ventrocaudally projecting lip (Fig. 5) separated from the head by an extended capital groove. As a result, the caudal surface of the proximal humerus end of S. muizoni sp. nov. is planar and shows no notch in dorsocaudal view. The lack of this ventrocaudally projecting lip is a common feature of S. muizoni sp. nov., S. urbinai, and S. chilensis, but contrasts all extant species of Spheniscus as well as those of Pygoscelis. The preaxial angle is weak (but stronger than in ?Palaeosphe niscus huxleyorum and D. hendeyi) and situated about the mid−point of the shaft and thus is placed more proximally than in all other species of Spheniscus, Palaeospheniscus, Eudyp tes, Pygoscelis, or Aptenodytes.

The ventral tubercle is strong and protruding, separated by a broad capital groove. S. muizoni sp. nov. differs from S. chilensis by a distinct concave indentation in the bicipital crest between ventral tubercle and shaft (Fig. 4A1, B1). This indentation is also present only in S. magellanicus, hinted in S. demersus and slightly in Pygoscelis adeliae, but is lacking in S. humboldti, Eudyptes, Py. papua, A. patagonicus, and Ps. huxleyorum. In proximal view, the sulcus for the trans versal ligament is short and deep; dorsal to it, a small, shal low dent is present like in Spheniscus, but which can vary in depth in other recent penguins and which can also form a short sulcus running distally, like in A. patagonicus or Py. papua. The fossa on the proximal caudal surface is as shal low as in all studied Spheniscus species, but is deeper when compared to Eudyptes, Pygoscelis, and Aptenodytes. As is typical for Spheniscus, the line defined by ventral and dorsal condyle (in cranial/caudal view) is steeper than in Eudyptes, Pygoscelis, and Aptenodytes. On the ventral epi condylus, the tip of the caudal−most process, caudally border ing the humerodtricipital sulcus, is slightly rising upwards (ventroproximally), most notably its distal margin; this differs from the ventral orientation in Spheniscus chilensis (Fig. 4A2, B2). The tip of the caudal−most process barely surpasses the ventral margin of the distal shaft (in cranial/caudal view), which corresponds to extant Spheniscus and Eudyptes species; it is longer and strongly surpassing in Aptenodytes patago nicus, moderately long in Pygoscelis papua, and shorter in Inguza predemersus. The middle process−like crest, caudally bordering the scapulotricipital sulcus, is slender, slightly poin ted, and extends relatively far distoventrally; it is somewhat longer and oriented slightly more distally than in S. chilensis (Fig. 4A2, B2). The cranial−most process−like crest, cranially bordering the scapulotricipital sulcus, is incomplete and miss ing its tip; however, the distal margin of the preserved part ends more proximally than that of the caudal process−like crest, unlike in S. chilensis (Fig. 4A1, B1). In distal view, the dorsal end of the scapulotricipital sulcus is curved caudally (Figs. 2C3, 4A4).

Ulna (Fig. 2B).—Preserved are a complete left ulna (MNHN IPP 147d), belonging to the associated skeleton, and a slightly smaller fragmentary left ulna (MNHN IPP 149), with damaged proximal end and proximal caudal margin. They show no important differences from those of extant Sphe niscus. The bone is flattened and expanded caudally into a crest−like flange; the ulna of S. muizoni sp. nov. is like the extant species of Spheniscus and like Eudyptes, relatively broader than the more slender ulnae of the representatives of Pygoscelis and Aptenodytes. In several extant penguin taxa the flange−like caudal margin shows in ventral or dorsal view in its distal half a slight indentation to which distally the bone becomes more slender; this recess is very weak in S. muizoni sp. nov. and situated at about the mid of the shaft, whereas it is situated more distally in S. humboldti, S. demersus, Eu dyptes chrysochome, E. chrysolophus, Py. adeliae, Py. pa pua, and A. patagonicus. The dorsal side carries at least two rows of marked pits for the attachment of feather quills. This can also be observed in most specimens of extant Sphe niscus, Pygoscelis, Eudyptula, and Eudyptes. The ventral side is marked by a longitudinal shallow concavity along the caudal margin of the flange−like caudal margin. The cranial side of the proximal end is flat; distoventrally to that, close to the cranial margin, there is a slender longitudinal brachialis impression. The distal shaft is dorsocranially swollen. Radius (Fig. 2D).—An almost complete left radius (MNHN PPI 150), only missing the cranial distal edge, is present; it shows no important differences from the radii of the com pared extant penguin species. In most of the latter the flange−like extended cranial crest ends proximally in a pro jecting peak. This is lacking in A. patagonicus and Py. ade liae, where the proximal end only forms an angle without projecting peak. This condition can also be observed in the radius of S. muizoni sp. nov., but it cannot be excluded that this character is liable to variation. As typical for penguins the dorsal side is marked by two furrows, one parallel to the cranial crest−like margin, the other one diagonal transversing the distal third of the radius from distocranially to proximo caudally.

Carpometacarpus (Figs. 2G, 4C).—An almost complete right carpometacarpus (MNHN PPI 154) is morphologically very similar to that of extant Spheniscus. Concerning limb bone length relations (Fig. 6), S. muizoni sp. nov. has a relatively short carpometacarpus, but which is relatively long for S. chilensis. The carpometacarpus of S. muizoni sp. nov. differs from S. humboldti and S. demersus by a short but distinct step between the proximal carpal trochlea and the extensor process of the alular metacarpal; such a step is low but present in S. magellanicus, and well developed in the compared species of Eudyptes, Pygoscelis, and Aptenodytes. The caudal margin of theminor metacarpal is less recessed below the carpal trochlea (Fig. 4C) in S. muizoni sp. nov. than in S. chilensis. In S. muizoni sp. nov. the line of fusion between alular digit and the major metacarpal is marked only by one foramen dorsally and none ventrally; thus, the fusion of these bones seems to be more progressive than in extant Spheniscus and also S. chi lensis, which show at least two but often more foramina on both the ventral and dorsal side (Fig. 4C). The carpometa carpus is broadest at its distal shaft. The spatiumis narrow and open, and shows no synchondroses, but which are present at least in both specimens of S. chilensis that were available for comparisons (UF 144110 figured in Emslie and Correa 2003: fig. 2C, and UF 144112).However, this character is more con stant in extant species (own observation), but is observed to be variable, e.g., for S. urbinai.

Femur (Figs. 3B, 4I).—The right femur is represented by a proximal portion and a distal end (MNHN PPI 147e); in addi tion, there is a distal end of a right femur of another individual preserved (MNHN PPI 151). Both distal ends are slightly damaged. The femur corresponds well with those of the extant species of Spheniscus. The proximal portion of the femur is relatively straight, in contrast to that of Py. papua whose prox imal half is curved medially. The femoral head and trochanter reach the same level in proximal direction as typical for all fos sil and extant species of Spheniscus, but unlike in Eudyptes chrysolophus, E. chrysochome, Py. papua, Py. adeliae, and A. patagonicus, in which the trochanter surpasses the head more or less distinctly. The trochanteric crest projects cranially but not proximally on the trochanter. A large and deep obturator impression is laterally bordered by a prominent crest. The depth and size of this impression seem to vary within the com pared extant species. The neck of the head is, in distal view, slightly thinner than in all compared extant specimens. Char acteristic for S. muizoni sp. nov. is the cranially very low, barely projecting medial condyle (in distal and medial view; Fig. 4I), lower than in the compared extant species of Sphe niscus as well as in S. chilensis (Fig. 4J) and S. urbinai. As in extant and fossil Spheniscus species, the lateral condyle crani ally reaches further proximally than the medial one, whereas the condyles reach about equally far proximally in most of the compared specimens of Eudyptes, Pygoscelis, and Apteno dytes. In distal view, the tibiofibular crest of the lateral condyle caudally projects relatively high and is sharp−crested; its me dial flank is steep like in extant Spheniscus, but steeper com pared to Eudyptes, Pygocelis, and Aptenodytes specimens. The sulcus of the fibular trochlea is relatively deep. There is a deep cranial impression for the attachment of the cruciate liga ment in the intercondylar sulcus.

Tibiotarsus (Figs. 3A, 4E).—The right tibiotarsus is almost complete, only the proximal end of the lateral and cranial cnemial crest are somewhat damaged, the left tibiotarsus is represented by its proximal end, with also damaged crests (MNHN PPI 147f−1+2). The proximal end is caudally charac terized by two fossae, separated by a longitudinal crest, which is relatively sharp−edged and proximally quite thin. The lateral longitudinal flexor fossa is also more or less developed in ex tant penguins; however, characteristic is the deep and distinct medial fossa (impression) of the collateral ligament (Fig. 4E2); it is mostly not or only very weakly developed in most com pared extant species (weak in S. magellanicus and in one of the two compared specimens of Py. papua). Also characteris tic is a distinct furrow along the lateral side of the distal shaft, starting on the caudal side distal to the fibular crest and turning to the laterocranial side in distal direction; this furrow is much weaker in all of the compared extant penguin specimens. Remarkable is the straight distal end of the tibiotarsus of S. muizoni sp. nov., contrasting most extant and fossil penguins (S. urbinai, Palaeospheniscus), which distal end is bent medi ally (in cranial view). Furthermore, the medial condyle is ver tical in S. muizoni sp. nov. and not laterally inclined and lacks a swollen tubercle for the retinaculum of the fibularis muscle forming the lateral bulge on the distal shaft (Fig. 4E1, E2); in stead of the latter there is only a very weak vertical line. In contrast, this lateral bulge is present in most extant Spheniscus species, Pygoscelis, Eudyptes, and Aptenodytes. Only S. ma gellanicus and S. chilensis also show a quite straight distal end of the tibiotarsus, provoked by very weak lateral tubercle for the retinaculum of the fibularis muscle, but differ by an in clined medial condyle (Fig. 4F1).

Unlike extant and other fossil penguin taxa, the distal end of the tibiotarsus of S. muizoni sp. nov. lacks the furrow be tween the tubercle for the retinaculum of the fibularis muscle and the extensor sulcus, and both the medial and lateral tuberosity for the extensor retinaculum on both sides of the extensor sulcus; the latter distinguishes S. chilensis, which has a distinct lateral tuberosity for the extensor retinaculum (Fig. 4E1, E2, F1, F2). The medial border of the extensor sulcus is relatively thin and straight. Like other fossil and ex tant species of Spheniscus, S. muizoni sp. nov. lacks the craniolaterally protruding lip−like crest (tuberositas retina culi extensoris medialis), but which is present in Eudyptes and also sometimes in Pygoscelis and Aptenodytes. The lat eral condyle is sub−rounded, the medial one is more longer craniocaudally; unlike in S. chilensis and Eudyptes, the distal margin of the medial condyle (in medial view) is not distally notched but round (Fig. 4E3, F3). A weak tubercle is present proximal to the lateral condyle, which is more or less devel oped also in extant penguin species. A tiny foramen is situ ated lateral to this tubercle. It is present in most penguins (ex cept Aptenodytes patagonicus), but its position varies from proximal to the tubercle in S. humboldti and S. demersus, to lateral of it in S. magellanicus, Py. papua, and Py. adeliae. Fibula.—The fibula is represented by a proximal portion of a right specimen (MNHN PPI 147g). The proximal articular facet for the femur is slightly damaged but was cranio caudally concave. The medial articular facet for the tibia is flat. The fibula of S. muizoni sp. nov. is characterized by a distinct and very deep concavity on the caudal side, which is less deep or almost absent in the extant species. The proximal shaft of the fibula is laterally marked by a distinct furrow run ning obliquely from proximocranial to distocaudal. This fur row is very well developed and also present in all compared species but can vary intraspecifically.

Tarsometatarsus (Figs. 3C, 4G).—One complete right tarso metatarsus (MNHN PPI 147h) is present. Fig. 6 shows that the tarsometatarsus of S. muizoni sp. nov. is relatively long in comparison with most extant Spheniscus species. It is slender and extends especially distally.

The elongation index of S. muizoni sp. nov. (2.07) is higher than that of S. urbinai (1.7–1.95; see Table 2), below that of Palaeospheniscus, which is diagnostically about 2.2–2.4 (Simpson 1972: 8) and distinctly below that of E. tonnii (elongation index: 2.66, Acosta Hospitaleche et al. 2004: 235), but is in the range of that of Paraptenodytes (elongation index: 1.8–2.1). However, Pa. robustus and Pa. antarcticus are distinctly larger species. The elongation in dex of S. chilensis is not known, but its ratio of length/least breadth of shaft (length: 33.5 mm, least breadth of shaft: 16 mm, unpublished data provided by Steve Emslie) is 2.09, whereas that of S. muizoni sp. nov. is 2.24.

The extensor sulci in S. muizoni sp. nov. are long and deep, as in Spheniscus. In contrary, both sulci are shallower in Eudyptes, Pygoscelis, and Aptenodytes and in Palaeo spheniscus (Ps. patagonicus and Ps. bergi) the medial sulcus is shallower. However, the medial sulcus of S. muizoni sp. nov. is in its distal part deeper than in extant species of Sphe niscus. Emslie and Correa (2003) described the extensor sulci of S. chilensis as shallow below the proximal foramen; from own observations, I would surmise that they are deep proximally but shallow only in their distal half. However, this is different from S. muizoni sp. nov., in which the sulci are also deep distally (Fig. 4G1, H1).

There are very probably two proximal vascular foramina, unlike in the smaller−sized species E. tonnii (Acosta Hospi taleche et al. 2004: 235) and N. insolitus (from the early Plio cene of Duinfontain, South Africa [Simpson 1979a]), which are characterized by only one foramen. In S. chilensis the me dial proximal foramen of the tarsometatarsus is described as “…greatly reduced or absent…” (Emslie and Correa 2003: 313). In S. muizoni sp. nov. the lateral proximal vascular fo ramen is probably situated in the middle of the lateral dorsal extensor sulcus, but where the bone is damaged; the medial one is small, situated more proximally than the lateral one in the most proximomedial edge of the infracotylar fossa and opens plantarly on the medial surface of the hypotarsus. The arrangement of the proximal foramina corresponds more to that of Spheniscus than to any other extant genus, but the me dial foramen is situated more proximal than in extant species. The position and size of these foramina also differs from Inguza predemersus.

The dorsal infracotylar fossa is very deep in that the proxi mal portion of the metatarsal III is distinctly lower than that of metatarsals II and IV. The proximal end of the metatarsal II dorsally carries a very pronounced impression of the extensor retinaculum (defining a distinct fossa) and a little distally to another small muscle or tendon scar; the arrangement and dis tinctiveness of these impressions and scars also corresponds mostly with the genus Spheniscus, whereas it is much weaker in Eudyptes chrysolophus, E. chrysochome, Py. papua, and A. patagonicus. The metatarsal III carries dorsally a long tubero sity for the attachment of the tibialis muscle.

The hypotarsus is relatively weak in comparison with ex tant Spheniscus species. The medial hypotarsal crest is rela tively thin and oriented medioplantarly (not lateroplantarly as in Palaeospheniscus); in proximal view it is bent hook−like medially with a concave furrow along its medial side; laterally attached to it (along its distal half) is a second very weak crest. The lateral hypotarsal crest is very weak and hardly project ing; laterally it is limited by a ridge, which runs distally pass ing medially to the lateral proximal foramen. Lateral to the ridge, the plantar surface of the proximal shaft is concave. The medial metatarsal (II) is more slender than the middle (III) and the lateral one (IV). In lateral view, a plantarly con cave ridge borders the lateroplantar side of the metatarsal IV. As typical for all extant and fossil Spheniscus species, the lateral intertrochlear incision is longer (deeper incised proxi mally) than the medial one (in dorsal view), whereas in A. patagonicus, E. chrysochome, and Py. papua they incise about the same length.

In S. muizoni sp. nov. as well as in all extant and fossil spe cies of Spheniscus, the trochlea II and IV are of about equal length distally; this is different in extant Eudyptes, Pygoscelis, and Aptenodytes and fossil Palaeospheniscus (Ps. patago nicus and Ps. bergi), which trochlea VI is somewhat shorter than II.

Trochlea III is very robust and swollen dorsally. In distal view, trochlea III reaches dorsally at the same level as the lat eral condyle of trochlea VI or slightly surpasses it (also in lat eral view). S. muizoni sp. nov. shares this feature with S. urbinai (and with E. chrysolophus), but differs from S. chi lensis and all extant species of Spheniscus, which trochlea IV dorsally surpasses the lateral condyle of trochlea III (Fig. 4G3, H3). Additionally, the lateral bulge of trochlea IV is dorsoplantarly lengthened and clearly surpasses trochlea III in plantar direction. The medial and lateral condyle of tro chlea III are dorsally about the same level, whereas in extant Spheniscus species, the medial one is clearly surpassing the lateral one. In lateral view, the lateral condyle of trochlea IV is distally somewhat flattened and shorter than the medial one (like in Pygoscelis papua and Aptenodytes), unlike in ex tant Spheniscus. Trochlea II is oriented relatively obliquely, so that the medial fovea is apparent in dorsal view.

Sternum (Fig. 2E).—Only a small, cranial−most center por tion of the sternum with both sulci for the coracoid articula tion is present. The cranial spines (spina interna and externa) and the carina are broken off.

Cervical vertebrae (Fig. 3H).—The probable two caudal most cervical vertebrae (C12, C13) are represented. C12 is quite complete, but lacking the transverse processes, of C13 only the body is preserved. Both vertebrae are strongly heterocoel and characterized by deep lateral dents on the lat eral concavities of the body, as typical for these vertebrae po sitions. C12 has a short spinosus and ventral processes. Thoracic vertebrae (Fig. 3F, G).—Seven thoracic vertebrae out of normally eight are preserved; they probably represent the cranial−most ones (T1–T7). All of them show the lateral depressions for the rib articulation. First and second verte brae (T1, T2) are preserved fragmentary; T1 lacks the dorsal arc, T2 its caudal half. T3–T7 (Fig. 3F) are more complete, but lack all ventral or dorsal spines and the transverse pro cesses; the cranial and caudal zygapophyses are mostly pre served. The first two vertebrae, even if fragmentary, show moderate heterocoel articulation facets, whereas the further five vertebrae are opisthocoel.

Synsacrum (Fig. 3D).—The synsacrum is represented by two fragments, a short portion of the cranial end and the rest of the fragmentary synsacrum, but both portions lack contact. The cranial portion is represented only by the laterally strongly flattened body with a worn convex cranial articular facet and ventrally forms a longitudinal crest. The larger cau dal portion lacks the body of the synsacrum, thus the verte bral canal is open; additionally, both lateral edges of the synsacrum are badly damaged; in its cranial half the spinosus crest is preserved, which becomes higher in the cranial direc tion. The sinoid structure forming the iliosynsacral suture is only partly preserved by their caudal portions, which are lat erally concave. The number of vertebrae involved in the synsacrum is not reconstructible.

Pygostyle (Fig. 3E).—About the cranial half of the pygostyle is preserved. It is laterally strongly flattened. The cranial ar ticulation is oval and higher than broad. The basis of the pygostyle is damaged.

Estimating the living weight based on the relationship of the body mass of birds with some hindlimb measurements (circumference of the shafts of femur and tibiotarsus), a method presented by Campbell and Marcus (1992), the fe mur leads to a weight of about 3500 g, and the tibiotarsus to a weight of about 3800 g for S. muizoni sp. nov.; these calcula tions were made by considering penguins as swimmers (SW in Campbell and Marcus 1992). The estimated living weight of S. muizoni sp. nov. corresponds well with the body mass of S. demersus, 3310 g for males and 2960 g for females, but is somewhat smaller than that for S. magellanicus (4500 g) and distinctly smaller than that for S. humboldti (4500 g), but is larger than that for S. mendiculus (2500 g; body masses by Dunning 1993).

Stratigraphic and geographic distribution.—Only known from the latest middle or earliest late Miocene of the type lo cality Cerro la Bruja.


Detailed comparisons with extant and fossil species of Sphe niscus suggest that the available postcranial bones of S. mui zoni sp. nov. morphologically correspond best with those of S. urbinai, aside from that the latter is distinctly larger. Some shared features are found to be exclusive for these two species within Spheniscus, such as the lack of a ventrocaudally pro jecting lip proximally bordering the pneumotricipital fossa on the humerus, or a tarsometatarsuswhich trochlea III is dorsally at about the same level as the lateral condyle of trochlea VI or slightly surpasses it. Although of almost identical size, S. muizoni sp. nov. can be distinguished by several osteological features from S. chilensis, such as the limb bone proportions of the carpometacarpus (Fig. 6) and morphological differences at the humerus, carpometacarpus, femur, tibiotarsus, and tarso− metatarsus. Based on the morphological similarity of S. mui zoni sp. nov. and S. urbinai and their stratigraphical succes sion within the Pisco Formation, it can be supposed that the first gave directly rise at least to the latter.

S. muizoni sp. nov. is the only known penguin species from the Cerro la Bruja locality and is unknown from any older or younger deposit in or outside the Pisco Formation. As pen guins are shorebirds—breeding on land (or ice) and feeding in fairly adjacent marine waters (del Hoyo et al. 1992), the pen guin fossils complement the exclusively marine vertebrate fauna represented at Cerro la Bruja, comprising fossil taxa of marine fishes (Teleostei) and sharks (Selachii), a turtle and a crocodile, different marine mammals such as fossil odonto cetes (Kentriodontidae, Pontoporiidae),mysticetes (Cetotheri idae, Balaenopteridae), pinnipeds (Phocidae), and a boobie (Aves, Sulidae; Muizon 1988: table 1).

To date, all fossil penguins reported from the different levels of the Pisco Formation belong to the extant genus Spheniscus. The today living species of Spheniscus include: S. mendiculus (Galapagos), S. humboldti (coast of Peru and Chile), S. magellanicus (coast of Chile and Argentina, Islas Malvinas or Falkland Islands), and S. demersus (coast of South Africa and Namibia; del Hoyo et al. 1992).

However, S. muizoni sp. nov. from the latest middle/earli est late Miocene of Cerro la Bruja is not only the most ancient penguin species in the Pisco Formation, but also the strati graphically oldest record for the extant genus Spheniscus in general.

The fossil record indicates a faunal change of Spheniscus specieswithin the stratigraphically older sequence of the Pisco Formation. S. muizoni sp. nov., exclusive in the oldest deposits (latest middle/earliest late Miocene, 13–11 Ma) is replaced by the persistant S. urbinai, existing from the late Miocene to the early Pliocene (9 to 3.5Ma) deposits (S. megaramphus is only known from a 6 Ma old locality). This replacement might be linked to changes of the climatic conditions at the Pacific coast of South America during the latest middle Miocene. Tsuchi (2002: 269) dated a short warm episode in between cool epi sodes of the surface marine climate of the latest middle Mio cene at ca. 11.5 Ma, which is supported by deposits rich in warm water planktonic foraminifera in northern Chile. Extant penguins (on a species level) require a relatively stable water temperature (del Hoyo et al. 1992: 144) and change of water temperature (e.g., as in the phenomenon known as El Niño) is known to have far reaching effects on the Humboldt and Galapagos Penguin. It is not known yet, if there is a faunal consequence also within some other fossil vertebrate groups represented in Cerro la Bruja and younger localities in the Pisco Formation because the vertebrate fauna of Cerro la Bruja has not been studied in detail as yet.

Apart from Spheniscus, there is only very scanty penguin material known from other Peruvian localities, which is de scribed as Palaeospheniscus sp. and Spheniscidae indet. (Acosta Hospitaleche and Stucchi 2005). However, the latter comes from deposits older than the Pisco Formation, the lower middle Miocene Chilcatay Formation and the upper Eocene–lower Oligocene Otuma Formation, respectively. Outside the Pisco Formation, fossil representatives of Sphe niscus have been reported only from Chile: S. chilensis from the late Pliocene and Spheniscus sp. from the late Miocene early Pliocene. There was only one report on a fossil Sphe niscus outside of South America, S. predemersus Simpson, 1971 described from South Africa. This species was, however, later referred by Simpson (1975b) to the new genus Inguza. Thus, the fossil record of Spheniscus is restricted to the lat est middle/earliest late Miocene to the late Pliocene of South America, more precise to the Pacific coast of South America (Peru and Chile). The modern distribution of Spheniscus cor relates with cold marine currents—the west−east directed Cir cumpolar Antarctic Current and their northern upturns along the west−coasts of South America (Humboldt Current) and southern Africa (Benguela Current). However, the restriction of the Mio−Pliocene record of Spheniscus to the Pacific coast of South America was not due to a geographic barrier, as the Drake passage was already open by the late middle Eocene around 40 Ma ago (Scher and Martin 2006). However, no re cord of Spheniscus is known along theAtlantic coasts of South America or southern Africa during the Neogene. However, the presence of this genus cannot be ruled out if their restriction to the Pacific coast during the Neogene is linked to other ecologi cal or alimentary factors—perhaps interrelated to the still open Panama Isthmus (closure starts during the early Pliocene, about 4–3 Ma, Kameo and Sato 2000) and thus corresponding marine currents, which might had an effect on the distribution of potential prey. In all extant Spheniscus species pelagic school fish (mostly anchovies) is the dominant prey; cephalo pods and crustaceans are subordinate (del Hoyo et al. 1992). Because the skull of S. muizoni sp. nov. is lacking, no informa tion is available on the shape of the bill, which varies at least in some extant penguin genera and their preferred diet. Neither fossil nor extant Spheniscus species show any obvious interre lation of body size and latitude of their habitat.


Acosta Hospitaleche, C. 2003. Paraptenodytes antarcticus (Aves: Sphenisci− formes) en la Formación Puerto Madryn (Mioceno tardío temprano), provincia de Chubut, Argentina. Revista Española de Paleontología 18: 179–183.

Acosta Hospitaleche, C. 2005. Systematic revision of Arthrodytes Ame− ghino, 1905 (Aves, Spheniscidae) and its assignment to the Parapteno− dytinae. Neues Jahrbuch für Geologie und Paläontologie, Monatshefte 7: 404–414.

Acosta Hospitaleche, C. and Canto, J. 2005. Primer registro de cráneos de Palaeospheniscus (Aves, Spheniscidae), procedentes de la Formación Bahía Inglesa (Mioceno Medio−Tardío), Chile. Revista Chilena de Historia Natural 78: 489–495.

Acosta Hospitaleche, C. and Stucchi, M. 2005. Nuevos restos terciarios de Spheniscidae (Aves, Sphenisciformes) procedentes de la costa del Perú. Revista Española de Paleontología 20: 1–5.

Acosta Hospitaleche, C. and Tambussi, C. 2004a. Fossil penguins from South America. Abstracts Book, Vth International Penguin Conference, Ushuaia, Argentina, 48.

Acosta Hospitaleche, C. and Tambussi, C. 2004b. Systematic revision of South American fossil penguins (Sphenisciformes). 6th International Meeting of the Society of Avian Paleontology and Evolution, Quillan, France, 3.

Acosta Hospitaleche, C., Chávez, M., and Fritis, O. 2006. Pinguinos fósiles (Pygoscelis calderensis sp. nov.) en la Formación Bahía Inglesa (Mioceno Medio−Plioceno), Chile. Revista Geológica de Chile 33: 327–338.

Acosta Hospitaleche, C., Tambussi, C., and Canto, J. 2005. Catálogo comentado de los pingüinos (Aves, Sphenisciformes) fósiles del Museo Nacional de Historia Natural de Santiago, Chile. Boletín del Museo Nacional de Historia Natural de Santiago 54: 141–151.

Acosta Hospitaleche, C., Tambussi, C., and Cozzuol, M. 2004. Eretiscus tonnii Simpson 1981 (Aves, Sphenisciformes): materiales adicionales, status taxonómico y distribución geográfica. Revista del Museo Argentino de Ciencias Naturales 6 (2): 632–637.

Acosta Hospitaleche, C., Tambussi, C., Donato,M., and Cozzuol,M. 2007. A newMiocene penguin from Patagonia and its phylogenetic relationships. Acta Palaeontologica Polonica 52: 299–314.

Ameghino, F. 1891. Enumeración de las aves fósiles de la República Argen− tina. Revista Argentina de Historia Natural 1: 441–453.

Ameghino, F. 1895. Sur les oiseaux fossiles de Patagonie. Boletín del Instituto Geografia Argentina 15: 501–602.

Ameghino, F. 1899. Sinopsis geológico−paleontológica. Suplemento (Adi− ciones y correcciones), 1–13. La Libertad, La Plata.

Ameghino, F. 1901. L’âge des formations sédimentaires de Patagonie. Anales de la Sociedad Científica Argentina 51: 20–39, 65–91.

Ameghino, F. 1905. Enumeración de los impennes fósiles de Patagonia y de la Isla Seymour. Anales delMuseo Nacional de Buenos Aires 3 (6): 97–167.

Ameghino, F. 1920. Sur les édentés fossiles de l’Argentine. Examen cri− tique, révision et correction de l’ouvrage de la M.R. Lydekker. Obras Completas y Correspondencia Cientifica 11: 447–909.

Baumel, J.J., King, A.S., Breazile, J.E., Evans, H.E., and Van den Berge, J.C. 1993. Handbook of avian anatomy: Nomina Anatomica Avium. Publications of the Nuttall Ornithological Club 23: 1–779.

Bonaparte, C.L. 1831. Saggio di una distribuzione metodica degli animali vertebrati. Giournale Arcadico di Science Lettere ed Arti 52: 155–189.

Campbell, K.E. Jr. and Marcus, L. 1992. The relationship of hindlimb bone dimensions to body weight in birds. In: K.E. Campbell Jr. (ed.), Papers in Avian Paleontology—Honoring Pierce Brodkorb. Natural History Museum Los Angeles County, Science Series 36: 395–412.

Cheneval, J. 1993. L’avifaune Mio−Pliocène de la formation Pisco (Pérou). Étude préliminaire. In: M. Gayet (ed.), Paléontologie et stratigraphie d’Amérique latine. Table ronde européenne (Lyon 1992). Documents des Laboratoires de Géologie Lyon 125: 85–95.

del Hoyo J., Elliott, A., and Sargatal, J. (eds.) 1992. Handbook of the Birds of the World 1. 696 pp. Lynx Edicions, Barcelona.

Driesch, A. von den 1976. Das Vermessen von Tierknochen aus vor− und frühgeschichtlichen Siedlungen. 114 pp. Institut für Paläoanatomie, Domestikationsforschung und Geschichte der Tiermedizin der Uni− versität München, Munich.

Dunning, J.B. Jr. (ed.) 1993. CRC Handbook of Avian Body Masses. 371 pp. CRC Press Boca, Raton.

Emslie, S.D. and Correa, C.G. 2003. A new species of penguin (Sphe− niscidae: Spheniscus) and other birds from the late Pliocene of Chile. Proceedings of the Biological Society of Washington 116: 308–316.

Forster, J.R. 1781. Historia aptenodytae. Generis avivm orbi avstrali proprii. Commentationes Societatis Regiae ScientiarumGottingensis 3: 121–148.

Kameo, K. and Sato, T. 2000. Biogeography of neogene calcareous nanno− fossils in the Caribbean and eastern equatorial Pacific—floral respons to the emergence of the Isthmus of Panama. Marine Micropaleontology 39: 201–218.

Linnaeus, C. 1758. Systema naturae per regna tria naturae, secundum classes, ordines, genera, species, cum characteribus, differentiis, syno− nymis, locis. Tomus I. Editio decima, reformata. 824 pp. Laurentius Salvius, Stockholm.

McDonald, G. and Muizon, C. de 2002. The cranial anatomy of Thalassocnus (Xenarthra, Mammalia), a derived nothothere from the Neogene of the Pisco Formation (Peru). Journal of Vertebrate Paleontology 22: 349–365.

Meyen, F.J.F. 1834. Beiträge zur Zoologie, gesammelt auf einer Reise um die Erde. 4. Abhandlung. Novorum Actorum Academiae Caesareae Leopol− dino−Carolinae Naturae Curiosorum 16 (Supplement 1): 1–312.

Moehring, P.H.G. 1758. Avium genera [Latin MS in 1752, 58 pp. Bremen]. Published Dutch translation: C. Nozeman and A. Vosmaer (eds.) 1758.

Geslachten der Vogelen. 97 pp. Pieter Meijer, Amsterdam. Moreno, F.P. and Mercerat, A. 1891. Catálogo de los pájaros fósiles de la República Argentina. Anales del Museo de La Plata (Paleontología Argentinia) 1: 8–71.

Muizon, C. de 1984. Les vertébrés fossiles de la formation Pisco (Pérou). II – Les odontocètes (Cetacea, Mammalia) du Pliocène inférieur de Sud− Sacaco. Travaux de l’Institut Français d’Études Andines 27: 1–188.

Muizon, C. de 1988. Les vertébrés fossiles de la formation Pisco (Pérou). III – Les odontocètes (Cetacea, Mammalia) du Miocène. Travaux de l’ Institut Français d’Études Andines 42: 1–244.

Muizon, C. de and DeVries, T.J. 1985. Geology and paleontology of late Ce− nozoicmarine deposits in the Sacaco area (Peru). Geologische Rundschau 74: 547–563.

Muizon, C. de, McDonald, G., Salas, R., and Urbina, M. 2003. A new early species of the aquatic sloth Thalassocnus (Mammalia, Xenarthra) from the Late Miocene of Peru. Journal of Vertebrate Paleontology 23: 886–894.

Noriega, G. and Tambussi, C. 1989. Un Spheniscidae (Aves, Sphenisci− formes) del Mioceno Tardio de la costa sur de Peru. VII Jornadas Argentinas Paleontologá Vertebrados. Abstract. Ameghiniana 26 (3–4): 247.

Olson, S.L. 1985. An early Pliocene marine avifauna from Duinefontein, Cape Province, South Africa. Annals of the South African Museum 95 (4): 147–164.

Scher, H.D and Martin, E.E. 2006. Timing and climatic consequences of the opening of Drake passage. Science 312: 428–430.

Sharpe, R.B. 1891. A review of recent attempts to classify birds. Proceedings of the Second International Ornithological Congress 2: 90. Budapest. Simpson, G.G. 1941. Large Pleistocene felines of North America. American Museum Novitates 1136: 1–27.

Simpson, G.G. 1946. Fossil penguins. Bulletin of the American Museum of Natural History 87: 1–99.

Simpson, G.G. 1970. Miocene penguins fromVictoria, Australia, and Chubut, Argentina. Memoirs of the National Museum of Victoria 31: 17–24.

Simpson, G.G. 1971. Fossil penguin from the Late Cenozoic of South Af− rica. Science 171: 1144–1145.

Simpson, G.G. 1972. Conspectus of Patagonian fossil penguins. American Museum Novitates 2488: 1–37.

Simpson, G.G. 1973. Tertiary penguins (Sphenisciformes, Spheniscidae) from Ysterplaats, Cape Town, South Africa. South African Journal of Sciences 69: 342–344.

Simpson, G.G. 1975a. Fossil penguins. In:B. Stonehouse (ed.), The Biology of Penguins, 19–41. University Park Press, Baltimore.

Simpson, G.G. 1975b. Notes on variation in pinguins and on fossil penguins from the Pliocene of Langebaanweg, Cape Province, South Africa. An− nals of the South African Museum 69 (4): 59–72.

Simpson, G.G. 1979a. Tertiary penguins from the Duinefontein, Cape Prov− ince, South Africa. Annals of the South African Museum 79 (1): 1–7.

Simpson, G.G. 1979b. A new genus of Late Tertiary penguin from Lange− baanweg, SouthAfrica. Annals of the SouthAfricanMuseum 78 (1): 1–9.

Simpson, G.G. 1981. Notes on some fossil penguins, including new genus from Patagonia. Ameghiniana 18: 266–272.

Stephan, B. 1979. Vergleichende Osteologie der Pinguine. Mitteilungen aus dem Zoologischen Museum in Berlin 55 (3): 1–98.

Stucchi, M. 2002. Una nueva especie de Spheniscus (Aves: Spheniscidae) de la Formatión Pisco, Perú. Boletín de la Sociedad Geológica del Perú 94: 17–24.

Stucchi, M., Urbina, M., and Giraldo, A. 2003. Una nueva specie de Spheniscidae del Mioceno Tardío de la Formacíon Pisco, Perú. Bulletin de l’Institut Français d’Études Andines 32 (2): 361–375.

Sundevall, C.J. 1871. On birds from the Galapagos Islands. Proceedings of the Zoological Society of London 1871: 124–129.

Tsuchi, R. 2002. Neogene evolution of surface marine climate in the Pacific and notes on related events. Revista Mexicana de Ciencias Geológicas 19 (3): 260–270.

Walsh, S.A. and Hume, J.P. 2001.Anew Neogene marine avian assemblage fromnorth−central Chile. Journal of Vertebrate Paleontology 21: 484–491.

Walsh, S.A. and Suárez, M.E. 2006. New penguin remains from the Plio− cene of Northern Chile. Historical Biology 18: 115–126.