Phoberodon arctirostris, CABRERA, 1926

PHOBERODON ARCTIROSTRIS CABRERA, 1926

( FIGS 2–10 View Figure 2 View Figure 3 View Figure 4 View Figure 5 View Figure 6 View Figure 7 View Figure 8 View Figure 9 View Figure 10 ; TABLES 1–3)

Holotype: MLP 5-4: incomplete skull (including maxillae, premaxillae, palatines, nasals, ethmoid, vomer, frontals, a small portion of the supraoccipital and exoccipitals, squamosals and a small portion of the sphenoids), almost complete mandibles, 35 vertebrae (five cervical, eight thoracic, 12 lumbar and ten caudal), four chevrons, ribs, incomplete right scapula and right humerus and radius. Cabrera (1926) figured the lacrimojugals, but both are now lost.

Referred specimens: MLP 5-3: nearly complete skull (including most of the maxillae and premaxillae, palatines, nasals, ethmoid, vomer, frontals, a portion of the supraoccipital and exoccipitals, squamosals, part of the pterygoid and sphenoids), incomplete mandible, 34 vertebrae (four cervical, ten thoracic, 11 lumbar and ten caudal), one chevron, ribs and manubrium. Cabrera (1926) mentioned 36 vertebrae for this specimen, but only 34 could be located.

MPEF-PV 10883: incomplete skull (including maxillae, premaxillae, nasals, frontals, ethmoid, a portion of right lacrimojugal, a portion of the vomer and right orbitsphenoid and a small portion of the supraoccipital), incomplete left mandible, 15 vertebrae (three cervical, six thoracic, four lumbar and two caudal), ribs, a right humerus and both incomplete scapulae. The skull presents a little diagenetic deformation on the rostrum and to a lesser degree in the cranium.

Geographical and stratigraphic occurrence: The holotype and MLP 5- 3 specimens are from Cerro Castillo (43°20 ′ 12.3 ″ S, 65°19 ′ 10.7 ″ W), to the south of Trelew city, whereas MPEF-PV 10883 was collected on Magagna beach (43°23 ′ 52 ″ S, 65°2 ′ 55 ″ W). All localities are in Chubut province, Patagonia, Argentina; Gaiman Formation (early Miocene; Mendía & Bayarsky, 1981; Scasso & Castro, 1999; see ‘Geological setting’ section for more details on the stratigraphic occurrence of the specimens).

Emended diagnosis: Phoberodon arctirostris differs from OU 21798, Prosqualodon davidis Flynn, 1923 , Squalodon calvertensis Kellogg, 1923 , I. vertizi and platanistoids (the contents of Platanistoidea of the present work are defined in the ‘Phylogenetic analysis’ section below) in the following unique combination of apomorphies: long and wide rostrum with wide premaxilla at mid-length, maxilla and premaxilla are unfused along the rostrum, medially straight mandible, absence of an antorbital process of maxilla, presence of an expanded tympanosquamosal recess (except I. vertizi ), lacrimal restricted below the supraorbital process of maxilla (except I. vertizi ), frontal forming the dorsolateral edge of the ventral infraorbital foramina (except I. vertizi ), pointed postorbital process of frontal (except OU 21798), absence of a maxillary crest (except OU 21798) and convex ventral edge of zygomatic process (except I. vertizi ). It further differs from P. davidis , S. calvertensis and platanistoids in having: a posterolateraly oriented postorbital process of frontal, wide right premaxilla posterior to premaxillary foramen, lack of a maxillary intrusion, shallow neck muscle fossa, most posterior portion of alisphenoid–squamosal suture anterior to foramen ovale and acromion horizontally oriented. Phoberodon arctirostris further differs from P. davidis , S. calvertensis and OU 21798 in having: a longitudinal groove on the ventral side of the mandible, an anteromedially oriented orbit and a dorsal infraorbital foramina near the posterior portion of the premaxilla. Phoberodon arctirostris differs from P. davidis , S. calvertensis , I. vertizi and platanistoids by having: a wide rostrum at the base (except I. vertizi ), rostral constriction anterior to the maxillary flange, vertex skewed to the left (except I. vertizi ), a premaxillary foramen anterior to the antorbital notch (except P. davidis and I. vertizi ) and squared-off postglenoid process of squamosal (except P. davidis ). It differs from OU 21798, S. calvertensis , I. vertizi and platanistoids in having: the maxilla forming the posterior wall of the antorbital notch and in the moderately concave posterior sinus fossa. Phoberodon arctirostris differs from OU 21798 and platanistoids in the absence of a posterior portion of the periotic fossa. Phoberodon arctirostris further differs from P. davidis and S. calvertensis in having: a transversely inflated premaxilla at the rostrum, nasals with a concave posterior margin and presence of a supraspinous fossa of the scapula; and from OU 21798 and S. calvertensis in having a deep posterolateral sulcus. Phoberodon arctirostris differs from OU 21798 and I. vertizi in having a longer maxilla on the rostrum. It further differs from OU 21798 in having: three anterior dorsal infraorbital foramina, a dorsally exposed ethmoid, a shallow premaxillary cleft, an unfused mandibular symphysis and presence of a pterygoid sinus fossa anterior to the internal nares. Phoberodon arctirostris differs from P. davidis in having: striated tooth enamel, mandibular teeth of about the same size, a premaxilla of constant width along the cranium, lateral lamina of the palatine fused to the maxilla, shallow anterior slope of the scapula, a coracoid process not expanded distally plus a distally expanded acromion; and from S. calvertensis in having: teeth with an entocingulum, nasals at the same height as frontals, frontals forming the anterodorsal wall of the braincase, temporal fossa shorter than anteroposteriorly long, ventrally exposed palatine, and a dorsolaterally developed dorsal transverse process of the atlas. Phoberodon arctirostris differs from I. vertizi in having: the premaxillae widely separated by the mesorostral groove, presence of anteriorly oriented and procumbent incisors, both premaxillae extending posterior to nasals, presence of a premaxillary sac fossa, temporal fossa not roofed over by lateral expansion of the maxilla, and a wide external auditory meatus. It further differs from platanistoids and I. vertizi in having a high coronoid process on the mandible; and from platanistoids in having heterdont teeth with a short crown.

Physical maturity: The preserved cranial sutures are closed but distinct in all the specimens; in MPEF-PV 10883, the frontal–nasal suture is also closed. The external surface of the left occipital condyle preserved on the holotype and the articulation surfaces of the skull and scapula of all specimens, and the humerus and radius of the holotype have smooth surfaces, without the coarse pitting that might indicate juvenile bone. The nuchal crest is distinct in the holotype and MLP 5-3, but in MPEF-PV 10883 it is poorly developed (whether ontogenetic or postmortem bioeroded is uncertain). The epiphyses of all the vertebrae of the holotype and MPEF-PV 10883 are fused, and of all the vertebrae except three thoracic vertebrae (one of them has the epiphyses partly fused) and two lumbar vertebrae (one only the posterior epiphyses) of MLP 5-3 are fused. All these features suggest that P. arctirostris is represented by stage V individuals, following Perrin’s (1975) stages for Stenella attenuata (Gray, 1846) .

Body size: Given that none of the specimens has an articulated squamosal, we estimated the bizygomatic width of the holotype skull with the specimen on its metal mount. We applied the formula proposed by Pyenson & Sponberg (2011) for stem Odontoceti: log(TL) = 0.92 × [log(BIZYG) − 1.72] + 2.68. The estimated bizygomatic width (BIZYG) of P. arctirostris is 331 mm, giving a reconstructed total length (TL) of 3.13 m, that is, a little more than three times the cranial length. Cabrera reported a holotype skull length of 994 mm, whereas we obtained 1200 mm. Also, Cabrera (1926: 389) indicated a body length of ~4.25 m. If we sum the skull condylobasal length (CBL) plus the length of all the vertebrae preserved, we estimate a total length of 3.3 m for the holotype and 3.1 m for MLP 5-3. Given that the postcranial skeleton is incomplete, the estimated total length from the formula of Pyenson & Sponberg (2011) might be slightly underestimated. Nonetheless, the estimated total length is comparable amongst living odontocetes to Delphinapterus leucas Pallas, 1776 , Pseudorca crassidens Owen, 1846 , Mesoplodon hectori Gray, 1871 and Kogia breviceps de Blainville, 1838 , among others ( Perrin et al., 2008).

DESCRIPTION