Talpanas lippa Iwaniuk, Olson & James, 2009, sp. nov.

Talpanas lippa, sp. nov. (Olson & James)

Holotype. USNM 535683; complete neurocranium lacking quadratojugals and any elements of the palate (Fig. 1) and with the surface of the basitemporal plate eroded away.

Locality. Hawaiian Islands, Kauai, south coast. Maha'ulepu/Pa'a district, Makauwahi Cave (Burney et al. 2001), 5 km ESE of the town of Koloa, 21˚53'17.5'' N 159˚25' 08.5''W.

Chronology and stratigraphy. Mid to late Holocene (Burney et al. 2001). All specimens are from Excavation EP (5) at depths of 2.6 to 3.5 m below datum (= 1.4 to 1.7 m below sea level). A fruit of Cordia from the same square and level as the holotype had a calibrated radiocarbon age range of 5490-5305 YBP. The paratypes were all from somewhat deeper in the deposits.

Etymology. L. lippus, nearly blind.

Distribution. Known only from the type-locality on the island of Kauai, Hawaiian Islands.

Diagnosis. As for the genus. Body mass estimates compared with the mass of living species (Dunning 2008) as well as standard markers, such as diameter of the acetabulum (see Materials and Methods), indicate a medium-large duck (692–1072 g) about the size of a Northern Pintail (Anas acuta) or female Mallard (Anas platyrhynchos), but with a much shorter, stouter leg.

Measurements (mm) of holotype. Overall length 36.5; width of brain case 26.7; depth of brain case 19.9; minimum distance between the two orbits (taken dorsally) 10.5; width of nasofrontal hinge 12.5; width across lacrimals 15.4; width across supraorbital processes 12.7; width across postorbital processes 29.7.

Paratypes. Figure 2 shows the six paratypes in comparison with elements of the Laysan Duck (Anas laysanensis). The paratypes are as follows: fused vomer and right and left palatine bones (broken and repaired) USNM 535684 (Figs. 2 B,C); right postdentary portion of mandible USNM 535685 (Figs. 2 E, G); fragment of left innominate with antitrochanter and portion of acetabulum USNM 535688; proximal end of right tibiotarsus USNM 535689 (Fig. 2 L); right tarsometatarsus USNM 535690 (Fig. 2 I); left tarsometatarsus USNM 535691 (Fig 2 J); and right innominate lacking most of pubis and anterior margin of ilium USNM 535686 (Fig. 2 N).

Measurements (mm) of paratypes. Palatines: length 18.6; anterior width 2.9, posterior width (lateral edge of palatines) 19.3, posterior width (medial edge of palatines) 12.0. Mandible: least depth of ramus 4.2; width of articulation including medial process 8.4; length of retroarticular process 5.3; depth of retroarticular process at base 3.2. Pelvis: length from posterior margin of acetabulum to posteriormost edge of ischium 30.4; depth of ilium just anterior to acetabulum 12.4; width of ilium just posterior to antitrochanter 7.5; depth of antitrochanter 3.9; greatest diameter of acetabulum 6.0; length and depth of ilioischiadic foramen 16.1 x 7.4. Tibiotarsus: greatest width and depth of articular surface not including cnemial crests 8.7 x 10.0; width and depth of shaft proximal to fibular crest 5.7 x 3.8. Tarsometatarsi (535691 in parentheses): length 32.9 (28.7); proximal width 8.9 (8.0); proximal depth through hypotarsus 8.9 (---); least width and depth of shaft 5.7 x 3.4 (5.0 x 2.9); distal width 9.8 (8.5); width of middle trochlea 3.5 (3.3).

Description. The length of the tarsometatarsus, relative to its midpoint circumference, was significantly smaller in Talpanas than in other anseriforms (Fig. 3 A), thus indicating a relatively short and stout leg. The disparity in size of the two known tarsometatarsi of T. lippa (Fig. 2 I, J) suggests that the species may have been sexually dimorphic, as in many waterfowl in which the males are larger (Dunning 2008). The morphology of the pelvic appendage argues against underwater foraging behavior in Talpanas. The lack of medio-lateral compression or medial flattening of the tarsometatarsus, the lack of elevation and retraction of the inner trochlea, the shallow cnemial crests, and the broad posterior half of the pelvis are opposite to the modifications observed in typical foot-propelled diving ducks. The trochleae of the tarsometatarsus are rather widely splayed and, although that for digit II is somewhat elevated it is not retracted with a bladelike posterior edge, as in diving ducks. The overall similarity of the pelvic appendage lies with graviportal terrestrial species.

The retroarticular process of the mandible is much shorter and less curved than in typical waterfowl (Fig. 2), probably reflecting a decrease in use of M. depressor mandibulae for the rapid bill movements that occur during feeding in waterfowl (Goodman & Fisher 1962). The vomeropalatines are also highly distinctive; the palatines are narrowly tapered posteriorly and are fused, unlike those of typical waterfowl, and the palatines diverge at a very wide angle anteriorly (Figs. 2 B, C), which is significantly greater than in other anseriforms (Fig. 3 B, Table 2). This may indicate a broadening of the base of the bill, but may also be a reflection of the great foreshortening of the anterior portion of the cranium so that a greater angle would be required to reach the same point of attachment to the base of the bill.

FIGURE 5. Scatterplots of the (A) maxillo-mandibular (nV) foramen and (B) optic foramen (nII) cross-sectional areas (mm2) against foramen magnum cross-sectional area (mm2). The maxillo-mandibular foramen is significantly larger (A) and the optic foramen is significantly smaller (B) in Talpanas (‘T’) relative to foramen magnum area.

The braincase of Talpanas is unique in several respects. Plots of the first three principal components from an analysis of eight measurements show that Talpanas is significantly different from other anseriforms in multivariate space (Fig. 4). Specifically, Talpanas has a wide nasofrontal hinge, interorbital region, braincase and postorbital processes combined with a short skull, narrow supraorbital process and lacrimal width, and a shallow braincase. Based on the position and size of the fossa for the nasal gland, which demarcates the lateral border of the orbit (Witmer 1995), the eyes were small and laterally displaced (Fig. 1) compared with other anatids. Talpan as lacks the fenestra on either side of the foramen magnum (fonticulus occipitalis) characteristic of many anatids, but these are absent in a number of species and may even be individually variable.

The most unusual aspects of Talpanas concern the relative size of the foramina associated with the optic (nII) and trigeminal (nV) nerves as shown by the cross-sectional area of the maxillo-mandibular (M-M) foramen and the optic foramen. The cross-sectional area of the M-M foramen, the primary exit point of the trigeminal nerve, is significantly larger in Talpanas relative to skull length and foramen magnum area than in any other anseriform (Fig. 5A). In contrast, the relative cross-sectional area of the optic foramen is significantly smaller in Talpanas than in other anseriforms (Fig. 5B). Thus, Talpanas has both the smallest optic foramen and largest trigeminal foramen of any anseriform examined.