Cranium of an Eocene/Oligocene pheasant-sized galliform bird from western North America, with the description of a vascular autapomorphy of the Galliformes

We describe the fossil cranium of a pheasant-sized galliform land bird from latest Eocene or earliest Oligocene marine rocks of the Jansen Creek Member, Makah Formation (Washington State, USA), which is the only three-dimensionally preserved cranium of a Paleogene representative of the Galliformes. The specimen was freed from a hard calcareous nodule with dilute formic acid. Micro-computed tomography provided further osteological details and a virtual cranial endocast. The fossil exhibits a plesiomorphic temporal morphology, lacking an ossified aponeurosis zygomatica, a feature characterizing some extant Cracidae and most Odontophoridae and Phasianidae. Overall, the fossil is most similar to the skull of the Asian phasianid taxon Arborophila, but this resemblance may well be plesiomorphic for a more inclusive clade. Still, we consider it possible that the fossil represents an archaic member of the Phasianoidea, in which case it would be the earliest record of this taxon from the New World. The fossil exhibits a previously unnoticed cranial autapomorphy of galliforms, a foramen in the temporal region that enables the vena profunda to enter the braincase, for which the name foramen temporale venosum is here introduced. Consistently present in all studied extant galliform taxa and absent in all other extant birds, this foramen enables a vascular connection between the brain and the ophthalmic rete, the latter playing an important role in thermoregulation of the avian brain.

Galliforms have a fairly comprehensive Paleogene fossil record from Europe, but fossils from the New World are scarce (Mayr 2009). Gallinuloides wyomingensis from the early Eocene Green River Formation in Wyoming is based on complete skeletons and belongs to the Gallinuloididae, which constitute the earliest diverging Cenozoic galliform taxon known so far (Mayr and Weidig 2004). A small middle Eocene stem group galliform from Utah is represented by a partial coracoid and was assigned to the taxon Paraortygidae (Stidham et al. 2020). Another small galliform from the late Eocene (Chadronian) of Canada, known from a fragmentary coracoid as well as a distal tarsometatarsus, was described as Nanortyx inexpectatus by Weigel (1963). Assignment of this latter species to crown group Galliformes, viz. the Odontophoridae, is not well based (Mayr 2009).
Various larger sized Paleogene galliform birds from North America were described from the late Eocene and early Oligocene of South Dakota and Nebraska. The best represented of these is Procrax brevipes, which is based on a partial skeleton from the late Eocene Chadron Formation of South Dakota (Tordoff and MacDonald 1957). Two further similar-sized species, Archaealectrornis sibleyi and Palaeonossax senectus from the early Oligocene Brule Formation of Nebraska and South Dakota, respectively, are only represented by isolated humeri (Wetmore 1956; Crowe and Short 1992). The affinities of these species are poorly resolved. Originally, they were assigned to the Cracidae (Procrax and Palaeonossax) and Gallinuloididae (Archaealectrornis), but neither classification is well based and detailed comparisons with Paleogene stem group Galliformes from Europe have not yet been performed (Mayr 2009).
Here we describe the skull of a pheasant-sized galliform discovered in marine sediments of the latest Eocene or earliest Oligocene (about 34 million years ago) Jansen Creek Member, Makah Formation, from the Olympic Peninsula in Washington State, USA. The Jansen Creek Member has yielded abundant remains of the Plotopteridae-large, flightless divers-as well as a few other marine birds (Goedert and Cornish 2002;Mayr 2015;Mayr et al. 2015;Mayr and Goedert 2016. The new specimen is the first land bird found in the Makah Formation and the only threedimensionally preserved skull of a Paleogene galliform bird. We, furthermore, report a previously unrecognized vascular character, which represents a cranial autapomorphy of Galliformes and is likely to be physiological significance in the living species.

Materials and methods
Institutional abbreviations: SMF, Senckenberg Research Institute, Frankfurt, Germany; UWBM, Burke Museum of Natural History and Culture, University of Washington, Seattle, USA.
Anatomical terminology follows Baumel et al. (1993). In addition to comparisons with all extant avian higher level taxa, the skulls of the following galliform species Tympanuchus cupido, Tympanuchus phasianellus. In addition, skulls of Macrocephalon maleo and Oreophasis derbianus were examined based on photographs of specimens in the Museum für Naturkunde, Berlin (Germany), and the Natural History Museum, Tring (UK), respectively. Comparisons with Leipoa ocellata and Aepypodius arfakianus (both Megapodiidae) are based on illustrations in Zusi and Livezey (2000) and Weber (1996).
Like other fossils from the Makah Formation, the skull was encased in a calcareous concretion. The latter was split by the collector and glued together before the matrix was dissolved with a 4% solution of formic acid buffered with tricalcium phosphate.
For micro-computed tomography (μCT), the specimen was scanned at the Senckenberg CT-Lab Frankfurt/Main with the ProCon-X-Ray-Micro-CT with a 100 kV Finefocus tube. The scans were performed with 90 kV and 89 μA, generating 1600 projections (endocast reconstruction) and 2400 projections (surface reconstruction) with 1500 ms per scan. To increase the signal to noise ratio, 3 exposures per image were averaged. While recording, the ring artefact reduction mode was applied (projection on different areas of the X-ray detector for each image). The resolution (voxel size) was 19.41 and 20.49 μm, respectively. The data sets were segmented and visualized with VGStudio MAX, version 3.4 (Volume Graphics, Heidelberg, Germany).

Locality and horizon
Beach terrace approximately 320 m northwest of the mouth of Jansen Creek, south shore of the Strait of Juan

Description and comparisons
The specimen consists of a cranium lacking the palatal bones as well as the rostrum, broken away at the nasofrontal hinge. As indicated by the completely co-ossified sutura frontoparietalis, it is from an adult bird (fissures along the dorsolateral surfaces represent cracks owing to the fact that the fossil was fractured in the discovery process). Because skulls of other Paleogene galliforms, that is, the gallinuloidid taxa Paraortygoides and Gallinuloides and the phasianoidean taxon Palaeortyx, are too poorly preserved for detailed comparisons, the following description is mainly based on comparisons with crown group Galliformes. The fossil agrees well with the skulls of some crown group Galliformes in proportions and morphological features (Fig. 1). No traces of an ectethmoid can be discerned, but the corresponding area is not well preserved (the ectethmoid is well developed in the Megapodiidae but reduced in the Cracidae and Phasianidae). Likewise, there are no well-defined articular facets for the lacrimals (in the Cracidae, the caudal portion of the lacrimal articulates with the frontal and is attached to the orbital rim; Fig. 1g).
Unlike in the Megapodiidae, the fonticulus orbitocranialis is well developed; in contrast to the Cracidae (Fig. 1h), there is no laterally open sulcus olfactorius. On the left side of the specimen, a kidney-shaped foramen is located caudal of the fonticulus orbitocranialis (Figs. 1d, 2g); this foramen is not present on the right side of the fossil. A fonticulus interorbitalis is absent. The large foramen opticum is similar in size and position to the foramen of crown group Galliformes.
The interorbital section formed by the frontal bones is mediolaterally narrow and measures only about one third of the maximum width of the skull. The interorbital section is equally narrow in many taxa of the Phasianidae (Fig. 1e), whereas it is wider in the Megapodiidae, Cracidae (Fig. 1g), and Numididae. The cranium has a sub-cylindrical shape and its dorsal surface is not as vaulted as it is in most crown group Galliformes. An equally flat dorsal surface of the cranium is found in the phasianid taxa Ammoperdix and Alectoris. Cross sectional images generated by micro-CT scans show that the bone walls of the dorsal portion of the neurocranium are very thick in the area of the parietals, as they are in at least some crown group Galliformes (only a cross section of the skull of Gallus gallus was available to us), whereas the skull roof of the fossil has thinner bone walls than G. gallus in the area of the frontals ( Fig. 3a-d).
As in other galliforms, the processus postorbitalis is a laterally protruding, subtriangular projection rather than a rod-shaped process. In the fossil, the tip of the right processus postorbitalis is broken, whereas the left one is complete. The lateral margin of the processus postorbitalis forms a ridge, as it does in many crown group Galliformes.
Overall, the morphology of the temporal region of the fossil (Fig. 4a) is most similar to that of extant Arborophila (Fig. 4g), which is one of the earliest diverging taxa of the Phasianidae (Wang et al. 2013). The processus zygomaticus is a dorsoventrally oriented, flange-like embossment, and there is no indication of an ossified aponeurosis zygomatica ( Fig. 4f, i), which characterizes many crown group Galliformes (Zusi and Livezey 2000;Elzanowski and Mayr 2018). An ossified aponeurosis zygomatica evolved multiple times within crown group Galliformes (see discussion) and is absent in the early Eocene stem group galliform Gallinuloides (Mayr and Weidig 2004) and the Oligocene Palaeortyx (Mayr et al. 2006), which is an early diverging representative of the Phasianoidea. A short processus suprameaticus is visible on the left side of the fossil, immediately caudal of the cotyla quadratica squamosi (of which only the lateral rim is exposed). Unfortunately, the tip of this process is broken, so we cannot determine whether it contributed to an osseous bridge delimiting the cranial margin of the tympanic cavity, as it does in many crown group Odontophoridae (except Ptilopachus; Fig. 4e) and Phasianidae (including Arborophila; Fig. 4g).
As in most extant Galliformes, the fossa temporalis is poorly developed. Furthermore, as in crown group Galliformes, there is a small vascular foramen in this fossa, between the processus postorbitalis and the processus zygomaticus ( Fig. 4; see discussion). This foramen is of similar size and position to the foramen in crown group Galliformes and is macroscopically visible (Fig. 4a). Owing to the fact that it is infilled by sediment of similar density to the bone, the foramen is not well delimited in the surface model created by micro-CT data (Fig. 2h), even though cross sectional images of this area show a canal that passes into the brain cavity (Fig. 2j). Unlike in extant galliform birds, the fossil exhibits a marked, pit-like depression rostral to this foramen (Figs. 2h, 4a).
The rostrum parasphenoidale exhibits rostrocaudally elongate and sessile processus basipterygoidei (Figs. 1b, 2f), a derived character of galloanserine birds. In the fossil, the articular facets of these processes are poorly developed, which is especially true for the right one that appears to be eroded. Other features of the basicranium, as well as details of the otic region, are not preserved in the fossil.  (red color). e, f Cranium of the extant Arborophila brunneopectus (Phasianidae; SMF 14588) in e dorsal and f left lateral view. g, h Cranium of the extant Ortalis canicollis (Cracidae; SMF 14247) in g dorsal and h left lateral view. dpr pitlike depression rostral to foramen temporale venosum, fio fonticulus interorbitalis, foc fonticulus orbitocranialis, fop foramen opticum, for kidney-shaped foramen caudal of fonticulus orbitocranialis, ftp fossa temporalis, ftv foramen temporale venosum, lac lacrimal, pbt processus basipterygoideus, ppo processus postorbitalis, pzg processus zygomaticus, sof sulcus olfactorius, uie unidentified element. The fossil specimen was coated with ammonium chloride. The scale bars equal 5 mm. [Color online] ◂ An elongate element situated next to the left otic region (Fig. 1d) cannot be identified. This structure may be mistaken for an ossified aponeurosis zygomatica, but in contrast to the latter, it is not flattened and sheet-like but tubular. The apparent bifurcation of this element and its clear separation from the processus zygomaticus furthermore preclude its identification as an ossified aponeurosis zygomatica. Judging from its slightly darker color, it may not be bone but a small piece of wood; fossilized wood is occasionally found together with avian remains in the Makah Formation.
The virtual cranial endocast (Fig. 5a, b) exhibits a morphology similar to that of extant Megapodiidae and Phasianidae ( Fig. 5c-f; Handley and Worthy 2021, see also Early et al. 2020). As in the latter, the shapes of the two telencephalic hemispheres give the rostral portion of the endocast a heart-shaped outline in dorsal and ventral view. In addition, pit-like depression rostral to foramen temporale venosum, for kidneyshaped foramen caudal of fonticulus orbitocranialis, ftv foramen temporale venosum, pbt processus basipterygoideus, ppo processus postorbitalis, pzg processus zygomaticus. Scale bar equals 5 mm in a-f and 1 mm in g-j. [Color online] as in crown group Galliformes, the wulsts are weakly developed. The mesencephalon (optic lobe) and cerebellum are likewise of similar shape and size to the corresponding brain parts of crown group Galliformes. Whereas there was a clear differentiation between the bones and the rostral portion of the endocast (telencephalic hemispheres) in the micro-CT scans, no unambiguous distinction between bone and surrounding matrix was possible in the area of the flocculus and our reconstruction of the shape of the latter structure is tentative. The bulbus olfactorius is not completely preserved.

Discussion
SMF Av 666 exhibits sessile basipterygoid processes, which are an apomorphy of galloanserines. The vascular foramen in the temporal fossa, between the postorbital and zygomatic processes, represents a distinctive autapomorphy of at least crown group galliforms, which has not been mentioned in earlier works on the galliform skull (e.g., Parker 1869; Jollie 1957) or more recent treatises of avian anatomy (e.g., Gheţie 1976;Baumel et al. 1993). It is identical to the "middle meningeal foramen" of Sedlmayr (2002: 182) and Porter andWitmer (2016: 1474) and opens into the skull cavity, being situated at the boundary between the squamosal and the laterosphenoid (Fig. 4h). The foramen enables the vena profunda to communicate between the ophthalmic rete and   (SMF 14306). apz ossified aponeurosis zygomatica, cqs cotyla quadratica squamosi, dpr pit-like depression rostral to foramen temporale venosum, fmc foramen musculi columellae, ftv foramen temporale venosum, mxm foramen nervi maxillomandibularis, oss osseous bridge delimiting the cranial margin of the tympanic cavity, ppo processus postorbitalis, psm processus suprameaticus, pzg processus zygomaticus, qdr quadrate, rcc recessus columellae. Scale bar in a represents 5 mm, images of extant taxa not to scale. [Color online] the cerebrotectal ("sphenotemporal") sinus (Sedlmayr 2002;Porter and Witmer 2016). Porter and Witmer (2016) noted its presence for Meleagris gallopavo and Gallus gallus, and here it is reported to be present in all examined species of crown group Galliformes; conversely, in all non-galliform birds this foramen is absent. To facilitate future anatomical discourse, the term foramen temporale venosum is introduced for this unique galliform character.
The physiological significance of this foramen has yet to be determined, but it is likely to be functionally correlated to the rete ophthalmicum, which is a vascular heat exchange system involved in the regulation of the temperature of the brain that also mitigates heat loss through the large eyes of birds (Kilgore et al. 1976;Midtgård 1983Midtgård , 1984. In rare cases, the foramen temporale venosum can be fairly large (a particularly large foramen was found in one individual of Odontophorus capueira), but usually its relative size shows little variation across different galliform taxa. Extant Galliformes live in a wide variety of habitats and climates, so an immediate physiological significance of the foramen is not apparent. However, galliforms exhibit other derived characteristics of the vascular system of the skull and, for example, lack the occipital ramus of arteria ophthalmica externa, with the latter artery supplying the ophthalmic rete (Mayr 2020). Because the ophthalmic rete plays a role in thermoregulation, a functional correlation may exist between the foramen temporale venosum and integumentary features, such as naked skin parts or wattles, even though there appears to be no obvious correlation between the size of the foramen and the aforementioned structures. A determination of the physiological significance of the foramen temporale venosum is certainly a promising venue for future studies, which also need to assess a possible functional interdependence of some autapomorphic vascular characters in the skull of galliforms.
Future studies will also have to show when exactly a foramen temporale venosum evolved in the stem lineage of galliform birds, that is, whether it is an apomorphy of crown group Galliformes or whether it is present in some taxa of the stem lineage. Currently, skulls of Cenozoic stem group galliforms are only known from the Gallinuloididae (Mayr and Weidig 2004) and Sylviornithidae (Mourer-Chauviré and Balouet 2005; see Mayr 2011 andWorthy et al. 2016 for the presumed stem group position of this taxon). However, it is not possible from the published photographs to assess the occurrence of a foramen temporale venosum in these birds and direct examinations of the fossils are required for firm conclusions about the presence or absence of this foramen in the above taxa.
The fossil (SMF Av 666) is from a species the size of a female Common Pheasant, Phasianus colchicus. In its size, it would correspond to the late Eocene and early Oligocene Procrax brevipes, Archaealectrornis sibleyi, and Palaeonossax senectus from Nebraska and South Dakota (see introduction), but because the skulls of these species are unknown, a species-level assessment of SMF Av 666 is not possible.
Whereas Procrax and Palaeonossax were likened to the Cracidae by earlier authors, various late Eocene and early Oligocene taxa from Europe clearly represent stem group Galliformes (Mayr 2009). With the phasianoid taxon Palaeortyx, however, the first unambiguous crown group Galliformes likewise occur in the early Oligocene of Europe. From an evolutionary and biogeographic point of view, it is therefore of interest whether galliforms from coeval North America faunas represent stem or crown group taxa.
In the stem group galliform taxon Gallinuloididae and in extant Megapodiidae, the ectethmoid bone is well developed. Presumably, this represents the plesiomorphic condition for Galliformes, because a well-developed ectethmoid is found in most non-galliform neognathous birds. A reduced ectethmoid bone would suggest a position of SMF Av 666 in the galliform crown group, within a clade formed by the Cracidae and Phasianoidea. However, the corresponding area of the fossil is not well preserved, and we cannot exclude the possibility that the ectethmoid is broken away. Other osteological characters visible in SMF Av 666 do not allow an unambiguous assignment to the galliform stem or crown group.
Galliform and anseriform birds are characterized by a short or completely reduced processus zygomaticus (Zusi and Livezey 2000). In most extant Galliformes, an ossified aponeurosis zygomatica inserts on the short processus zygomaticus, which often co-ossifies with the postorbital process to form an orbitozygomatic junction. This junction evolved independently in some Cracidae, all Odontophoridae except for Ptilopachus (which was long assigned to the Phasianidae, but is now recovered as an Old World representative of the Odontophoridae; Cohen et al. 2012), and many Phasianidae (Elzanowski and Mayr 2018).
Ossification of the zygomatic aponeurosis occurs in postnatal development (Zusi and Livezey 2000;Elzanowski and Mayr 2018), and among the extant galliform species studied, an ossified aponeurosis zygomatica is absent in adult individuals of Macrocephalon (Megapodiidae), some Cracidae (Ortalis, Oreophasis), all Numididae, as well as Ptilopachus (which is the sister taxon of New World Odontophoridae; Cohen et al. 2012) and Arborophila (which, together with Rollulus and Caloperdix, is one of the earliest diverging taxa of the Phasianidae; Wang et al. 2013). As noted in the description, an ossified aponeurosis zygomatica is absent in the early Eocene stem group galliform Gallinuloides and in the Oligocene phasianoidean taxon Palaeortyx (Mayr and Weidig 2004;Mayr et al. 2006). Most likely, therefore, the absence of an ossified aponeurosis zygomatica is plesiomorphic for crown group Galliformes and this character evolved multiple times independently, in the Megapodiidae, Cracidae, Odontophoridae, and Phasianidae. Functionally, this variation of the temporal region may be correlated with different feeding strategies of galliform birds, even though comparative functional morphological studies still have to be performed.
The absence of an ossified aponeurosis zygomatica in SMF Av 666 is also likely to be a plesiomorphic trait. With regard to the morphology of the temporal region, in particular the shape of the processus zygomaticus, the fossil most closely resembles the Asian Arborophila (Phasianidae), with which it also agrees well in the overall shape of the cranium (Fig. 1). We consider it possible that the fossil represents an archaic member of the Phasianoidea, in which case it would be the earliest record of this taxon from the New World. However, we note that most similarities to Arborophila are plesiomorphic and it is not possible to unambiguously constrain the affinities of the fossil with derived characters.
As long as the affinities between SMF Av 666 and the coeval North American taxa Procrax, Palaeonossax, and Archaealectrornis are not better constrained, a robust phylogenetic placement of the new fossil is not possible. At least Procrax brevipes lacks derived characteristics of the Phasianoidea, such as a well-developed intermetacarpal process (carpometacarpus), and Archaealectrornis is likely to even be outside crown group Galliformes (Mayr 2009).
Further light on the affinities of SMF Av 666 may be shed by another fossil from the Jansen Creek Member of the Makah Formation east of Neah Bay (Washington State), of which only a cast is available (UWBM 88804-C1009). The original unprepared specimen was in a private collection but its current whereabouts are unknown. It consists of wing and pectoral girdle bones in a split nodule and-judging from the shapes of the humerus and furcula as well as the short ulna-appears to represent a galliform (Fig. 3e). With a humerus length of approximately 56 mm, it is from a species that was smaller than Procrax brevipes (79 mm; Tordoff and MacDonald 1957) and Archaealectrornis sibleyi (77.3 mm; Crowe and Short 1992). Palaeonossax senectus is only known from the distal end of the humerus, which is not well exposed in UWBM 88804-C1009 making meaningful size comparisons impossible. The furcula of UWBM 88804-C1009 is more strongly U-shaped that that of crown group Galliformes (which is V-shaped) and the apophysis furculae is proportionally smaller than in most extant galliform species. It is not possible to reliably assess the affinities of the fossil from a study of the cast alone, because critical parts of the bones are not exposed. Still, we are confident that if a future examination of the original fossil eventually becomes possible, it will contribute to a better understanding of the phylogenetic affinities of Eocene and Oligocene North American galliforms. Washington) for discovery and collection of the fossil specimen and donating it to SMF. Olaf Vogel (SMF) is thanked for its preparation. The photos were made by Sven Tränkner (SMF). Lawrence Witmer and Ruger Porter are thanked for confirming the identification of the temporal foramen as the structure they termed "middle meningeal foramen". Christian Sidor and Katie Anderson (UWBM) enabled the study of fossil specimens in UWBM. Comments from two anonymous reviewers improved the manuscript.
Funding Open Access funding enabled and organized by Projekt DEAL.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.