Definition: “Pan-Charadriiformes” refers to the clade containing all birds more closely related to Charadrius hiaticula and Sterna hirundo than to flamingos and grebes (Mirandornithes) based on the Prum et al. (2015) reference phylogeny.

Referred Specimen: IGM 100/1435 (distal end of left humerus; fig. 2).

Measurements: Dorsoventral width of distal end, 4.7 mm.

Remarks: IGM 100/1435 is broken immediately proximal to the dorsal supracondylar process. This humerus is from a bird slightly smaller than the smallest extant sandpiper, Calidris minutilla (Least Sandpiper). The fossil exhibits a well-developed dorsal supracondylar process, which projects from the shaft well proximal to the dorsal condyle. The tip is rounded and is directed proximally, and it has a slight sigmoidal curve along the base of its dorsal margin. The fossa for the m. brachialis is preserved as a shallow scar with a well-demarcated ventral terminus and fainter margins proximally and dorsally (fig. 2A, B). The dorsal and ventral condyles are of similar maximum length (fig. 2A, B, D, E), and the flexor process is short and broad. The olecranon fossa is deep and undercuts the edge of the flexor process, giving that process a dorsally hooked margin in caudal view. A ridge separates the sulcus for the m. scapulotricipitalis from the sulcus for the m. humerotricipitalis. The sulcus for the m. scapulotricipitalis is narrower than the sulcus for the m. humerotricipitalis (fig. 2D, E).

Comparisons: A well-developed dorsal supracondylar process as in IGM 100/1435 is observed in shorebirds, tubenoses (Procellariiformes), swifts and hummingbirds (Apodi-formes), and songbirds (Passeriformes) among extant birds. However, the new fossil humerus is otherwise dissimilar in size and proportion to those found in swifts, hummingbirds, and songbirds; specifically, the dorsal supracondylar process is more strongly projected as well as being angled more cranially than dorsally. In the new fossil, this process is narrow and projects proximally, consistent with shorebirds (fig. 2C); by contrast, in tubenoses this process is broad and projects more cranially (Hou and Ericson, 2002). Additionally, this fossil shows a distinct sulcus for the m. scapulotricipitalis, which is observed in shorebirds but not in tubenoses (Zvonok et al., 2015). The new fossil can be further differentiated from most tubenoses by the presence of a shallow fossa for the m. brachialis; among nonalbatross tubenoses, this fossa is deep. IGM 100/1435 also exhibits a ventral condyle that projects distally past the dorsal condyle, consistent with shorebirds but not tubenoses, in which these two condyles show similar distal projections (Zvonok et al., 2015). Features of the distal humerus alone are not diagnostic for most lineages within shorebirds (Zvonok et al., 2015; Mayr, 2016), so further taxonomic assignment for the new specimen is not possible. Because this fossil cannot be assigned to crown-group charadriiforms, we place this new fossil within Pan-Charadriiformes.

A slightly younger (∼2–3 Ma) distal humerus (SMF Av 619) was reported from the early Eocene Nanjemoy Formation of Virginia (Mayr, 2016). This specimen is similar in size to IGM 100/1435 but has a shallower fossa for the m. brachialis and the dorsal supracondylar process has a slightly narrower base.

Referred Specimen: IGM 100/1268 (furcula missing both omal extremities; fig. 3).

Measurements: Maximum preserved length of right ramus, 13.7 mm; maximum length of furcular apophysis, 3 mm; maximum preserved interclavicular width from lateral margins of rami, 7.5 mm.

Remarks: The apophysis of IGM 100/1268 is slightly crushed but retains its shape in lateral view. Mechanical preparation of the fragile rami and apophysis is impossible, and so these parts of the specimen remain encased in rock. Digitally “prepped” renderings derived from high-resolution CT data were used for comparison with other birds. The fossil is from an individual slightly smaller than the extant Least Sandpiper. The furcula is a narrow U-shape in cranial view. Toward the apophysis, the rami mediolaterally broaden and dorsoventrally flatten (fig. 3A, C). The furcular apophysis is bladelike and sharply projects dorsally and perpendicularly to the rami (fig. 3A). The ventral margin is curved in lateral view (fig. 3A).

Comparisons: IGM 100/1268 has a well-developed, bladelike furcular apophysis, similar to those of charadriiforms, passeriforms, and galliforms (e.g., Fürbringer, 1888; Nesbitt et al., 2009; Smith, 2014). Among these groups, the new fossil is most similar to small extant charadriiforms (e.g., Calidris, Charadrius, Turnix), which also exhibit furculae with dorsally directed apophyses as well as dorsoventrally compressed and gently ventrally curved rami (Smith, 2014; fig. 3B, D). By contrast, the apophysis in galliforms is elongate and projects sternodorsally, well past the sternal margin of the rami, resulting in more V-shaped furculae (Fürbringer, 1888; Nesbitt et al., 2009). Galliforms also do not exhibit the curvature of the rami observed in IGM 100/1268 in lateral view. Passeriforms have furculae with rami that are variably arched or curved in lateral view and are highly dorsoventrally compressed, with no appreciable broadening toward the sternal extremity, unlike in IGM 100/1268. Based on these comparisons we assign IGM 100/1268 to Pan-Charadriiformes.

Definition: “Pan-Mirandornithes” refers to the clade including all birds more closely related to Phoenicopterus ruber and Podiceps cristatus than to shorebirds (Charadriiformes) based on the reference phylogeny in Prum et al. (2015).

Referred Taxa: In addition to crown-group Mirandornithes (flamingos and grebes), Pan-Mirandornithes also includes the proposed stem mirandornithine Juncitarsus (Olson and Feduccia, 1980; Mayr, 2014).

Referred Specimen: IGM 100/1418 (left quadrate lacking most of the orbital process).

Measurements: Lateromedial width of mandibular articulation, 6 mm; maximum dorso-ventral height, 9.4 mm.

Remarks: IGM 100/1418 shows slight abrasion of the otic process and body. It is similar in size to the quadrate of the extant grebe Podiceps grisegena (Red-Necked Grebe). We used digital renderings derived from high-resolution CT scan data to aid comparison with other taxa.

The otic and squamosal facets are both slightly abraded, and the intercapitular incisure is wide and shallow (fig. 4A, B, K, L). A distinct tympanic crest runs ventromedially from the lateral margin of the squamosal facet to the medial face of the body where it meets the medial crest (fig. 4A, B, F, G, L, K). The medial crest is also well defined and runs from the medial margin of the otic facet down to the medial condyle (fig. 4F, G). Between these crests lies a deep furrow, containing four small dorsoventrally aligned pneumatic foramina (fig. 3B). The basiorbital fossa, bounded by the medial crest and the base of the orbital process, contains three additional basiorbital pneumatic foramina also arranged in a line (fig. 4F, G). No caudomedial pneumatic foramina are present.

The pterygoid articulation is conformed as a facet on the medial base of the orbital process (fig. 4F, G). The specimen lacks a pterygoid condyle on the medial process, but it is unclear whether it is missing due to breakage (fig. 4F, G, K, L). The mandibular articulation consists of distinct lateral, caudal, and medial condyles. The medial condyle is convex mediorostrally with a small lip and is developed laterally as a concave articular facet (fig. 4K, L, P, Q). The lateral and caudal condyles are nearly confluent, separated by a shallow depression (fig. 4K, L, P, Q). The lateral process is moderately projected and bears a deeply concave quadratojugal cotyle with a complete, rounded margin (fig. 4A, B). The ventral margin of this cotyle is markedly expanded into a rostroventrally oriented lip in lateral view (fig. 4A, B, K, L).

Comparisons: We assign IMG 100/1418 to Pan-Mirandornithes based on the following suite of characters: a wide and shallow intercapitular incisure; a gracile quadrate body; a deep furrow bounded by the tympanic and medial crests; a concave articular facet lateral of the medial condyle; a rostrally projecting lip formed by the medial condyle; and a deeply concave quadratojugal socket. The combined presence of rostromedial and basiorbital pneumatic foramina is observed in both extant and stem flamingos (e.g., Palaelodus). These foramina are absent in most grebes (fig. 4I); however, a loss of pneumatic foramina is commonly observed in diving birds (Witmer, 1990; Mayr, 2015). This same combination of foramina placement has otherwise been reported only in some gulls (Laridae; Samejima and Otsuka, 1987; Mayr, 2015). However, in gulls, these foramina are relatively much larger than those observed in the new fossil and in stem and crown flamingos.

IMG 100/1418 exhibits additional similarities with the much younger (by ∼20 Ma) stem flamingo Palaelodus ambiguus (Milne-Edwards, 1863; Cheneval and Escuillie, 1992), known from the Oligo-Miocene of Europe, which are not observed in extant flamingos and grebes. These characters include a distinct furrow bounded by the medial crest and the base of the orbital process (fig. 4J), as well as a pterygoid articular facet located on the medial base of the orbital process (Mayr, 2015; fig. 4J). It differs from Palaelodus by exhibiting a concave caudal margin in lateral and medial views, whereas in Palaelodus this margin is straight (fig. 4E, J). Palaelodus also has a more ventrally projecting lip on the quadratojugal facet (fig. 4E) and a more rostrally projecting lip on the medial condyle (fig. 4J) than in the new fossil. The new fossil can be differentiated from Presbyornis spp., the most abundant avian taxon from this locality, by its trituberculate mandibular articulation (presence of a caudal condyle). In contrast, Presbyornithidae bears only two mandibular condyles (lateral and medial), consistent with other galloanserines (e.g., Cracraft and Clarke, 2001; Mayr and Clarke, 2003). The isolated nature of the new fossil and the lack of relevant comparative material available for other proposed pan-mirandornithines outside the crown group (e.g., Juncitarsus; Olson and Feduccia, 1980; Peters, 1987; Mayr, 2014) prevent a confident taxonomic assignment within Mirandornithes.

Referred Specimen: IGM 100/1371 (omal end of left coracoid).

Measurements: Maximum length as preserved, 13.6 mm; maximum mediolateral width of omal end, 4.7 mm.

Remarks: The dimensions of IGM 100/1371indicate that it is from a bird significantly smaller than the smallest extant galliform, Excalfactoria chinensis (King Quail). The shaft is narrow with negligible sternal expansion. It bears a deeply concave scapular cotyle with a rounded, slightly teardrop shape (fig. 4B). There is no foramen associated with the passage of the n. supracoracoideus. The acrocoracoid process is hooked with its lip forming a near-perpendicular (∼100°) angle to the shaft (fig. A, B). The clavicular articular face is flat. The well-projected glenoid facet terminates at the omal margin of the scapular cotyla (fig. 5B). The procoracoid process is reduced, appearing only as a slight ridge along the medial margin mediosternal to the scapular cotyla (fig. 5B).

Comparisons: The morphology of IGM 100/1371 is consistent with Galliformes in its narrow shaft, hooked acrocoracoid, greatly reduced procoracoid process, and lack of a supracondylar nerve foramen. A concave scapular cotyle like that of the new fossil is not observed among crown-group galliforms but is present in stem members of the clade and considered the ancestral condition in birds (Mayr, 2000; Clarke, 2002, Mayr and Weidig, 2004). Based on these traits we assign IGM 100/1371 to the clade Pangalliformes.

Among stem galliforms, the coracoid is most like the middle-late Eocene Quercymegapodiidae with respect to the subcircular scapular cotyle, parallel medial and lateral margins of the shaft, and a near-perpendicular angle formed by the hooked acrocoracoid and the shaft. By contrast, other stem galliforms (e.g., the early Eocene Gallinuloididae and the middle-late Eocene Paraortygidae) exhibit more mediolaterally broad scapular cotylae, sternally diverging lateral and medial margins of the midshaft, and more obtuse acrocoracoid-to-shaft angles (Mourer-Chauviré, 1992; Mayr, 2000, 2006; Mayr, and Weidig, 2004; fig. 5D). The new specimen can be differentiated from previously reported Quercymegapodiidae by its subplanar clavicular articular face, which in other quercymegapodiids is slightly concave (e.g., Quercymegapodius; fig. 5C) or convex (e.g., Ameripodius) (Mourer-Chauviré, 1992, Mourer-Chauviré, 2000). The acrocoracoid process is also more hooked in IGM 1371 than in these taxa or in Taubacrex (Alvarenga, 1988; Mourer-Chauviré, 1992, Mourer-Chauviré, 2000).

Referred Specimen: IGM 100/1360 (distal end of left humerus).

Measurements: Dorsoventral width of distal end, 5 mm.

Remarks: IGM 100/1360 is broken proximal to the dorsal supracondylar tubercle. It is from a bird slightly larger than Porzana carolina (Sora Rail). The dorsal supracondylar tubercle is a proximodistally elongate rectangular projection with a slight depression on its dorsal margin (fig. 6A, B). The fossa for the m. brachialis is elongate and ventrally positioned, with clearly demarcated ventral and distal edges (fig. 6A, B). Both the ventral and dorsal condyles are weakly cranially projected. The ventral condyle is roughly half the length of the dorsal condyle and extends distal to the dorsal condyle (fig. 6A, B). The flexor process is similar in distal extent to the dorsal condyle (fig. 6A, B, D, E). Both the m. scapulotricipitalis sulcus and the olecranon fossa are broad and shallow (fig. 6D, E).

Comparisons: IGM 100/1360 exhibits a combination of conditions consistent with a position within the clade Messelornithidae + Ralloideae (rails and allies) (Mayr, 2004), including: craniocaudal compression, a proximodistally elongate dorsal supracondylar tubercle bearing a concavity on its margin, a narrow and ventrally situated fossa for the m. brachialis, and a well-developed flexor process (fig. 6C, F). These characters can be observed in several fossil and extant taxa of this group; the Paleocene “messel rail” Wal-beckornis creber (Messelornithidae), the early Oligocene Bellgirallus (Rallidae), the early Miocene Palaeoaramides (Rallidae), and in the extant Aramides saracura (Rallidae) and Podica senegalensis (Heliornithidae) (Mayr, 2007; Mayr and Smith, 2002; Louchart et al., 2005). However, many of these characters are not apomorphic to the clade and are also seen outside this group. Therefore, while this new fossil may show affinities to a clade formed by messelornithids, rails, and finfoots (Mayr, 2004), we do not feel it can be assigned confidently at this time. For example, a weakly projected but elongate dorsal supracondylar tubercle with a concavity along the margin is found in the taxa listed above but is not found in all rails; in the Miocene Baselrallus intermedius and Porzana carolina the tubercle is entirely convex, and this feature shows variability in form throughout the clade (De Pietri and Mayr, 2014). An elongate, concave dorsal supracondylar tubercle is also present in the Sunbittern Eurypyga helias (Eurypygidae), and even in charadriiforms like Burhinus and members of the Alcidae (Mayr, 2007; Smith, 2014). Traits such as the development of the flexor process and the situation of the fossa for the m. brachialis are also variable.