Bavaricarbo brevirostris sp. nov.

The generic name is derived from the Neo-Latin term Bavaria and carbo (Lat.): coal, a common component of the scientific names of cormorants; it refers to the geographic provenance of the type species.

The new genus differs from:

Comparisons with the Late Oligocene Praecarbo strigoniensis, which was described by Kessler and Horváth (2023) on the basis of a distal tibiotarsus, are not possible owing to the lack of overlap in the known bones.

GPIT/AV/00224 (incomplete right humerus lacking proximal end and ventral portion of distal end; Figs 1H, 2O, 2Q).

Medium-sized cormorant the size of the extant Phalacrocorax capensis, with a proportionally shorter beak (Figs 1, 3). Differs from the species of the taxa Borvocarbo, Nectornis, Oligocorax, and Nambashag as detailed above. Distinguished from:

The specific epithet is derived from brevis (Lat.), short, and rostrum (Lat.), beak, and refers to the short beak of the taxon.

Hammerschmiede clay pit near Pforzen, Allgäu region, Bavaria, Germany (47.923°N, 10.588°E); early Late Miocene, earliest Tortonian, MN 8; regional stratigraphic level HAM 4.

SNSB-BSPG 2020 XCIV-8500 and 8499 (partial rostrum and associated rostral portion of left ramus mandibulae; Figs 1A, 3A, 3B, 3D, 3F, 3G); SNSB-BSPG 2020 XCIV-17519 (rostral portion of left ramus mandibulae); SNSB-BSPG 2020 XCIV-16498 (partial right coracoid); SNSB-BSPG 2020 XCIV-7605 (extremitas omalis of right coracoid); SNSB-BSPG 2020 XCIV-16184 (fragmentary extremitas omalis of left coracoid); SNSB-BSPG 2020 XCIV-16865 (extremitas sternalis of left coracoid; Figs 1E, 4H); SNSB-BSPG 2020 XCIV-1360 (extremitas omalis of left coracoid); GPIT/AV/00194 (extremitas omalis of left coracoid; Figs 2M, 4A–C); GPIT/AV/00105 (fragmentary right coracoid; Fig. 1C); GPIT/AV/00222 (right humerus lacking distal end; Figs 1G, 4M, 4N); SNSB-BSPG 2020 XCIV-8028 (distal end of left ulna of an immature individual; Fig. 1J); GPIT/AV/00124 (proximal end of right carpometacarpus; Fig. 1K); SNSB-BSPG 2020 XCIV-3535 (right tibiotarsus of an immature individual lacking proximal and distal ends; Fig. 1L); SNSB-BSPG 2020 XCIV-5114 (left tibiotarsus lacking proximal and distal ends); SNSB-BSPG 2020 XCIV-0652 (fragmentary distal end of left tarsometatarsus); SNSB-BSPG 2020 XCIV-5073 (fragmentary distal portion of left tarsometatarsus); SNSB-BSPG 2020 XCIV-0697 (partial distal end of left tarsometatarsus; Figs 1O, 5A–C); SNSB-BSPG 2020 XCIV-16174 (phalanx of hallux).

All of the above material is assigned to the new species based on a corresponding size to the holotype humerus; the morphology of the referred bones is consistent with the assignment of the fossils to a single species.

Hammerschmiede clay pit near Pforzen, Allgäu region, Bavaria, Germany (47.923°N, 10.588°E); early Late Miocene, earliest Tortonian, MN 8. Most specimens are from the regional stratigraphic level HAM 4 except for GPIT/AV/00124, which is from HAM 5.

Rostrum (SNSB-BSPG 2020 XCIV-8500), length as preserved, 50.1. Humerus (GPIT/AV/00224), length as preserved, 101.0; estimated total length ~115. Tibiotarsus (SNSB-BSPG 2020 XCIV-3535), length as preserved, 80.4; estimated total length, ~95–110. Tarsometatarsus, length as preserved, 32.1; estimated total length, ~50–55.

The partial rostrum SNSB-BSPG 2020 XCIV-8500 (Fig. 3A, D, F) lacks the tip and is longer than that of the extant Microcarbo species, which are the smallest living cormorant. Judging from the caudal widening and dorsal upturning, not much of the caudal portion is missing, and the rostrum is proportionally shorter than that of Phalacrocorax capensis, which corresponds to the new species in the size of the postcranial bones (Fig. 1). A lateral rostral furrow, which characterizes extant Phalacrocoracidae and other seabird taxa is visible.

The mandibular ramus (SNSB-BSPG 2020 XCIV-8499) bears a ventral flange at its rostral end (Fig. 3B), which indicates that only a short portion of the symphysis is missing. The ramus is proportionally deeper dorsoventrally than in extant Phalacrocoracidae (Fig. 3B, C).

Unlike in Anhinga pannonica, the facies articularis scapularis of the coracoid is flat. On the extremitas sternalis, the linea intermuscularis, which delimits the insertion site of musculus supracoracoideus from that of m. coracobrachialis caudalis, is situated father laterally than in crown group Phalacrocoracidae (Fig. 4H–L); this plesiomorphic feature was first noted by Lambrecht (1933: 296) to characterize the taxon Oligocorax. Compared to Oligocorax littoralis, the extremitas omalis of the coracoid of Bavaricarbo brevirostris is dorsoventrally narrower and more pointed (Fig. 2M, N).

The humerus is represented by two specimens in complementary states of preservation, which lack the proximal and distal ends, respectively (Fig. 1G, H). It differs in some subtle features from the corresponding bone of extant Phalacrocoracidae. Most notably, the crista bicipitalis is less extensive proximodistally than in extant Phalacrocoracidae and the fossa musculi brachialis is deeper and proximodistally narrower. The crus dorsale fossae is broken, but it can still be discerned that, unlike in the Anhingidae, it was well developed and overhung the fossa pneumotricipitalis. The cranial surface of the crista bicipitalis is weakly convex. On the distal end of the bone, the proximal end of the condylus dorsalis forms a ventrally directed lobe as it does in extant Phalacrocoracidae. The tuberculum supracondylare ventrale is mediolaterally narrower than in Oligocorax littoralis (Fig. 2O, P).

Only the distal portion of the ulna has been found, and, judging from the rough surface texture, the specimen is from an immature individual.

The carpometacarpus lacks the distal portion and is poorly preserved; the processus extensorius and the processus pisiformis are broken. The fovea carpalis cranialis is less pronounced than in P. africanus.

All tibiotarsi lack the distal and proximal ends, which prevents detailed comparisons. The preserved portions of the bone show, however, that it was proportionally longer than the tibiotarsus of Oligocorax littoralis (compare Fig. 1 and Fig. 2).

The tarsometatarsus agrees with the corresponding bone of Microcarbo spp. The sulcus for musculus abductor digiti IV is more extensive than in extant Phalacrocoracidae other than Microcarbo, and the lateral margin of the bone forms a sharp ridge in lateral view. Also, as in the species of Microcarbo and unlike in other Phalacrocoracidae, the trochlea metatarsi II projects distally beyond the trochlea metatarsi III.

Overviews of Cenozoic Phala­crocoracidae were given by Mlíkovský (2002), Worthy (2011), Göhlich and Mourer-Chauvire (2010), and Mayr (2015a). Since these studies, Phalacrocorax ibericus Villalta, 1963 from the Late Miocene (MN 9) of Spain was synonymized with the galliform species Palaeortyx gallica by Sánchez-Marco (2021), and Phalacrocorax intermedius Milne-Edwards, 1867 was considered to likely be a representative of the Anhingidae (Mayr et al. 2020a). A very small cormorant from the Late Oligocene or Early Miocene of Germany, which was initially described as ?Limicorallus carbunculus by Mayr (2009), was tentatively assigned to Borvocarbo guilloti by Mayr (2010); the holotype of this latter species stems from the Late Oligocene of France (Mourer-Chauviré et al. 2004).

Bavaricarbo brevirostris is readily distinguished from the species of the Late Oligocene and Early Miocene taxa Borvocarbo and Nectornis in the flat facies articularis scapularis of the coracoid (which is plesiomorphically concave in Borvocarbo) and the shorter processus procoracoideus (which is distinctly longer in Nectornis), respectively. In size and morphology, the new species resembles Oligocorax littoralis from the Early Miocene (MN 1–3) of Europe. A specimen of O. littoralis in the collection of SMF, which was first mentioned by Lambrecht (1933: 290), consists of wing and leg bones of a single individual that were found around 1885 during construction work for the Frankfurt Westhafen (western harbor). This specimen, SMF Av 133 (Fig. 2A–L), is one of the most substantial records of an Early Miocene (Aquitanian) cormorant and is here for the first time figured. While similar to B. brevirostris in size and overall morphology, it differs in several features listed in the differential diagnosis.

SNSB-BSPG 2020 XCIV-10084 (processus mandibularis of left quadratum; Fig. 3H, I); GPIT/AV/00275 (fragmentary omal portion of right coracoid; Fig. 4G); GPIT/AV/00256 (fragmentary distal end of left tarsometatarsus; Fig. 5O, P); GPIT/AV/00164 (fragmentary distal end of right tarsometatarsus; Fig. 5Q, R).

Hammerschmiede clay pit near Pforzen, Allgäu region, Bavaria, Germany (47.923°N, 10.588°E); early Late Miocene, earliest Tortonian, MN 8. Most specimens are from the regional stratigraphic level HAM 4 except for GPIT/AV/00164, which is from HAM 5.

Tarsometatarsus, GPIT/AV/00164, length as preserved, 15.3; estimated total length, ⁓33–35.

These bones are from a very small cormorant, the size of the extant Microcarbo pygmaeus. An equally small cormorant from the Late Miocene (?MN 9) of Moldova was described as Phalacrocorax lautus by Kurochkin and Ganea (1972). Kessler (1984) referred a coracoid and a partial carpometacarpus from Moldova to this species, but the latter bone was subsequently identified as belonging to a species of the Anatidae (see Mlíkovský 2002). Because of the fragmentary preservation of the coracoid from the Hammerschmiede clay pit, meaningful comparisons with the similar-sized specimens from Moldova (proximal femur and distal ulna) are not possible, and there may well have been other small-sized cormorants in the Late Miocene of Europe. Ph. lautus was assigned to Microcarbo by Zelenkov and Kurochkin (2015). The small size of the Hammerschmiede fossils also suggests an assignment to Microcarbo, even though a small size may be plesiomorphic for the Phalacrocoracidae.

SNSB-BSPG 2020 XCIV-13269 (tip of rostrum; Fig. 3J–L), SNSB-BSPG 2020 XCIV-8677 (fragmentary distal end of left ulna); SNSB-BSPG 2020 XCIV-17035 (partial left coracoid of an immature individual); SNSB-BSPG 2020 XCIV-3536 (partial left coracoid); SNSB-BSPG 2020 XCIV-8872 (partial left coracoid); SNSB-BSPG 2020 XCIV-8695 (partial distal portion of left humerus); GPIT/AV/00279 (partial right humerus lacking proximal and distal ends); SNSB-BSPG 2020 XCIV-18824 (partial right humerus lacking proximal and distal ends); SNSB-BSPG 2020 XCIV-7927 (fragmentary distal portion of left tarsometatarsus; Fig. 5G–I).

SNSB-BSPG 2020 XCIV-16163 (rostrum Fig. 3N).

Hammerschmiede clay pit near Pforzen, Allgäu region, Bavaria, Germany (47.923°N, 10.588°E); early Late Miocene, earliest Tortonian, MN 8. All new specimens are from the regional stratigraphic level HAM 4.

Tarsometatarsus (SNSB-BSPG 2020 XCIV-7927), length as preserved, 28.8; estimated total length, ~45–50.

The rostrum (SNSB-BSPG 2020 XCIV-13269) resembles that of extant Anhingidae (Fig. 3M). As in the latter, the ventral surface of the specimen is flat, whereas it is concave in, e.g., the Gaviiformes (see below). The tentatively referred specimen SNSB-BSPG 2020 XCIV-16163 is somewhat smaller, which may reflect a sexual dimorphism in size.

The tarsometatarsus of A. pannonica was previously unknown. SNSB-BSPG 2020 XCIV-7927 shows a close resemblance to the tarsometatarsus of the extant Anhinga anhinga (Fig. 5J–L), but is from a distinctly larger species. As in A. anhinga, the foramen vasculare distale is very small and the fossa metatarsi I bears an ovate articular facet. Also as in extant Anhingidae, the lateral margin of the bone forms a sharp ridge in lateral view, and the sulcus for musculus abductor digiti IV is marked and wide.

SNSB-BSPG 2020 XCIV-17516 (left humerus; Fig. 7A–C).

Hammerschmiede clay pit near Pforzen, Allgäu region, Bavaria, Germany (47.923°N, 10.588°E); early Late Miocene, earliest Tortonian, MN 8; regional stratigraphic level HAM 4.

Maximum length, 9.9; distal width, 3.9.

Apus gaillardi was described by Ennouchi (1930) from the late Middle Miocene (MN 7–8) of La Grive-Saint-Alban in France. The holotype and only known specimen is a humerus (Fig. 7D, E). The species closely resembles the Hammerschmiede apodiform, and because of the similar ages of the fossils, the latter is here tentatively referred to Apus gaillardi.

The specimen from the Hammerschmiede clay pit is also of the same size as the humerus of Apus wetmorei from the Miocene-Pliocene transition (MN 13–14) of the Gargano peninsula in Italy (Ballmann 1976). Morphologically, it is likewise similar to the latter species, which was not differentiated from A. gaillardi by Ballmann (1976). However, in the Hammerschmiede specimen the processus supracondylaris dorsalis is situated somewhat farther distally than in A. wetmorei, which suggests a distinctness on species level.

Apus gaillardi is the oldest record of the Apodini, but the exact affinities of the fossil species within the latter clade still need to be determined. Skeletons of Tachornis, Panyptila, and Aeronautes were not available to us, and the assignment of A. gaillardi and A. wetmorei to the taxon Apus is anything but well established. The crista bicipitalis of SNSB-BSPG 2020 XCIV-17516 does not form the ventral bulge found in Apus apus and A. pallidus, but the bone is damaged in this area, so it is possible that this bulge is broken in the fossil.

Apus gaillardi and A. wetmorei were very small species, the size of Apus caffer and A. affinis, both of which are distinctly smaller than the extant European Apus apus (see measurements in Pavia et al. 2025). An equally small but considerably younger swift, Apus baranensis, is known from the Late Pliocene of Hungary and Bulgaria (Boev 2000). Because early Cenozoic stem group Apodidae were even smaller (Mayr 2015b), a small size is likely to be plesiomorphic for the Apodidae.

Mioaetus pforzenensis sp. nov.

The generic name is derived from aetos (Gr.): eagle; the prefix “Mio-“ refers to the age of the taxon.

The new taxon is distinguished from all other Accipitridae by the plantar deflection of the trochlea metatarsi II and the mediolaterally wide trochlea metatarsi IV.

SNSB-BSPG 2020 XCIV-3657 (distal end of left tarsometatarsus; Fig. 6D–F).

As for genus. The new species was of similar size to a male individual of the extant Ornate Hawk-eagle, Spizaetus ornatus.

The specific epithet refers to Pforzen, the municipality to which the Hammerschmiede clay pit belongs.

Hammerschmiede clay pit near Pforzen, Allgäu region, Bavaria, Germany (47.923°N, 10.588°E); early Late Miocene, earliest Tortonian, MN 8, regional stratigraphic level HAM 4.

Length as preserved, 40.3; distal width, 13.4.

This bone exhibits a very distinctive morphology in that the trochlea metatarsi II is more plantarly deflected than in other Accipitridae (Fig. 6F, I), and has a more convex lateroplantar articular surface. The trochlea metatarsi III is damaged and its medial rim is broken, but the trochlea appears to have been mediolaterally unusually narrow. The trochlea metatarsi IV is proportionally larger and wider than in other accipitriforms. The shaft has a subtriangular cross section. The fossa metatarsi I is situated far proximally (much more so than in strigiforms).

We considered the possibility that the unusual morphology of SNSB-BSPG 2020 XCIV-3657 is due to taphonomic distortion of the specimen. However, although the avian bones from the Hammerschmiede clay pit are often crushed to various degrees, they do not show plastic deformation. The trochlea metatarsi IV of SNSB-BSPG 2020 XCIV-3657 is broken and has been somewhat displaced in the specimen, but this is not the case for the trochlea metatarsi II, whose plantar deflection is one of the most unusual features of the new species. Direct comparison of SNSB-BSPG 2020 XCIV-3657 to a “typical” accipitrid from the Hammerschmiede (Fig. 6G–K) serves to illustrate the unusual features of Mioaetus pforzenensis gen. et sp. nov.