One of the most unforeseen results of sequence-based analyses of avian interrelationships, and one that has not been proposed by earlier anatomists, is a sister group relationship between the Passeriformes (passerines) and the Psittaciformes (parrots). A clade including both taxa was first found in analyses of Ericson et al. (2006) and has since been recovered in all phylogenies based on nuclear gene sequences (Hackett et al. 2008; Jarvis et al. 2014; Prum et al. 2015; Kuhl et al. 2021; Stiller et al. 2024; Wu et al. 2024). A clade including the Psittaciformes and Passeriformes is also supported by retroposon data and was termed Psittacopasserae (Suh et al. 2011), which was emended to the grammatically correct spelling Psittacopasseres by Sangster et al. (2022). Sequence-based studies, furthermore, identify the Falconiformes (falcons) as the sister taxon of the Psittacopasseres. Together with the Cariamiformes (seriemas), the Falconiformes and Psittacopasseres form the clade Australaves (Ericson et al. 2006; Hackett et al. 2008; Jarvis et al. 2014; Prum et al. 2015; Kuhl et al. 2021; Stiller et al. 2024; Wu et al. 2024).

Because stem group representatives of the Passeriformes, the aptly named Zygodactylidae (Fig. 1C), have zygodactyl feet (i.e., the fourth toe was permanently reversed as it is in parrots), it was hypothesized that a zygodactyl foot may be a plesiomorphic trait of passeriform birds (Mayr 2009: 184). A developmental mechanism for a reversal of this foot morphology into the plesiomorphic anisodactyl configuration was subsequently identified by Botelho et al. (2014; see, however, Mayr 2017: 213f. for a potential caveat regarding this hypothesis). As detailed by Mayr (2015), the new view on the evolutionary history of passeriforms has implications for the affinities of various early Cenozoic birds with “parrot-like” feet and led to a reinterpretation of various parrot-like taxa.

The occurrence of true parrots in the early and middle Miocene of Europe is well documented (Milne-Edwards 1867-1871; Cheneval 2000; Mayr and Göhlich 2004; Mayr 2010; Pavia 2014), and fossil Psittaciformes were also reported from the early Miocene of Siberia (Zelenkov 2016). Psittaciform affinities were, furthermore, assumed for a multitude of recently described fossils from the Paleogene of Europe and North America, but the exact phylogenetic positions of these are afflicted with uncertainty.

Harrison (1982) identified fragments of a putative psittaciform from the early Eocene British London Clay and described it as Palaeopsittacus. However, this taxon has since shown to be misclassified (Mayr and Daniels 1998) and may be a stem group representative of the Nyctibiiformes (potoos; Mayr and Kitchener 2025). A decade later, more substantial remains of parrot-like birds were reported by Mourer-Chauviré (1992) from the late Eocene of the Quercy fissure fillings in France; these fossils were assigned to the taxon Quercypsitta.

Mayr and Daniels (1998) described putative stem group Psittaciformes from the latest early/earliest middle Eocene of Messel in Germany and the early Eocene British London Clay. The species from Messel was classified in the taxon Psittacopes, which is the type genus of the Psittacopedidae (Fig. 1B). The fossils from the London Clay, which were revisited by Mayr (2020) and Mayr and Kitchener (2023a), belong to the taxa Parapsittacopes and Psittacomimus; one species from the London Clay was, furthermore, tentatively assigned to Psittacopes. Based on the new hypotheses on the interrelationships of passerines, Psittacopes, Parapsittacopes, and Psittacomimus are now considered to be zygodactyl stem group representatives of the Passeriformes, together with the Morsoravidae and Zygodactylidae (Mayr 2015; Ksepka et al. 2019, 2025; Mayr and Kitchener 2023a, 2023b, 2023c).

The taxon Morsoravis (Fig. 1A) from the Danish Fur Formation – the type genus of the Morsoravidae – was initially described as a charadriiform bird (Bertelli et al. 2010), but Mayr (2009: 115) noted close affinities to the latest early/earliest middle Eocene Pumiliornis from the Messel site, and a phylogenetic analysis performed by Mayr (2015) supported a clade including Morsoravis, Pumiliornis from Messel, Psittacopes, and the Zygodactylidae (see also Mayr 2011a). The Morsoravidae now include the taxa Morsoravis, Sororavis (from the London Clay), Consoravis (from the North American Green River Formation), and Pumiliornis (Mayr and Kitchener 2023b; Ksepka et al. 2025). The Zygodactylidae are well represented in early Eocene sites of Europe and North America and existed in Europe until the middle Miocene (Mayr 2017, 2022). A clade including the Zygodactylidae and the Passeriformes, which was termed Parapasseres by Mayr (2015), is supported by all current analyses, even though monophyly of zygodactylids is not recovered by some of these (Mayr 2015, 2020, 2021; Ksepka et al. 2019, 2025; Mayr and Kitchener 2023a, 2023b, 2023c).

Most current analyses also support a clade including the Morsoravidae, Psittacopes, Psittacomimus, Parapsittacopes and the Parapasseres (Ksepka et al. 2019, 2025; Mayr 2021; Mayr and Kitchener 2023a, 2023b, 2023c), but the interrelationships of these taxa are controversially resolved. Whereas the analysis of Ksepka et al. (2019) found a clade formed by the Morsoravidae and Psittacopedidae, analyses by Mayr (2021) and Mayr and Kitchener (2023b, 2023c) recovered a clade including the Morsoravidae and Parapasseres to the exclusion of Psittacopes, Psittacomimus, and Parapsittacopes. An analysis by Ksepka et al. (2025) resulted in a clade formed by Psittacopes, Psittacomimus, Parapsittacopes, Eofringillirostrum (see below), and the Parapasseres, to the exclusion of the Morsoravidae.

Dyke and Cooper (2000) described Pulchrapollia, another putative psittaciform bird from the London Clay. This taxon was subsequently assigned to the Halcyornithidae (“Pseudasturidae”; Mayr 2002), which include fossils from Messel (Pseudasturides, Serudaptus; Fig. 1E), the Green River Formation (Cyrilavis), and other early and middle Eocene localities in Europe and North America (Mayr 1998, 2000a, 2026; Ksepka et al. 2011; Mayr 2022; Mayr and Kitchener 2024). At times considered to be true stem group representative of the Psittaciformes (Dyke and Cooper 2000; Mayr 2002), the affinities of halcyornithids are elusive. An earlier analysis with a restricted ingroup sampling supported a sister group relationship to crown group Psittaciformes (Ksepka et al. 2011), whereas more recent analyses did not recover close affinities of halcyornithids to psittaciforms. Some analyses found the Halcyornithidae to be outside crown group Psittacopasseres (Mayr 2015; Ksepka et al. 2019; Mayr and Kitchener 2023c), but others identified them as stem group Passeriformes (Mayr and Kitchener 2023a; Ksepka et al. 2025).

Another taxon for which psittacopasserine affinities were assumed are the Messelasturidae, which include Messelastur from Messel and Tynskya from the London Clay and the Green River Formation (Fig. 1F; Peters 1994; Mayr 2000b, 2005a, 2011b; Mayr and Kitchener 2023c). Messelastur was initially assigned to the Accipitridae (Peters 1994), whereas Tynskya was likened to the Strigiformes (owls; Mayr 2000b). Possible strigiform affinities were also proposed for Messelastur (Mayr 2005a), but messelasturids were subsequently considered to be most closely related to the Halcyornithidae and both taxa were tentatively identified as stem group representatives of the Psittaciformes (Mayr 2011b). However, and like that of halcyornithids, the placement of messelasturids is poorly resolved in current analyses, some of which do not even support psittacopasserine affinities (Mayr 2011b, 2020; Mayr and Kitchener 2023a, 2023b, 2023c).

Vastanavis is a bird with a parrot-like tarsometatarsus from the early Eocene of India (Mayr et al. 2010, 2013), whose phylogenetic affinities are not unambiguously resolved in current analyses. Vastanavis was likened to Quercypsitta by Mayr et al. (2010), and recent studies supported a clade including Vastanavis, Avolatavis from the London Clay and the Green River Formation, and Eurofluvioviridavis from Messel (Fig. 1D), with this clade – the Vastanavidae – being only distantly related to psittacopasserines (Mayr 2015, 2020; Mayr and Kitchener 2023a, 2023b, 2023c; the analysis of Ksepka et al. 2025 did not include representatives of the Messelasturidae and Vastanavidae). Avolatavis resulted as the sister taxon of Quercypsitta in the analysis of Ksepka and Clarke (2012), and the clade formed by both taxa was recovered as the sister taxon of crown group Psittaciformes. The higher-level affinities of Eurofluvioviridavis could not be resolved in the original description of the taxon (Mayr 2005b).

Apart from Vastanavis, the only other taxon from Eocene deposits outside North America and Europe for which psittaciform affinities were assumed is Namapsitta from the middle Eocene of Namibia (Mourer-Chauviré et al. 2015, 2017). This bird is currently considered to be a stem group representative of the Psittaciformes.

A further early Eocene zygodactyl taxon assigned to the Psittacopasseres is Eofringillirostrum from the Green River Formation and Messel (Ksepka et al. 2019), which resulted as the sister taxon of either Pumiliornis (Ksepka et al. 2019), a clade formed by the Zygodactylidae and Passeriformes (Ksepka et al. 2025), or a more inclusive clade of putative stem group Passeriformes (Mayr and Kitchener 2023a).

The above taxa show disparate morphologies and can be grouped into two morphotypes, which are here informally termed “parrot-like” and “near-passerine”. The “parrot-like” morphotype (Vastanavis, Avolatavis, Eurofluvioviridavis, as well as the Halcyornithidae and Messelasturidae) is characterized by stocky tarsometatarsi and abbreviated proximal phalanges of the fore toes, whereas most representatives of the “near-passerine” morphotype (Morsoravidae, Zygodactylidae) features elongate tarsometatarsi, a comparatively short humerus, and long and slender toes with unabbreviated phalanges.

All current analyses support a clade including the Passeriformes, Zygodactylidae, Psittacopedidae, and Morsoravidae (Mayr and Kitchener 2023a, 2023b, 2023c; Ksepka et al. 2025), but the interrelationships of most taxa of the “parrot-like” morphotype are elusive. The present study revisits the phylogenetic affinities of early Eocene zygodactyl birds and identifies potential apomorphies of some major clades.

Phylogenetic analyses were performed on the basis of the emended and revised character matrix of Mayr and Kitchener (2023c; see File S1 for character descriptions and and File S2 for the character matrix). Four taxa and eight characters were newly added. The early Oligocene taxon Eocuculus, which was likened to Pumiliornis by Mayr (2008), was excluded, because it is uncertain whether the referred specimen (Mayr 2008) – on which most scorings of the wing and pectoral girdle elements in previous analyses (e.g., Mayr and Kitchener 2023a, 2023b, 2023c) are based – belongs to the taxon.

The analyses were run with the heuristic search modus of PAUP*4.0a169 (Swofford 2002). The primary analysis is based on the unconstrained data set. In a subsequent run, the analysis was constrained to the tree topology of Kuhl et al. (2021) as a molecular backbone phylogeny. Bootstrap support values were calculated with 200 replicates and the stepwise addition search. The trees were rooted with the anseriform Anhimidae. Three characters were scored as ordered. Tree length (L), consistency index (CI), and retention index (RI) were calculated.

The figured fossils are deposited in H.N.B. Garhwal University, Department of Geology, Uttarakhand, India (GU/RSR/VAS); National Museums Scotland, Edinburgh, United Kingdom (NMS); Senckenberg Research Institute Frankfurt, Germany (SMF); Staatliches Museum für Naturkunde Karlsruhe, Germany (SMNK); Bayerische Staatssammlung für Paläontologie und Geologie, Munich, Germany (SNSB-BSPG); and Université Claude Bernard, Lyon, France (UCBL).

The unconstrained analysis resulted in 1530 most parsimonious trees (L = 352; CI = 0.32; RI = 0.68), the strict consensus tree of which is shown in Figure 2A. The constrained analysis yielded 2152 most parsimonious trees (L = 367; CI = 0.31; RI = 0.69), the strict and majority rule consensus trees of which are shown in Figures 2Band C. In all phylogenies, there was no bootstrap support (> 50%) for most nodes including fossil taxa, with one notable exception noted further below.

Both analyses supported a clade including the taxa Morsoravis, Sororavis, Consoravis, and Pumiliornis, that is, a monophyletic Morsoravidae sensu Mayr and Kitchener (2023b). All analyses, furthermore, yielded a clade formed by Parapsittacopes and Psittacomimus to the exclusion of Psittacopes, as well as a clade including Vastanavis, Avolatavis, and Eurofluvioviridavis. Notably, even the unconstrained analysis recovered a sister group relationship between the Psittaciformes and Passeriformes, although monophyly of the Psittacopasseres has not yet been established with unambiguous apomorphies that are not also found in other avian clades (see also Chen et al. 2025).

Both analyses found a clade formed by Psittacopes, Parapsittacopes, Psittacomimus, and the Morsoravidae, Zygodactylidae, and Passeriformes. They also congruently supported a clade including most Halcyornithidae and the Messelasturidae, even though in the unconstrained analysis, the halcyornithid-like taxa Scopsoides and Serudaptus were recovered in a polytomy together with “typical” halcyornithids, messelsturids, and vastanavids.

The unconstrained analysis supported a sister group relationship between the Morsoravidae and Parapasseres, whereas the majority rule consensus tree of the constrained analysis recovered a clade formed by the Parapasseres and a clade including Parapsittacopes and Psittacomimus. In concordance with an analysis by Ksepka et al. (2025), the unconstrained analysis found zygodactylids as recognized by Mayr and Kitchener (2023a) to be non-monophyletic and successive sister taxa of passerines; in the majority rule consensus tree of the constrained analysis, the affinities of zygodactylids are unresolved.

The position of Psittacopes differed in the analyses. In the unconstrained analysis it was found to be the sister taxon of the clade (Parapasseres + Morsoravidae), whereas it was recovered as the sister taxon of the Morsoravidae in the majority rule consensus tree of the constrained analysis.

The analyses likewise did not unambiguously resolve the higher-level affinities of the Vastanavidae and the clade (Halcyornithidae + Messelasturidae). The unconstrained analysis resulted in a clade comprising these three taxa, which was shown to be the sister taxon of a clade including the Psittacomimidae, Morsoravidae, and Parapasseres. The majority rule consensus tree of the constrained analysis, by contrast, supported a position of the Vastanavidae outside Psittacopasseres and even Eufalconimorphae (the clade formed by the Psittacopasseres and Falconiformes). Both analyses did not support monophyly of the Halcyornithidae as currently recognized, because Serudaptus was recovered outside a clade formed by other halcyornithids and messelasturids in the unconstrained analysis, whereas the majority rule consensus tree of the constrained analysis found the Messelasturidae to be within the Halcyornithidae.

The taxon Eofringillirostrum was recovered as a stem group psittaciform in the unconstrained analysis, whereas it resulted in an unresolved polytomy in the constrained analysis.

Because all analyses supported a clade formed by Parapsittacopes and Psittacomimus to the exclusion of Psittacopes, a new family-level taxon is introduced (see also Ksepka et al. 2025 for the proposal to name this clade). It is diagnosed as follows: