The earliestfossils are known from the very end of theCretaceous with the oldest known members beingTeviornis andVegavis,[6][7] but molecular clocks suggest that neognaths originated sometime in the first half of theLate Cretaceous, about 90 million years ago.[8] Since then, they have undergone adaptive radiation, producing the diversity of form, function, and behavior that exists today. Neognathae includes theorderPasseriformes (perching birds), one of the largest orders of land vertebrates, containing some 60% of living birds. Passeriformes is twice as species-rich asRodentia and about five times as species-rich asChiroptera (bats), which are the two largest orders ofmammals. Neognathae also contains some very small orders, often birds of unclear relationships like thehoatzin.
The neognaths have fusedmetacarpals, an elongate third finger, and 13 or fewervertebrae. They differ from the Palaeognathae in features like the structure of their jawbones.Neognathae means "new jaws", but it seems that the supposedly "more ancient" paleognath jaws are among the fewapomorphic (more derived) features of the palaeognaths, meaning that the respective jaw structure of these groups is not informative in terms of comparative evolution. However, a neognath-like palate is also present inornithuran birds likeIchthyornis.[9]
Neognathae was longranked as asuperorder subdivided into orders. Attempts to organise this group further, as in theConspectus ofCharles Lucien Bonaparte, were never accepted by a significant majority ofornithologists. Until the 1980s, there was little subdivision of the Aves in general, and even less ofphylogenetic merit. Since then, the availability of massive amounts of new data fromfossils (especiallyEnantiornithes and otherMesozoic birds) and molecular (DNA andprotein) sequences allowed scientists to refine the classification. With new groups of neognath orders being verified, the taxonomic rank of the group needed to shift. Most researchers have now employed the unrankedtaxa ofphylogenetic nomenclature.[10]
Neognathae is now universally accepted to subdivide into two lineages, the "fowl" cladeGalloanseres and theNeoaves (sometimes called "higher neognaths"). The formalPhyloCode definition given to Neognathae by George Sangster and colleagues in 2022 is "the least inclusive crown clade containingGallus gallus andPasser domesticus".[11]
The subdivisions of the latter are still not well resolved, but severalmonophyletic lineages have been proposed, such as theMirandornithes,Cypselomorphae,Metaves, andCoronaves. Although groups such as the former two (uniting a few closely related orders) are robustly supported, this cannot be said for the groups Metaves and Coronaves for which there is no material evidence at present, while theMesozoic record of Neognathae is at present utterly devoid of birds that should have been present if these proposed clades were real.[12]
The orders are arranged in a sequence that attempts to follow the modern view on neognathphylogeny. It differs from the widely usedClements taxonomy as well as from theSibley-Ahlquist taxonomy, combining those elements from each that more modern research agrees with while updating those that are refuted. Most of the changes affect those "higher landbirds" that are sometimes united asnear passerines.[13]
^Mayr, Gerald; De Pietri, Vanesa L.; Love, Leigh; Mannering, Al; Scofield, Richard Paul (2019). "Oldest, smallest and phylogenetically most basal pelagornithid, from the early Paleocene of New Zealand, sheds light on the evolutionary history of the largest flying birds".Papers in Palaeontology.7 (1):217–233.doi:10.1002/spp2.1284.
^Torres, Christopher R.; Clarke, Julia A.; Groenke, Joseph R.; Lamanna, Matthew C.; MacPhee, Ross D. E.; Musser, Grace M.; Roberts, Eric M.; O’Connor, Patrick M. (2025). "Cretaceous Antarctic bird skull elucidates early avian ecological diversity".Nature.638 (8049):146–151.doi:10.1038/s41586-024-08390-0.ISSN1476-4687.
^For a draft phylogeny of Neoaves that is based on a review of massive amounts of published sources, and probably rather close to "the real thing", see Mindellet al. (2005)