The TiF Checklist
You can view the annotated version of the TiF list by clicking on the list of bird orders in the right column, or by going to the family index, or by clicking on the family names in the various tree view pages. If you're interested in a particular genus, the genus index may be helpful.
In the annotated list, recently extinct species and species whose taxonomic placement is particularly uncertain are color-coded. In some cases, superfamilies, subfamilies, tribes, and other groups have been added to help show how the birds are related.
Modern Birds in 47 Orders
The TiF checklist currently (mid-2024) groups the extant bird species into 47 Orders and 251 families. The pdf of the ordinal-level tree has been updated based on Stiller et al. (2024). An older family-level tree is also available in pdf format. Due to its length, the family tree is split into 5 parts.
 |
|
| 47 Orders | 251 Families |
|---|
What is an Avian Order?
When do I recognize a clade, a monophyletic group, as an order rather than a superfamily, family, genus, etc.? There are two main criteria. It has to be an old group. In the case of bird lineages, which mostly have originated since the Cretaceous, that means preferably originating no later than end of the Paleocene. That is, at least 56 million years old. Secondly, the group has to be sufficiently distinct from related orders. The later critereon argues against include screamers as distinct from the other waterfowl. Even if the screamer lineage were old enough, they just fit too comfortably with the other waterfowl.
With the elevation of Aegotheliformes (Owlet-Nightjars) to their own order, the TiF list contains 47 orders as of January 2024. This is a few more orders than most other lists. The Aegotheliformes were added because they are both ancient and distinctive. The other extra orders ensure that each of the orders is monophyletic, meaning that they include all descendants of a common ancestor. As it currently stands, I am highly confident that each TiF order is monophyletic.
The TiF orders are also old. The point estimates from Stiller et al. (2024) suggest that 36 of the orders originated at least 60 million years ago, and 45 (all but 2) originated no later than the Paleocene. Looked at from the other direction, every lineage that separated before 60 mya is either an order, or contains multiple orders.
The only orders that might be newer than the Paleocene are the Phoenicopteriformes (flamingos) and Podicipediformes (grebes). Their distinctiveness certainty supports ordinal status. Moreover, there is uncertainty about their dates and they too may have originated in the Paleocene.
A Hard Polytomy?
I froze the supraordinal portion of the TiF list in late 2016. It wasn't until May 2024 that I again updated it. All of the evidence pointed to a fairly hard polytomy in the Paleogene, a polytomy that more data couldn't satisfactorily resolve. Indeed, different papers had different solutions. The differences meant none could be convincing. Suh (2016) argued that these inconsistencies stemmed from a hard polytomy at the base of Neoaves. This polytomy could not be satisfactorily resolved with contemporary data and methods of analysis. I made my best guess at how it might work, and left at that for eight and a half years.
What Could Cause the Polytomy?
One plausible cause of the hard polytomy lies with the end-Cretaceous extinction. The scenario is that a single neoavian lineage survived (as did at least one lineage each of Galloanseres and Palaeognathae). The single lineage soon evolved into many lineages as it adapted to the new, post-apocalyptic ecology. The ancestors of the supra-ordinal groups diversified too quickly for it to be reflected in the genes (presumably due to incomplete lineage sorting) and may have also have continued to hybridize. This makes the genetics of the splits difficult or even impossible to sort out. It was certainly beyond what we could do in 2016.
When thinking about the polytomy, you should keep in mind that the ancestors of the ten groups would have orginally been subspecies, fully capable of interbreeding. Eventually they attained full species status, and over time became the orders we know now. There were probably other lineages that didn't survive to the present.
At present we don't know what these ancestral species were like, but they may have been some sort of shorebird, as in Feduccia's (1999) “transitional shorebirds” hypothesis. Near relatives to all modern birds include taxa such as Ichthyornis, Iaceornis, and the loonish foot-propelled diver Hesperornis. (see, e.g., Longrich et al., 2011). Interestingly, Longrich et al. found fossils of one such relative, which they designate “Ornithurine C”, on both sides of the K/Pg boundary (K/T boundary). Since no representatives of Neoaves were included in the analysis, we cannot say if this taxon was closer to Neoaves than to Galloanserae and Palaeognathae.
Further Thoughts on the Polytomy
A hard polytomy would explain why the relationships within Neoaves have been so problematic. The Vegavis fossil (Clarke et al., 2005) tells us that the Palaeognathae and Galloanserae lineages separated from the other birds in the Cretaceous. Neoaves seems to owe its origin to the consequences of the bolide that crashed into Chicxulub at the end of the Cretaceous (currently estimated at 66.04 million years ago, see Renne et al., 2013).
The resulting environmental devastation wiped out many species, genera, families, and even orders of living organisms. Even some of those that survived proved unable to thrive in the new world of the Paleogene.
The surviving birds rapidly diversified following the impact, creating problems such as incomplete lineage sorting at supra-ordinal levels within Neoaves. Untangling this has been a surprisingly protracted process. It is now thirty-five years after Sibley and Ahlquist, and the problem has still not been completely solved. Jarvis et al. (2014) threw a lot of data at the problem, three orders of magnitude more than Hackett et al. (2008) and still failed to crack it. In spite of all this, Suh (2016) informs us that we are still left with an intractable polytomy in Neoves. The situation remain unchanged until the 2020s.
Maybe the Problem isn't Really a Polytomy
Then Kuhl et al. (2021) appeared. By focusing on the 3'-UTR region, a noncoding region containing mRNA, messenger RNA that often affects gene expression. By focusing on the 3'-UTR regions, Kuhl et al. were able to get a pretty well-resolved tree for all of the TiF orders and many of the families. Here is the Kuhl ordinal tree:
Originally, I was unsure whether to go with that tree. I couldn't tell for sure if it really broke the polytomy, or if it was it just a another variation on the existing non-solutions..
I'd finally convinced myself Kuhl et al. was more likely real than illusory and started to work on rebuilding the TiF list when the “new Jarvis tree” appeared. This 52-author project (including Jarvis) has produced two papers. Stiller et al. (2024) presents the tree while Mirarab et al. (2024) explains the true nature of the problem, and, more importantly, how to solve it. The key is that they found a section of the bird genome (corresponding to the 4th chicken chromosone) that remained virtually unchanged for a time. This introduced a misleading phylogentic signal. They developed tools to deal this problem, and it has broken the polytomy. Cool!
Mirarab et al. (2024) calculated that normal gene recombination would mean the longest unchanged sequences would be around 4000 base pairs. Imagine their surprise when they found such a segment with an incredible 21,000,000 base pairs! Apparently, the genes lined up in such a way that they blocked most recombination. Mirarab et al. developed tools to find the troublsome regions and exclude them from the analysis. These methods were used to construct the Stiller et al. (2024) tree.
Stiller et al. (2024) and Kuhl et al. (2021) are in considerable agreement. This can be seen by comparing the TiF version of the
The ordinal-level disagreements are marked with red asterisks. Let's take a detailed look at them.
Its important to check the support values. The positions of Opisthocomimorphae, Gruimorphae, and Strisores involve softer support in both papers. The other taxa with soft support either don't affect the higher phylogeny, or are in Kuhl et al. One meaningful area of disagreement affects the position of the cuckoos and the owls. The rheas get high support from both papers, but are placed differently. This may be responsible for softening the support of the Cassowary-Kiwi clade in Stiller et al. I'm going with the Stiller et al. tree for TiF, maingly because of the huge amount of genetic data they use.
Neornithes: 12 Consensus Supra-Ordinal Groups
To understand the differences between Kuhl et al. (2021) and Stiller et al. (2024), we first focus on the similarities. I've highlighted 12 supraordinal taxa that are shared on the tree diagrams using green and magenta. The green highlights are applied to 4 taxa that have been recognized for some time. Except for Strisores, the 8 magenta taxa are recent, defined in the 21st century. Both Kuhl et al. (2021) and Stiller et al (2024) agree on their phylogenetic position. Finally, Kuhl et al. and Stiller et al. agree on the existence of the three blue (cyan) taxa, but disagree about their location on the tree.
If you examine my version of the Kuhl and Stiller trees, you'll also note some taxa are highlighted in either peach or yellow. The peach taxa either only make sense in one of the two phylogenies, or have a different composition (e.g., Otidimorphae and Columbimorphae) in the two phylogenies. The orders highlighted in yellow are in slightly different locations in the two trees.
The Main Supra-ordinal Groups
The modern birds, the Neornithes, are a clade containing all lineages that thrived following the end of the Cretaceous. The earliest divisions are the green taxa: Palaeognathae, Neognathe, Galloanserae, and Neoaves.
Palaeognathae and Neognathae
The division into Palaeognathae and Neognathe (all other living birds) was first proposed by Pycraft (1900), using Huxley's (1867) discovery of the palaeognathous (old jaw) palate. Although the ratites had long been considered related, Pycraft completed the group by including the Tinamous, creating the Palaeognathae. The key distinction for Pycraft was the palate, which differed between the two groups. He placed all other birds in the Neognathe. This has been confirmed many times in the genetic era, including both Kuhl et al. (2021) and Stiller et al. (2024). Kuhl et al. and Stiller et al. are in agreement about when this split occurred (doesn't mean they're right!). Kabout put it at 94±4 mya, Stiler et al. at 92.5±7.5 mya.
Palaeognathae and Neognathae may be poorly named. A recently described fossil suggests that the neognathous palate is much closer to the ancestral form and that the palaeognathous palate is the novelty. Oops! Unfortunately, it's too late to change the names.
Previously, there was no good fossil evidence on bird palates in the Cretaceous. That changed when Benito et al. (2022) described Janavis finalidens, a toothed bird fossil related to Ichthyornis that dated from about 67mya, almost the end of Cretaceous. It had a pterygoid bone that was similar to that of Galloanserae. Although Janavis is not in the direct ancestry of the Neornithes, it is a closely related lineage and their common ancestor predates the Neognathae/Palaeognathae split, which is all we need.
Galloanserae and Neoaves
The Galloanserae were recognized by P.L. Sclater in 1880. Sibley, Ahlquist, and Monroe (1988), using DNA hybridization techniques suggested that the Galloanserae and Palaeognathae were basal lineages among the living birds (Neornithes) and united them as Eoaves, with all other birds (except the then uncertainly placed Buttonquail) grouped in the novel taxon Neoaves. Uniting the Galloanserae and Palaeognathae turned out to be incorrect, but separating Galloanserae from Neoaves was spot on.
Sibley and Ahlquist (1990) corrected their error, and recognized that the Palaeognathae and Galloanserae were separate branches. However, they then expanded Neoaves to include the Galloanserae. This has not been followed by subsequent authors, and we use Neoaves to include all living birds but the Palaeognathae and Galloanserae. The Buttonquail are now known to be in Charadriiformes. Fossil evidence indicates that Galloanserae split occurred in the Cretaceous. Kuhl et al. estimate it happened about 82.2±4.6 mya.
Mirandornithes
Seven of the eight magenta taxa have been recognized by TiF for some time. Van Tuinen et al. (2001) seem to have been the first to recognize that flamingos and grebes were sister taxa. The pairing was later named Mirandornithes by Sangster (2005). Van Tuinen et al.'s reanalysis of Sibley and Ahlquist (1990) shows that other data to support this association were available earlier, but that the association had been missed by Sibley and Ahlquist, who did not include both of them in any of their individual trees. They did not suspect flamingos and grebes could be closely related.
Later analyses — Chubb (2004), Cracraft et al. (2004), Ericson et al. (2006a), Brown et al. (2008), Hackett (2008), Gibb et al. (2013), McCormack et al. (2013), Jarvis et al. (2014), Prum et al. (2015), Kuhl et al. (2021), and Stiller et al. (2024) — have also supported the Flamingo-Grebe clade, Mirandornithes. Further, Mayr (2004) identified morphological and oological evidence in favor of Mirandornithes. See also Mayr (2007, 2008). As with the genetic data, previous morphological investigations also did not consider the possibility of a relationship here. It was just too inconceivable.
Division into Columbaves and "Passerea"
The name Passerea is due to Jarvis et al. (2014). However, their delimitation is not a clade in either the Kuhl et al. (2021) or Stiller et al. (2024) phylogenies. Instead, I have separately reinterpreted the concept to make sense in the Kuhl and Stiller topologies, respectively. That's why the name is in quotes. I'm using the name in the same spirit, dividing the remainder of Neoaves into a part containing the pigeons and doves and a part containing the passerines. Although the included taxa differ, that division makes sense in both Kuhl et al. and Stiller et al. One version of "Passerea" works with Stiller et al., the other with Kuhl et al. I will only discuss the former.
In Stiller et al., Columbaves is comprised of 6 orders in two superorders: Otidimorphae and Columbimorphae. Otidimorphae consists of the turacos, bustards, and cuckoos and anis. Columbimorphae is composed of the Madagascan Mesites, the sandgrouse, and the pigeons and doves. I use the same two superordinal names with Kuhl et al., but with the cuckoos in Columbimorphae. For Stiller et al., Columbaves is sister to "Passerea", which is the remainder of Neoaves. Kuhl et al. is different, and Columbaves groups with Gruimorphae, Opisthocomimorphae, and Strisores, making a larger grouping, Columbae.
"Passerea": Elementaves and Telluraves
In Stiller et al., "Passerea" itself splits into two parts: Elementaves and Telluraves. Elementaves is the newest of these supra-ordinal taxa. It was introduced by Stiller et al. (2024). It includes the Hoatzin (Opisthocomimorphae), cranes, rails, and shorebirds (Gruimorphae), Kagu, Sunbittern, Tropicbirds (Eurypygimorphae), and Hackett et al.'s (2008) waterbird clade, called water-carnivores by Gibb et al. (2013). This last clade was named Aequornithes by Mayr (2011). Mayr (2010) also revived the name Strisores for the nightjar/swift/hummingbird clade (Cabanis 1847; Baird, 1858; G. Mayr 2010).
Telluraves: Afroaves and Australaves
Ericson (2012) defined both Afroaves and Australaves. Australaves includes the passerines as well as the closely related parrots and falcons. Afroaves consists of the higher landbirds together with the owls and hawks. Yuri et al. (2013) combined the two clades into Telluraves, which has no other members. Since Australaves and Telluraves both include the passerines, they are large groups.
Ages of Supra-Ordinal Taxa
Kuhl et al.'s calibrated dates suggest several of the 8 magenta taxa split off before the end-Cretaceous extinction. The much more numerous non-Neornithes were completely wiped out, so I consider this highly implausible. The confidence intervals allow for these taxa to have split soon after the dino-killer hit — just over 66mya. I expect this is actually what happened, a rapid diversification into the new ecosystems of the Paleogene. It's not clear if more than one lineage each from Galloanserae and Neoaves survived and prospered following the end of the Cretaceous. The extinct Lithornithiformes (false tinamous) are known from fossils from the middle Eocene, and may have also survived the extinction. They appear to have belonged to Palaeognathae.