Ratites to Waterfowl

The 45 Orders







The TiF Checklist

The checklist can be viewed in two ways. You can either view the annotated checklist on these web pages or download a list. The downloadable lists are Excel csv files that can be imported into spreadsheets such as Excel, or easily manipulated by programs such as perl. Four lists are available in csv format. The ABA and AOU lists above include only ABA or AOU species, but in TiF order, with TiF families. The S. American list has a slightly different species list than the AOU's SACC working list.

Further, Stephen Nawrocki has provided an excel version of the worldlist, version 2.79.

You can view an annotated version by clicking on the list of bird orders on the right, or by going to the family index, or by clicking on the family names in the various tree view pages. 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.

The 45 Orders

The TIF checklist currently groups the birds in 45 Orders and 241 families. Both a order-level and family-level trees are now availalbe in pdf format. Due to its length, the family tree is split into 5 parts.

Click for order-level tree Click for family-level tree
45 Orders 241 Families

The 45 orders of the TiF list include a few more orders than most other lists. However, I wanted to make sure that each of the orders was monophyletic, meaning that they include all descendants of a common ancestor. I did not have confidence that a shorter list would do. As it currently stands, I am highly confident that each order is monophyletic, with the possible exception of Ardeiformes (herons and ibises), where I am slightly less confident.

The orders are arranged in a way that is based on the results of Hackett et al. (2008). That arrangement is similar to that of Ericson et al. (2006a). So far, only the 4th edition Howard and Moore checklist (Dickinson and Remsen, 2013) has adopted a Hackett-like tree, but I expect others will follow it.

Not all of the Hackett tree is equally reliable. We can be pretty confident that 7 of the supra-ordinal groups are monophyletic. Moreover, we have a good idea of how the orders within each group relate, and their family structure. These groups are:









AEQUORNITHES (waterbirds)




PASSERAE (landbirds)


The deepest division in the avian tree is between the Paleognathes and everything else (the Neognathes). The next division is between the Galloanseres and the rest (Neoaves). It is not entirely clear how the other five groups relates. We also have 10 orders of uncertain affinities, the troublesome orders. These troublesome orders account for about 7% of all bird species.

The Hackett tree used here groups the first 5 troublesome orders with Mirandornithes and Strisores as Metaves, and the avian tree divides into Metaves and everything else (Coronaves). The other 5 troublesome orders group around Aequornithes. This group is then sister to Charadriae plus Passerae. The Metaves hypothesis was introduced by Fain and Houde (2004). Ericson et al. (2006a) refined it and produced results similar to the Hackett tree.

Although I am highly confidence about the 7 major groups, and the division into Paleognathes, Galloanseres, and Neoaves, that's as far as it goes. It has been clear for some time that the Metaves arrangement relies heavily on beta fibrinogen. In the recent paper by Kimball et al. (2013), the genetic signal from beta fibrinogen outweighs the signal (such as it is) from 49 other genes. Without it, the 7 major groups are still there, but you get a different picture for Neoaves.

Unfortunately, the other 49 genes give very little information about the affinities of the 10 troublesome orders. The support values are mostly under 50%, meaning it is virtually certain that it's wrong. There are serious doubts about the Hackett tree too. We can't really give a probability because it really comes down to the likelihood that beta fibrinogen gives a true signal, and there's not yet a good way to assess that.

Until we have more solid information concerning the troublesome orders and the correct arrangment of the major groups within Neoaves, the TiF list will continue to use a version of the Hackett et al. (2008) tree. It's mostly right, and three generations of analyses using beta fibrinogen have produced very similar results regardless of the set of additional genes and taxa investigated (Ericson et al., 2006; Hackett et al., 2008; Kimball et al., 2013).


Paleognath tree

The first major division among the living birds is between the Paleognaths and Neognaths. The taxonomy of the Paleognaths continues to be controversial. Traditionally, they have been divided into the flightless ratites and the volant tinamous. Some of the earlier generic evidence seemed to support this (e.g., Haddrath and Baker, 2001).

However, the recent studies of Chojnowski et al. (2008), Hackett et al. (2008), Harshman et al. (2008), and Phillips et al. (2009), Faircloth et al. (2012), Haddrath and Baker (2012), and J.V. Smith et al. (2013) have come to a very different conclusion. The ratites are not monophyletic. Some of the them are more closely related to the volant tinamous than to other ratites. In particular, the ostriches are no more closely related to the other ratites than to the tinamous, and the other ratites share a more recent common ancestor with the tinamous than with the ostriches.

Once we get beyond the ostriches, things are not so clear-cut. There are three major groups: (1) rheas, (2) tinamous (and probably the recently extinct moas), and (3) cassowaries, emus, and kiwis. The three groups seem to have separated over a short period of time. The next question is which of the three groups separated first?

Currently, this has not been conclusively resolved. However, it appears that it is either the rheas or the tinamou/moa clade (see J.V. Smith et al., 2013). Many analyses (e.g., Harshman et al., 2008; Phillips et al., 2009; J.V. Smith et al., 2013) favor putting the rheas first. However, when Haddrath and Baker (2012) considered retroposons, the balance shifted to the tinamaou/moa clade, and that is the topology we will follow here.

This topology implies that flightlessness has evolved at least 3 times in the Paleognaths: in ostriches, in moas, and at least once among the rheas, emus, cassowaries, and kiwis. Indeed, if the chronogram in Haddrath and Baker (2012) is correct, both the split of the rheas and Austalasian ratites and the split between the kiwis and emus/cassowaries occurred about 80 million years ago. This timing is consistent with the kiwis originating due to the separation of Australia and New Zealand and with the ancestral rheas walking their way across the then-temperate Antarctica to South America.

There is a possible complication besides the inherent uncertainties concerning ancient dates. There's some evidence that the extinct elephant birds of Madagascar are the closest relatives of the kiwis. This is a problem because Madagascar (and India) likely separated from Gondwana considerably earlier, perhaps 120 million years ago. In that case, the ancestor of the elephant birds would have had to fly, and we can infer that flight was indpendently lost in the rheas, emu/cassorwary, elephant bird, and kiwi lineages, a total of 6 losses of flight. This is not an unreasonable number. One only has to consider all of the flightless rails to see that.

The fossil evidence considered by Johnston (2011) suggests that the long-extinct Lithornithiformes were close relatives of the tinamous.

The Paleognaths are divided into several orders in recognition of the great antiquity of the various branches, some probably dating back to the Cretaceous period. Phillips et al. estimate the ostriches diverged from the rest of the Paleognaths about 60-95 million years ago, while Chojnowski et al. (2008) date this divergence around 64±22 million years ago, near the K/T (=K/Pg) boundary. Haddrath and Baker (2012) put it at 73-119 million years ago. The Cassowaries and Emus seem much more closely related than the other paleognath families, and so are placed in a single order: Casuariiformes. In fact, the molecular dates in Phillips et al. and in Haddrath and Baker suggest they are so closely related that they could be treated as a single family.


Struthionidae: Ostriches Vigors, 1825

The authors for family-group names are mostly based on Bock (1994), while the authors for order-group names are primarily based on the series by Brodkorb (1963-1978). In both cases, additional sources have been consulted (e.g., Livezey and Zusi (2007)). In a number of cases, I have managed to examine the original (thank you Google Books!). The ICZN does not regulate order-group names. However, for parvorders (-ida), infraorders (-ides), suborders (-i), orders (-iformes), and superorders (-imorphae), I am attempting to follow similar rules. In particular, they are based on priority and the orginal use must been at an ordinal level and must be based on an included genus name (type genus) actually used by that author. The endings have been adjusted to modern usage.

1 genus, 2 species HBW-1

RHEIFORMES Forbes, 1884

Rheidae: Rheas Bonaparte, 1849

1 genus, 2 species HBW-1


Dromaiidae: Emus Huxley, 1868

1 genus, 1 species HBW-1

Heupink et al. (2011) argue that the extinct King Island Emu, Dromaius ater, was quite closely related to the extinct Tasmanian subspecies of the Emu, and is best considered a dwarf subspecies of Dromaius novaehollandiae. The Kangaroo Island Emu, D. baudinianus, seems likely to have been no more different from the Emu than the King Island Emu was, so I am also considering it a subspecies of the Emu, D. novaehollandiae.

Casuariidae: Cassowaries Kaup, 1847

1 genus, 3 species HBW-1


The order of the Kiwis reflects the phylogeny in Burbidge et al. (2003). They also provide evidence for recognizing 3 species of brown kiwi, as is done here.

Apterygidae: Kiwis G.R. Gray, 1840

1 genus, 5 species HBW-1


Tinamidae tree

The Tinamiformes are the other branch of the Paleognaths. There is only one family. The taxonomy here is based on Bertelli and Porzecanski (2004) and SACC.

The Tinamidae are sometimes divided into two subfamilies: Rynchotinae and Tinaminae. This is consistent with the phylogeny shown, although there is some uncertainty about whether Nothocercus really groups with Tinamus and Crypturellus.

Tinamidae: Tinamous G.R. Gray, 1840

9 genera, 47 species HBW-1


The Neognaths divide into two parts: Galloanseres and Neoaves.



The Anseriformes are one branch of the Galloanseres, comprised of 3 families divided into 58 genera and 171 species. Linnaeus himself already recognized an order Anseres, but does not get priority because he did not base it on the genus Anser.

Anhimidae: Screamers Stejneger, 1885 (1831)

2 genera, 3 species HBW-1

Anseranatidae: Magpie-Goose P.L. Sclater, 1880

1 genus, 1 species Not HBW Family

Anatidae: Ducks, Geese, Swans Leach, 1820

59 genera, 170 species HBW-1

The large-scale organization of the ducks is based on Gonzalez et al. (2009b), with support from Bulgarella et al. (2010), Donne-Goussé et al. (2002), and Sorenson et al. (1999). The whistling-ducks are the basal group. There is uncertainty about whether Thalassornis belongs here or is basal to the rest of the ducks. The remaining ducks generally follow a cascade which is hard to break into convenient pieces. Not all of these pieces are included in these analyses, so information from other sources has to be added. These include Johnson and Sorenson (1999), McCracken et al. (1999), Worthy and Olson (2002), St. John et al. (2005), Pointer and Mundy (2008), and the discussion of Sraml et al. (1996) in Christidis and Boles (2008).

Anatidae tree
Click for genus-level tree
for Anseriformes

The resulting paste-up yields the tree to the right. There are some issues with this tree, but it was what I could do with the information I have.

It's generally thought that Stictonetta is the basal genus of the remaining ducks, but Sraml et al. (1996) found it in a clade with Cereopsis, which Donne-Goussé et al. (2002), St. John et al. (2005), Pointer and Mundy (2008), and Gonzalez et al. (2009b) group with Coscoroba. Here I separate them, putting Stictonetta in its own subfamily and grouping Cereopsis with Coscoroba in Anserinae.

Plectropterus is another duck of uncertain affinities. It may be quite basal, near Stictonetta, but there is less certainty about this, hence the blue color.

Gonzalez et al. (2009b) found that the remaining ducks fall into two clades: here designated Anserinae and Anatidae (they use a narrower Anserinae).

Anserinae contains four pieces. One is Oxyrunini. McCracken et al. (1999) made clear that the Musk Duck Biziura lobata is not one of the stiff-tailed ducks (Oxyrunini). In fact, it seems to be sister to Anserini + Oxyrunini (Gonzalez et al., 2009b). There's some evidence that the Nettapus Pygmy-Geese also belong in this subfamily (Sraml et al., 1996). Where the pygmy-geese go is not clear, so I have put them in a relatively basal position pending information on their true relatives.

I have also made some adjustment to Anserini. Although Gonzalez et al. (2009b) found that the Black-necked Swan groups with the Cygnus swans, Pointer and Mundy (2008) found it sister to the geese. I've compromised here by putting it between the two. The genus name Sthenelides (Stejneger 1884) is revived for this. It's possible that Chen should be merged into Anser. Donne-Goussé et al. (2002) and Gonzalez et al. (2009b) found that Chen is not monophyletic. However, they found different topologies for Anser + Chen. I use the Gonzalez et al. arrangement. This requires putting the Emperor Goose in a separate genus, the monotypic Philacte (Bannister, 1870). I have also arranged the species in Anser to conform with Gonzalez et al. (2009b). The arrangement of Branta follows Paxinos et al. (2002). Those curious about why the AOU allocated subspecies between Canada Goose and Cackling Goose the way they did should consult Paxinos et al. (2002) and Scribner et al. (2003). The other noteworthy point here is that Gonzalez et al. (2009b) placed Malacorhynchus basally in Anserini (which here includes the swans).

The other ducks belong to Anatinae, which contains almost three-fourths of the Anatidae. I split it into six tribes. How to order them is an issue. Donne-Goussé et al. (2002) present a couple of alternatives. A third in found in Sorenson et al. (1999), while Gonzalez et al. (2009b) have another and Bulgarella et al. (2010) yet another. Most pair Anatini and Aythyini. I've followed that, but left the relative position of Tadornini, Mergini, Cairinini, and Callonettini unresolved. The different analyses handle this differently, and I don't have sufficient reason to choose one over the other.

Within Tadornini, Gonzalez et al. (2009b) found that the Radjah Shelduck does not group together with the other Tadorna shelducks. This is handled by returning it to the monotypic genus Radjah (Reichenbach, 1852). Gonzalez et al. (2009b) have Cairinini sister to Tadornini. However, other studies have a different arrangement (e.g., Bulgarella et al. (2010) place Cairinini sister to Anatini + Aythyini).

One of the nice features of Gonzalez et al. (2009b) is that they have sufficient taxon sampling to reasonably resolve Aythyini. Note that Cairina and Asarcornis, formerly considered congeneric, end up in different tribes, Cairinini and Aythyini.

About one-third of all the ducks and geese are in the tribe Anatini, and many are conventionally placed in the genus Anas. The data presented by Johnson and Sorenson (1998, 1999) suggest that Anas should be split as some species of Anas end up closer to the Lophonetta-Tachyeres clade. Bulgarella et al. (2010) have a different arrangement with a monophyletic Anas, but only sample one of the relevant Anas species. The Lophonetta-Tachyeres clade is treated as in Bulgarella et al., but they note that alternative topologies cannot be ruled out.

Appropriate genus names exist for the Anas split, and I've used them for the ducks from Baikal Teal (now Sibirionetta formosa) to Northern Shoveler (Spatula clypeata). The point here is that the smallest monophyletic group that includes these species and the Mallard may be all of Anatini. Thus we either change the genus names or call all of them Anas. I prefer the former option.

Had they survived, the flightless Moa-nalos of Hawaii would be the basal group in the Anatini (Sorenson et al., 1999). Of the exant Anatini, the basal clade is comprised of two parts. One includes various South American ducks together with the Blue Duck of New Zealand and the New Guinean Salvador's Teal. The other includes Baikal Teal (Sibirionetta), Garganey (Querquedula) the silver teals (Punanetta), and the shovelers and blue-winged teals (Spatula).

The remaining ducks are left in Anas. This natural division results in a narrower form of Anas, but one that still has substantial structure within it. Within Anas, I follow the arrangement in Johnson and Sorenson (1999), not the similar but different Gonzalez et al. (2009b), which seems to mostly use the same data. Both agree that the wigeons are the basal group within Anas, strepera through sibilatrix. Some have argued these should also be placed outside Anas (the name Mareca would apply). I'm happy to think of them as a subgenus of Anas.

In the Johnson and Sorenson analysis, the other clades that successively break off start with the brown teals, chlorotis through nesiotis, which could be called Nesonetta. Then come the pintails, which sometimes go by Dafila. The gray teals follow. The name Virago could be applied to them. The green-winged teals are next, and could go by the name Nettion. The final split is between the African Black Duck (subgenus Melananas) and Anas in the narrow sense: the mallard complex.

Gonzalez et al. (2009b) have a slightly different take on the topology, although in most cases the differences are only weakly supported in both analyses. After the wigeons split off, the Mallard group is sister to the remaining Anas ducks. The brown and gray teals are a unified group (with a different internal arrangement that is the only well-supported conflict between the analyses) that is sister to the green-winged teals and the whole lot is sister to the pintails.

Before turning to the mallards, note that the Green-winged Teal, Anas carolinensis, and Eurasian Teal, Anas crecca, are not sister species. Rather, both Johnson and Sorenson (1999) and Gonzalez et al. (2009b) found that the Green-winged Teal is sister to the Andean Teal, Anas andium.

The Mallard complex continues to be a particular problem, with potential hybridization issues equalling those of the large gulls (see McCracken et al., 2001; Kulikova et al., 2004, 2005). Based on McCracken et al. (2001), I've decided to include the Mexican Duck as a separate species. Unless one takes an expansive view of the Mallards that includes Mottled and Black Ducks as subspecies, it's not a Mallard. It's closer to the Mottled Ducks and Black Ducks than to Mallards, so it doesn't make sense to list them separately and treat the Mexican Duck as a Mallard subspecies. In fact, there's a question about whether the Florida Mottled Ducks should be split from the other Mottled Ducks.

McCracken et al. (2001) also propose a solution to the problem that the Mallard appears to be in two separate group of mallard-type ducks. They argue that an mallard-type ducks have twice colonized North America. In that view colonization by a monochromatic ancestral mallards resulted in three monochromatic species (Black Duck, Mexican Duck, and Mottled Duck). The dichromatic Mallard we all known developed in the other clade, and subsequently colonized North America. It was able to hybridize with the existing mallard-type ducks and its North American descendents still carry the DNA of both ancestors. This makes the Mallard appear in two places on the tree. It also serves as a reminder that the phylogenetic network need not always form a tree, but may sometimes be more complex.

This of course creates problems handling the mallard complex. Avise et al. (1990) discovered two haplotypes in North American Mallards. It appears that the type B haplotype arises from hybridization with black ducks (including Mottled, Mexican, and Hawaiian), while pure Mallards are type A. Kulikova et al. (2004) note that the Eastern Spot-billed Duck has a variant of the type B haplotype, while the Indian Spot-billed and Philippine Ducks are type A. This suggests that the Eastern Spot-billed Duck should be grouped with the black ducks and that the Indian Spot-billed and Philippine Ducks go next to the Mallard. See Rhymer (2001) and Kulikova et al. (2004, 2005) for more details on the mallard complex.

Dendrocygninae: Whistling-Ducks Reichenbach, 1849-50

Stictonettinae: Freckled Duck von Boetticher, 1950

Plectropterinae: Spur-winged Goose Eyton, 1838

Anserinae: Geese, Swans Vigors, 1825 (1815)

Biziurini: Musk Duck Mathews, 1946

Nettapodini: Pygmy-Geese Bonaparte, 1856

Oxyurini: Stiff-tailed Ducks Swainson, 1831

Anserini: Geese, Swans Vigors, 1825 (1815)

Anatinae: Ducks Leach, 1820

Tadornini: Shelducks and Sheldgeese Reichenbach, 1849-50

Mergini: Sea Ducks Rafinesque, 1815

Cairinini: Perching Ducks von Boetticher, 1936-38

Callonettini: Ringed Teal Verheyen, 1953

Aythyini: Diving Ducks Delacour and Mayr, 1945 (1831)

Anatini: Dabbling Ducks Leach, 1820

Next Page