There are big changes underway in bird taxonomy. You've probably noticed that the checklist is always changing. Suddenly the ducks are have replaced the loons in front of the NGS guide. Not too long ago the shrikes and vireos moved from next to the warblers to next to the jays. Maybe you've heard that vultures are storks, or that they're not really storks after all. Then there are the Baltimore and Bullock's Orioles, or are they just Northern, Slate-colored and Oregon Juncos, or maybe it's Dark-eyed today. Not only are species being split or lumped, but everything is being rearranged. A new view of bird taxonomy takes shape as we watch.
There have always been questions and disagreements about bird taxonomy, and it has always been changing. Since Aristotle or before, people have tried to find a natural way to arrange bird species in order (the very words “genus” and “species” carry the Aristotelean tradition). Early attempts were based on obvious similarities, such as grouping all waterbirds together. But what constitutes a natural order?
Linnaeus simplified scientific nomenclature by naming each type bird with a genus and species, but was unable to tell us what order was natural. It is not until Darwin that we get a clearcut way to order the birds. They should be ordered according to descent. Birds with more recent common ancestors should be grouped together.
Thomas Huxley (1867) was the first to construct a comprehensive bird list on Darwinian lines. He focused on the characteristics of the bird's skull to determine common inherited features that he could base his taxonomy on. Numerous other ornithologists have used morphological traits to classify the birds, with Livezey and Zusi (2007) being the most recent effort of this type.
More recently, focus has shifted to molecular methods, the most important of which is to directly study the stuff of inheritance—DNA. The culmination of the first wave of DNA based taxonomy are the publications by Sibley, Ahlquist, and Monroe (Sibley, Ahlquist, and Monroe, 1988; Sibley and Ahlquist, 1990; Sibley and Monroe, 1990). Sibley, Ahlquist, and Monroe used DNA hybridization to try to classify birds. This, together with increasing emphasis on cladistic methods, has revitalized interest in bird taxonomy. DNA hybridization was soon replaced by the much more precise DNA sequencing. This has opened the possibility of constructing a “tree of life” showing the evolutionary relationships between all living organisms. It has brought the prospect of producing a completely accurate taxonomic tree of all living bird species.
A Guide to the New Taxonomy
This set of web pages contains a guess at what the avian part of the tree of life might look like. It examines recent taxonomic changes, and possible changes to come. Some of the projected changes are pretty solid, others are guesses based on current research. Keep in mind that some of my guesses will turn out to be wrong, and some of the research they are based will be wrong or misleading. There will doubtless be taxonomic surprises before the entire avian tree is worked out, just as there have been a number of recent surprises.
The tree I've put together is an interpretation of the genetic relationships studied by many authors. The higher level taxonomy (the non-passerines) particularly draws on Ericson et al. (2006a), Fain and Houde (2004), Cracraft, Barker, and Cibois (2003), the AOU's South American Classification Committee, and the Sibley-Monroe checklist. Its purpose is twofold: To present the new ideas that Fain and Houde's Metaves hypothesis has brought to the higher taxonomy (mostly above the family level), and to examine the wholesale changes that been going on in passerine taxonomy, which at times seems like it's been run through a blender.
I will use the families and order listing from Gill's Ornithology text (2nd ed., 1995) for comparison purposes. Gill uses a modified Sibley-Ahlquist-Monroe ordering. I picked Gill's list (which I will call the modified SAM list) because it presents a view of taxonomy that was up-to-date in the mid-90's, one that judiciously incorporates portions of Sibley, Ahlquist, and Monroe's taxonomy. In spite of this, we still see considerable change. Had I picked an older taxonomy, such as that used in Clement's 5th edition, it would have be almost impossible to track the changes.
This set of web pages includes both tree diagrams and a linear order for bird families. The linear order is constructed by finding the oldest division in the tree, and then placing one group first and the other last. This procedure is repeated for every branch of the tree. Normally, I put the smallest group first, but sometimes the weight of tradition pushes me the other way. You shouldn't infer that families listed sooner are somehow older than families listed later. An equal amount of time has passed for both. Which is actually genetically closer to birds at the time of the split depends on details of their history, details which we do not know.
Linnaeus established a hierarchy of ranked groups for use in taxonomy. The number of ranked groups has expanded since his time. At the bottom is the species. Its name has two parts, a genus name and a specific name. The genus, which is capitalized, denotes a group of similar birds, the specific name indicates which one. Thus Dendroica palmarum denotes the Palm Warbler, one of 29 warblers in the genus Dendroica. We group similar genera into families, families in orders, and orders into classes. The class Aves includes all birds, and nothing but birds.
The names of the various ranked groups are usually constructed from the generic names with their level distinguished by a suffix. The ranks commonly used in bird classification are, from highest to lowest, superorder (-imorphae), order (-iformes), suborder (-i), infraorder (-ides), parvorder (-ida), superfamily (-oidea), family (-idae), subfamily (-inae), and tribe (-ini). Thus Parulini is a tribe containing the genus Parula while Parulidae (the wood warblers) is a family containing Parula. The standard ranks are insufficient to organize the complexity of the tree of life, and various un-ranked groups will also be used, sometimes with informal names. Informal names that are unique to this essay or used in a non-standard way are indicated by single quotes.
Paleognathes and Neognathes
The most basal (oldest) division in the tree of living birds has been identified. It is the division between the Paleognathes (Ostrich, Rheas, Cassowaries, Emus, Kiwis, Tinamous) and the rest of the living birds (Neognathes). The name reflects differences in the skulls of the two types of bird.
The split may have occurred during the early Cretaceous, more than 100 million years ago. The Paleognathes also appear in the modified SAM list. More recent evidence has strengthened the notion that the Paleognathes form a natural group, and that the Paleognathes themselves have an ancient division between the Ratites (Struthioniformes) and Tinamous (see Cracraft, 2001 for a summary). The recently extinct Elephant Birds of Madagascar (Aepyornithidae) and Moas of New Zealand (Dinornithidae) are included in the Ratites. The taxonomy here draws on Haddrath and Baker (2001), although they did not include the Aepyornithidae in their analysis.
Following the publication of Sibley and Ahlquist's book, ornithologists have gradually realized that the deepest division in the Neognathes separates the waterfowl and gallinaceous birds (Galloanserae) from everything else (Neoaves). These divisions are recognized in recent checklists such as AOU, ABA, and Howard-Moore 3rd edition. In contrast, the modified SAM list places the Anseriformes after the Pelecaniformes with the Phoenicopteriformes, Ciconiiformes, and Falconiformes separating them from the Galliformes.
There are two branches of Galloanserae: Galliformes and Anseriformes. We'll take the Anseriformes first. The screamers (Anhimidae) diverge first from the rest of the ducks, followed by the magpie-goose. I've included the Presbyornithidae in the tree so you can see how these numerous and well-known fossils fit in a modern taxonomy (Ericson, 1997). The oldest Presbyornis fossils date back about 60 million years.
The ducks (Anatidae) have recently been studied by Donne-Goussé et al. (2003). Their DNA study generally supports conventional wisdom. The whistling-ducks are the basal group. The remainder breaks into two major groups, one (Anserinae) containing the swans and some of the geese, the other (Anatinae) containing the various ducks and the rest of the geese.
The Galliformes have undergone some restructuring, most notably by Crowe et al. (2006). We follow their recommendations for the Galliformes in these web pages. The Galliforme tree takes the form of a cascade, with one group breaking off at a time until you get inside the Phasianidae. It starts with the megapodes, then the cracidae (guans, chachalacas, and currassows), guineafowl, and finally new world quail, all before we get to the Phasianidae.
Crowe et al. found two species that were quite wrongly placed: the Stone Partridge (Ptilopachus petrosus) and Nahan's Francolin (now Ptilopachus nahani), both of which ended up in the new world quail (Odontophoridae). When Crowe et al. (1992, with a different set of co-authors) had reorganized the francolins, they noted that Nahan's Francolin didn't appear to be a francolin. That reorganization is mostly supported by the new paper, but one other francolin proved problematic. The Crested Francolin, which they had already reassigned to the genus Peliperdix, is now Dendroperdix sephaena, although it remains in the francolin subfamily, Gallininae.
Crowe et al. (2006) rearranged the Phasianidae internally (see the tree). One item of note is pairing the Perdix partridges (specifically, the gray partridge) with the turkeys. The data does not support this as strongly as it does many of their taxonomic recommendations, so it may be subject to future revision.
Metaves and Coronaves
The higher taxonomy of the remaining bird families is uncertain. There seem to be separate groupings that are primarily waterbirds and primarily non-passerine land birds (see the discussion of “higher water birds” and “higher land birds” in Cracraft et al., 2004). It is also fairly well established that the shorebirds, gulls, terns, and alcids group together, as do nightjars, hummingbirds, and swifts. Grebes and flamingos also seem related, as do kagu and sunbittern. How these and other deeply rooted groups in the avian tree relate is more difficult to answer. The groups we are trying to understand are not always well-defined, and the presence of would-be members that are wrongly classified complicates DNA studies just as much as it does in traditional morphological studies.
In 2004, Fain and Houde published their Parallel Radiations paper, which proposed that many hard-to-classify families were part of a basal branch in Neoaves, Metaves. Besides the always troublesome Hoatzin, Metaves contains pigeons and doves, sandgrouse, kagu, sunbittern, grebes, flamingos, mesites, tropicbirds, nightjars and relatives, swifts, and hummingbirds. It adds up to over 900 species, almost 10% of all living bird species. Fain and Houde called the other 85% of living birds, Coronaves.
There had been hints before that some of the Metaves were related (e.g. grebes and flamingos), and that some bird families might be wrongly placed (e.g. tropicbirds), but existence of a whole group like Metaves came as a surprise. Ericson et al. (2006a) followed this up with a more detailed analysis that probably does a decent job of portraying the higher taxonomy of much of Neoaves.
It is not clear yet whether Fain and Houde's Metaves hypothesis is correct. In particular, studies based on large amounts of mitochondrial DNA do not support it (e.g., Gibb et al. 2007; Slack et al. 2007; Brown et al. 2008; Morgan-Richards et al. 2008). Another alternative is given by Ericson et al. (2006a) in the supplementary material. In that tree, the Hoaztin is basal in Neoaves. Then next branch includes the rest of the ‘Columbimorphae’. After that, the next division is between the Cypselomorphae and Coronaves.
The taxonomy presented here presumes that Metaves is a natural grouping, containing birds that are more closely related to each other than to the rest of Neoaves. It relies heavily on Fain-Houde and especially Ericson et al. (2006a) for the organization of the non-passerines. I have inserted several higher level groups to organize Ericson's tree.
Ericson et al. (2006a) also presented a timeline of avian evolution. Their paper suggests that at the very minimum, a dozen avian lineages survived the mass extinction at the end of the Cretaceous (the K/T extinction, about 65 million years ago). It's likely that the Paleognathe lineages had already separated, so the number is probably higher. Indeed, Brown et al. (2007) argue that a minimum of two dozen lineages of modern birds survived the K/T extinction.
In any event, by the end of the Eocene (about 34 million years ago), all of the modern bird orders (and many of the families) were distinct lineages. Similar results have been obtained in other studies such as Cooper and Penny (1997).
It's interesting to note that Brown et al's point estimate for the Passeriforme lineage is almost 90 million years ago. It was not long ago that ornithologists considered the passerines to be a recent lineage, and many reacted with disbelief when DNA data suggested they might be near the base of Neoaves. The tree presented here places the root of the Passeriforme branch in a basal position in Coronaves II.
The molecular estimates should be viewed somewhat skeptically. This applies even to calibrated estimates as in Ericson et al. (2006a) Fossil support for these genetic estimates of timing has been very weak. This is a concern since molecular clocks cannot always be trusted (e.g., Groth and Barrowclough, 1999). Although there are possible Cretaceous fossils of modern birds, most are pretty fragmentary and one can't be sure they belong to modern birds (Neornithes). There are plenty of fossils of other birds—Ichthyornithes, Hesperornithines, and Enantiornithes—just not Neornithes.
Should the lack of fossils be a big concern? No. The fossil record is incredibly sparse. How sparse? Caley (2007) gives a figure of around 2000 birds known only through fossils, compared with an estimate of 1.6 million bird species that have existed. That means we have only found fossils of 1 out of every 800 bird species.
How many Neornithes were there in the Cretaceous? They seem to have been a minor component of the avifauna. Suppose there were 100 at the end of the Cretaceous. Recall that we only need one or two dozen survivors. Suppose this had been a typical figure for the previous 25 million years and that the average species last for a million years. Then there were about Neornithe 2500 birds in the Cretaceous, so we should expect about 3 fossils. The numbers could be off by an order a magnitude, so not finding fossils is not too surprising.
It has also been suggested that location contributes to the problem. Since the vast majority of fossil collection has been in the northern hemisphere, one possible explanation for the paucity of supporting fossils is that most modern birds initially had a Gondwanan distribution (Cracraft, 2001; Ericson et al., 2002a). In that case, the lack of fossils merely reflects low collection effort in the right location.
The best fossil candidate does come from Gondwana, more specifically, Antartica. It is Vegavis (Clarke et al., 2005). Clarke argues that Vegavis is very closely related to the Anatidae. It may even be in the Anatidae. Its existence implies that the Paleognathes, Galliformes, Anhimidae, and Anseranatidae were already separate lineages before the K/T extinction, as was at least one lineage in Neoaves.
Part two continues with more details of the new taxonomic order for non-passerines, and includes a list of all the bird families.
The Birds in 34 Orders
- Struthioniformes (Ostriches, Rheas, Cassowaries, Emus, Kiwis)
- Tinamiformes (Tinamous)
- Anseriformes (Screamers, Magpie-Goose, Ducks, Geese, Swans)
- Galliformes (Megapodes, Chachalacas, Currassows, Guans, Guineafowl, Quail, Turkeys, Grouse, Pheasants, Partridges)
- Opisthicomiformes (Hoatzin)
- Columbiformes (Sandgrouse, Doves, Pigeons)
- Eurypygiformes (Kagu, Sunbittern)
- Phoenicopteriformes (Tropicbirds, Mesites, Flamingos, Grebes)
- Steatornithiformes (Oilbird)
- Podargiformes (Frogmouths)
- Caprimulgiformes (Nighthawks, Nightjars, Potoos)
- Apodiformes (Owlet-Nightjars, Swifts, Treeswifts, Hummingbirds)
Coronaves I (waterbirds and waders)
- Charadriiformes (Shorebirds, Larids, Alcids)
- Gaviiformes (Loons)
- Sphenisciformes (Penguins)
- Procellariiformes (Albatrosses, Petrels, Shearwaters)
- Phalacrocoraciformes (Frigatebirds, Sulids, Cormorants, Anhingas)
- Pelecaniformes (Hammerkop, Shoebill, Pelicans, Herons, Ibis)
- Ciconiiformes (Storks)
- Gruiformes (Turacos, Finfoots, Rails, Trumpeters, Limpkin, Cranes)
- Otidiformes (Bustards)
- Cuculiformes (Cuckoos)
Coronaves II (land birds)
(Seriemas, Caracaras, Falcons)
Cathartiformes (American Vultures)
(Secretary-Bird, Osprey, Hawks)
- Leptosomatiformes (Cuckoo-Roller)
- Strigiformes (Owls)
- Coliiformes (Mousebirds)
- Troganiformes (Trogans)
- Upupiformes (Hoopoes, Woodhoopoes, Hornbills)
- Coraciiformes (Roller, Ground-Rollers, Bee-eaters, Todies, Motmots, Kingfishers)
- Piciformes (Puff-birds, Jacamars, Barbets, Toucans, Honeyguides, Woodpeckers)
Psittaciformes (Cockatoos, Parrots)