Note this is part 2 of a series. Part one is here
Bird watchers are interested in lots of aspects of the world of birds. One primary focus is identification. How do we tell one species from another? How do we identify a particular bird down a group of birds so we can find them in the field guide? Come to think of it, what determines the sequence that birds are shown in your copy of Peterson's or Sibley or whatever? Identification implies a classification system. And, whether you realize it or not, that classification is telling you something about evolution or at least what ornithologists think they know about evolution. The modern classification of birds, like that of all other organisms is rooted in evolutionary relationships. Species that are closely related, evolutionarily, are grouped together. These groups are then hierarchically grouped into larger groups containing species with a wider range of relationships.
So let's take a tour through a bit of the current work on understanding the evolutionary relationships of living birds.
I'll try and keep the technical details down to a minimum and get to the birds as soon as possible. However a bit of background is necessary. A good source of information on evolutionary relationships in general is the Tree of Life. The portion dealing with modern birds can be found here. You can follow the links to investigate specific groups of birds. The arrow at the base of the tree on each page will take you down a level to a larger and more inclusive level. Clicking on any of the linked names will move you up to a higher level in the tree.
The 'tree' in the tree of life is a phylogeny, an evolutionary family tree. A branching point represents a common ancestor of all the species at the end of the branches that are above that point. Species that share a more recent common ancestor (higher up in the tree) are said to be more closely related. The job of evolutionary biologists known as systematists is use both anatomical and molecular data to reconstruct phylogenies. This is a highly technical process involving very specialized statistical techniques.
In the previous diary I mentioned that one difficulty in understanding the evolutionary history of birds is the relatively poor fossil record owing to the fragile nature of bird bones. A second hurdle is that modern birds apparently underwent a period of very rapid evolution and diversification either in the late Cretaceous or the early Cenozoic (i.e. either just before or just after the last mass extinction). This is known as an adaptive radiation in which there is a lot of evolution to adapt to new environments/ways of life and lots of formation of new species. The adaptive radiation seems to have determined most of the major morphological types of birds (i.e. the dominant themes of bird evolution) and most evolution since then has been variation on those themes.
This poses a problem for people studying evolutionary history by building phylogenies using DNA/RNA sequence data. The reason is that the sequences being used in the analysis are probably not evolving particularly rapidly at this time. Only genes associated with the evolving anatomical changes would be evolving at a heightened rate. So you end up with something like this.
This is the phylogeny generated by the study we're going to discuss in just a minute (I promise). The color coding will be explained below. The crucial thing is to look at the length of the various branches. Both the red and purple clusters at the bottom are separated from the main tree by fairly long (horizontal) branches. The length of the branch corresponds to the number of estimated changes in the DNA sequences in the analysis (the length of the vertical branches in this tree are irrelevant - they are just spaced to make the branching pattern readable). The long branches at the base of each of those two groups indicate that there was quite a lot of change before the modern diversification within each group. So it is 'easy' to get strong data supporting the existence of those groups. If you look at the other color groups higher up you see that each one has a very short 'trunk' and then a lot of branching. This represents the very rapid evolution of all these different groups of birds. Because it happened quickly it is much more difficult to use the molecular data to sort out the order and pattern of splitting.
Those of you unfamiliar with the Linnean classification system might want to look here before continuing forward. The traditional classification of birds (at least all the birds occurring in North America, several orders such as Penguins have been omitted) can be seen here.
Many of the orders listed in the table are easily recognizable. Everyone knows what an owl is for example and can tell if any given bird is an owl or not. The order in which the orders are listed reflects an idea of their relationships with the most 'primitive' forms being listed first. The concept of primitive is not popular among modern evolutionary biologists because it is implying a directionality to evolution. Just over twenty years ago Sibley and Alquist published the first ever attempt to work out a bird phylogeny based on molecular characteristics (this Sibley is the father of the field guide author). It was based on the concept that officially named taxa (e.g. an Order) should be monophyletic - include all the descendants of a single common ancestor. It made a number of suggestions that were quite shocking at the time (e.g. New World Vultures were actually related to storks and herons rather than raptors) and has inspired a lot more work taking advantage of the increased ease of obtaining molecular data and more sophisticated statistical techniques. In the last decade a group of bird systematists teamed up to gather as much data as possible and analyze the hell out of it in order to get around the problem posed by the above mentioned rapid evolution. We'll get to what they found in just a minute. First a couple of comments.
About half of living birds are included in a single order the Passeriformes. It is common to talk about birds as passerines and non-passerines because this splits the birds in in approximately equal halves. It is a bit like dividing personal transportation into cars and not cars where not cars includes bicycles, motorcycles, a wide variety of boats and corporate jets. The Passeriformes are mostly birds that would fit the image most people would have when asked to visualize a generic bird. Examples of passerine birds include sparrows, thrushes, warblers, vireos, jays and crows, finches, chickadees, blackbirds, orioles, flycatchers, and so on.
Anyway I mostly brought up the passerines because the study I'm going to discuss in just a minute doesn't delve into their diversity very much, just including a few representatives. They argue that we are pretty sure based on a lot of previous work that passerines have a single evolutionary origin and that trying to delve into their evolutionary history in detail would have made the analysis too difficult. I'm not an expert on this but a perusal of several studies over the last decade would seem to back up their argument.
Three years ago the results of thismajor collaborative study of bird evolutionary relationships was published in Science with close to 20 authors (Hackett et al. 2008). A blog entry that goes over the results in detail can be found here and here is a second link to the paper (and many others) in case you can't access it directly. This publication represents a monumental amount of work on two fronts. First they used sequences from 19 different loci (genes) gathered from each species in the study. This means not just a lot of species but species spread out across the diversity of birds. Passeriformes were definitely under sampled given their diversity for reasons they gave that I described above. Second a lot of high-powered computers had to run for a long time, working out the best possible evolutionary tree from the DNA data. This is not a trivial matter - the number of possible evolutionary configurations in a phylogeny increases exponentially with the number of different organisms in the tree. With a relatively small number (a hundred or so) the number of possible outcomes exceeds the number of atoms estimated to exist in the entire universe.
So what did they find. Here's a summary tree.
The color coding in this tree is the same as the one above except that they switched black and gray for some reason. There are two important differences. The top tree has horizontal lines of different lengths representing different amounts of molecular evolution. So the branch ends don't line up on the right side of the figure. This tree is just showing the branching pattern so the branches are of arbitrary lengths to get a easy to read tree. Secondly the first tree gives the best branching pattern they found. Parts of it have more support than others. The parts with so little support that they are not statistically supported are collapsed to show uncertainty.
First their findings supported what a number of other recent studies have found - the division of the living birds into three distinct groups. These groups don't have common names and their official names are a bit daunting. So I'll give them nicknames and put the official names in parentheses.
Early Birds (Palaeognathae) - Purple. This is by far the smallest group and the only one with no representatives within North America (as defined by birders - i.e. Canada and the US excluding Hawaii). It includes the ratites which are the various large flightless birds of the southern Hemisphere both still living (Kiwi, Emus, Cassowaries, Rheas and the Ostrich (which ranges north of the equator)) and recently extinct (Moas and Elephant Birds). It also include the tinamous, a group of unassuming ground dwelling birds resembling quail or grouse that are found in Central and South America. Tinamous have striking calls and are usually quite secretive. Interestingly the tinamous show up in their analysis as being more closely related to the Rheas and the Australian/New Zealand ratites than any of those birds are related to the ostrich. This implies that either flight was lost multiple times in the ratites or the tinamous re-evolved flight! The latter seems very unlikely.
Great Tinamou and Emu
Ducks and Chickens (Galloanserae) - Orange. A striking finding of several studies of bird evolution has been that two orders of birds that had not traditionally been thought to be closely related are in fact both closely related and fairly distantly related to other birds, being the second living branch of avian evolution. The two orders are the Anseriformes (ducks, geese, swans, and a handful of other obscure birds) and the Galliformes (pheasants, grouse, quail, turkeys, chickens, and other similar birds). This finding is being reflected in the changing order in bird lists and should be reflected in new field guides.
Brush 'Turkey' and Ringneck Duck
New Birds (Neoaves) - This is everything else, or 95%+ of existing bird species. It is within this group that rapid evolutionary diversification took place. The massive study linked above was able to identify five major subgroups within this group. However the period in which they diverged from one another was relatively short and the evidence for the membership in each of these groups is not quite as strong as the evidence for each of the three major groups. All five groups contain common North American birds.
1. Land Birds - Green. This is by far the largest group as it includes the Passerine birds discussed above. It also includes a large number of other groups of mostly non-aquatic birds (I don't think land birds is a good term but that's what the authors used). Other examples of this group include falcons, hawks and eagles, vultures, owls, parrots, woodpeckers, toucans, hornbills, and kingfishers. One striking finding of this study is that their evidence points to a very close relationship between parrots, falcons, and the passerine birds. If this is true then falcons evolved their 'hawk-like' form and way of life independently from the other birds of prey.
Northern Mockingbird, Great Kookabura, Rainbow Lorikeet, Peregrine Falcon
2. Shorebirds and Relatives - Yellow. This group is closely related to the land birds and is equivalent to the traditional bird order Charadriiformes. It includes sandpipers, plovers, turnstones, oystercatchers, avocets and stilts, and a number of other sandpiper-like birds from around the world as well as the gulls and terns and their close relatives (skua, jaegers, auks, skimmers). This finding closely aligns with traditional taxonomy both in the inclusion of species within the group and also with its relationship to other birds.
Willet and Ring-billed Gull
3. Water Birds - Blue. This group contains almost all water associated birds not already mentioned (there are some exceptions which will be mentioned later). Within the water birds are two major groups: the penguins and tubenosed seabirds (albatrosses, petrels, shearwaters) which are a highly oceanic and predominantly southern Hemisphere group and a more heterogenous group of more familiar birds to North Americans (see next paragraph). Also in the water birds but more distantly related to both groups are the loons.
The second group I allude to contains both the long-legged wading birds (herons, storks, etc.) and the pelicans and birds long thought to be closely related to them (cormorants, anhingas, boobies, etc.). The surprising finding here is that pelicans are more closely related to herons than any of the birds that were their supposed relatives before and that herons and ibis are more closely related to not only pelicans but also cormorants etc than they are to storks.
Anhinga and Buller's Albatross
4. Superfliers (Caprimulgiformes and Apodiiformes). Brown. This group includes the swifts, hummingbirds, and the nightjars (nighthawks, whip poor wills, etc.). This is not a surprising finding as these groups have long been assumed to be related. What is a bit different about this result is that these birds fall outside of the land bird group and that the swifts and hummingbirds appear to have evolved from nightjars.
Tawny Frogmouth and Booted Rackettail
5. Freaks and geeks. Gray in first tree, Black in second. The last major group compromises most of the birds in two traditional orders: rails and their relatives and cuckoos and their relatives. Both of these were kind of quirky diverse groups to begin with so the fact that two such different groups of birds are combined makes a kind of perverse sense.
Limpkin and Weka. I don't seem to have a cuckoo picture.
There are a number of other birds that failed to fall into any of these groups and remain of uncertain position with the Neoaves. Here are some of the more interesting examples. They are in black in the first tree and gray in the second
a) Pigeons and Doves (Columbiformes). A widesprea and well known group of birds that was traditionally thought to be closely related to the parrots.
Squatter Pigeon
b) The Hoatzin. This bizarre South American bird has long baffled ornithologists. It most closely resembles members of the cuckoo group but has never been definitively placed in that order. It is the only living bird species with claws on its wings. This study only deepens the mystery of how this weird looking (and sounding) colonial bird that climbs using its wings fits in with the rest of the birds.
c) The Kagu and the Sunbittern. Two unusual birds found on opposite sides of the Pacific island. The Kagu is only found on New Caledonia while the Sunbittern is native to South America. Both have traditionally been thought to be related to the rails but this analysis finds them to be closely related to one another and to nothing else. Given that they are both sedentary tropical forest birds and not likely to disperse across the ocean it seems likely that they are the remnants of a more widespread group.
Kagu and Sunbittern
d) Grebes and Flamingos. These two types of birds wouldn't seem to have much in common other than living in association with water. Again they data strongly link them together.
Pied-billed Grebe and American Flamingoo
e) Tropicbirds. These beautiful birds of the tropical oceans had long been assumed to be closely related to boobies and pelicans. This study doesn't support that idea and places them off by themselves.
Phylogeny images from Hackett et al. 2008. Images of Tropicbird, Flamingo, Kagu, Sunbittern, Hoatzin, and Great Tinamou from Wikimedia Commons. All other images by the author.