On Avian Subspecies: Buyer Beware? - Part 1
by Nate Swick
Editor's Note: The ABA Blog is excited to offer a four part series by renowned field ornithologist and researcher Steve N.G. Howell on the history and relevance of avian subspecific identification. Steve is a senior international bird tour leader for WINGS and has written several books including A Guide to the Birds of Mexico and Northern Central America, Gulls of the Americas (with Jon Dunn), and the recently released and impressively received Petrels, Shearwaters, and Albatrosses of North America. He lives near Point Reyes, California.
These days, birders are more interested than ever before in subspecies, and thus taxonomy. Now that we think we can identify most birds we see to species, we want to go further down the path, to put names on the variation we see in the species around us. But how realistic is this notion? And are subspecies a useful means of describing variation?
I'd like to look at avian taxonomy from a philosophical and historical perspective, to help explain why subspecies are such a problem in modern ornithology, and by extension, in birding. The concepts are very simple, in theory. For better or worse, however, most of us live in a world of realities, not theories, and therein lies the paradox of taxonomy: if the world were simple enough for us to understand, we'd be too simple to understand it. Like the species concept, the subspecies concept might be viewed as a Procrustean bed in which Nature will not lie quietly - but as humans we have to try and make order out of chaos, one way or another.
Any discussion of a subject should be based on definitions, so how might we define taxonomy, species, and subspecies?
Taxonomy simply means classification, usually in reference to living things, such as avian taxonomy - the classification of birds. The units of taxonomy are taxa (singular: taxon) - for example, species and subspecies are different levels of taxa. Linked to taxonomy is nomenclature, the naming of things. It's basically human nature to put things in boxes - we are the taxonomic animal, and we need to classify things, to help make our world manageable.
But what is a species? Basically, a species is impossible to define to everyone's satisfaction and Charles Darwin's statement of 1859 holds true today: "No one definition has as yet satisfied all naturalists; yet every naturalist knows vaguely what he means when he speaks of a species." It's a bit like Justice Potter Stewart's famous "definition" of pornography: "I know it when I see it."
Birders should recognize that taxonomists themselves can't agree on how to define a species. Books and countless articles have been written on speciation and species concepts, with sides taken and lines drawn between proponents of different species concepts. But while there may be an answer, for now it is unknowable, or at least indefinable. Still, as humans we need a working definition, even if it's imperfect, and here I'll use the biological species concept (or BSC). The BSC remains the foundation for the taxonomy of the American Ornithologists' Union (AOU) Committee on Classification and Nomenclature (often abbreviated by birders simply to "the AOU" or "they" - as in: "Where the %$&# did they put longspurs?").
The classic definition of a biological species, as promoted by Ernst Mayr (1942) is basically: A population of actually or potentially interbreeding organisms, which are reproductively isolated from other such populations; the members of a species should be able to interbreed and produce fertile offspring. Within a biological species we can often distinguish variation. If this variation shows consistent patterns (beyond those associated with age and sex and season) it may be recognized by means of subspecies (often abbreviated as ssp.), which are designated formally by a trinomial, a third part of the scientific name.
A bird's scientific species name is two-part and italicized, based on Latin, and comprises the genus name (Capitalized) and the species name (lower case). A genus can be defined simply as a group of closely related species, or of species sharing a common ancestor; it's another subjective definition, but one that lies outside the scope of this essay. For example, the Song Sparrow is known scientifically as Melospiza melodia, and shares its genus, Melospiza, with Lincoln's Sparrow and Swamp Sparrow, because these three species are considered more closely related to one another than to other sparrows. Song Sparrow populations breeding in the south arm of San Francisco Bay, California, are considered distinguishable from other populations, and are given the subspecies name pusillula, as in Melospiza melodia pusillula. Usually the population of Song Sparrow described first is given the same subspecies name as the species name, as in Melospiza melodia melodia, which is known as the nominate subspecies. This nominate designation is purely administrative, it does not imply that the subspecies melodia is any more typical than any other subspecies, or more widespread - simply that is was named first.
A subspecies is synonymous with a race, despite the colloquial term "human race" - when really what people presumably mean is human species, of any race, creed, or color. Thus, for example, the race elegans means exactly the same as the subspecies elegans.
But as with a species, how do we define a subspecies? Ornithology is the science of the study of birds. For something to be considered a science, certain criteria should be met. These include repeatability, whereby something can be repeated and confirmed by other researchers, using universally accepted measures that are considered unambiguous. Thus, in describing a subspecies (or species) the ideal is that anyone else can recognize it and identify it by its characters.
Subspecies or species? Based on geographic location and season, in combination with bill size and shape, we can pretty safely say this Savannah Sparrow is of the taxon sanctorum, split in Handbook of the Birds of the World as a full species, San Benito Sparrow, but elsewhere considered a well-marked insular subspecies. Isla San Benito Oeste, Baja California, Mexico, 5 July 1999. © Steve N. G. Howell.
In 1949, Dean Amadon attempted to formalize "the 75% rule" as the standard for defining a subspecies. That is, to be recognized as a subspecies, at least 75% of a population should be distinguishable from >99% of any other population, on the basis of a defining character or set of characters. This is a purely arbitrary definition, but at least it's a definition - with numbers. Applying it is not always easy, however, and this formalized definition was proposed after the great majority of subspecies had been described. Moreover, most descriptions of avian subspecies in recent years have not used the 75% rule but instead have used the concept of detecting statistically significant mean differences between populations, not on determining the extent of overlap in characters. (Whether statistical significance is the same as biological or evolutionary significance is a minefield I’m not going to enter here!)
The AOU website (http://www.aou.org/committees/nacc/subspecies.php; accessed 4 April 2012) notes that "subspecies should represent geographically discrete breeding populations that are diagnosable from other populations on the basis of plumage and/or measurements, but are not yet reproductively isolated. Varying levels of diagnosability have been proposed for subspecies, typically ranging from at least 75% to 95%. Because subspecies represent relatively young points along an evolutionary time scale, genetic differentiation between subspecies may not necessarily parallel phenotypic divergence. ... Described subspecies that represent points along a phenotypic continuum (cline) probably would not warrant recognition given further study."
You'll note that this definition describes what the AOU would like subspecies to be, which is not how most subspecies were necessarily described. The AOU definition is also descriptive rather than prescriptive, and there remains no official template for describing subspecies. People do what they like, and it could be argued that avian taxonomy often falls short of science.
Several subspecies have been described for both Red Knot (front left) and Short-billed Dowitcher (front center and right) - but these subspecies are rarely separable other than in breeding plumage, and then sometimes only with the aid of complicated statistics. Yucatán, Mexico, 1 December 2010. © Steve N. G. Howell.
Subspecies and History
Humans like to collect things. Back in the day, especially throughout the 1800s, wealthy noblemen in Europe maintained collections of many things, bird specimens among them. They paid people to travel the world in search of new and novel additions to their collections. With specimens in hand, numerous new species of birds were described, usually based on the physical properties of structure and color - what has been termed the morphological species concept. The heyday of this naming lay in the 1830s to 1890s, and when something was named, its scientific name - and the person who described it - were enshrined forever; there was a certain prestige attached to the process.
As one might expect, humans being humans, there was rivalry, both personal and international, in getting names published first; Nigel Collar (1999) discusses some good examples. Single specimens with scarcely any description might be published as new taxa, and because of the rules of taxonomy, these names usually had to stay. New names might even be rushed out the next day in newspapers to cement "priority" of naming in print, rather than wait for the slower process of possible peer review and publication in respected journals or more formal outlets.
Virtually all subspecies diagnosis was based upon examination of specimens, that is, study skins made from dead birds. Other features, by which populations may be diagnosed, such as habitat, vocalizations, and breeding season, were rarely if ever used to define subspecies, yet such factors are often used in defining species. In essence, the subspecies developed as simply a morphological unit, one that could be identified in a museum tray by virtue of measurements and plumage. Perhaps paradoxically, some species appear indistinguishable by conventional measurements or plumage differences (think storm-petrels and owls, among others), and thus in a series of specimens they would not qualify as subspecies. Yet they can be distinguished as species by vocalizations, genetics, breeding seasons, habitat, and other criteria not apparent from a study skin!
Subspecies were most often described based on males, which in many species are more colorful than females (and thus more favored by collectors), and often more conspicuous and easier to collect, such as when singing in the breeding season. However, subspecies may also be described based on distinctive features of the female plumage or even of a seasonal plumage. For example, the three widely recognized subspecies of Short-billed Dowitcher are best separated in breeding plumage - in nonbreeding plumage they are often not distinguishable. Thus, a subspecies may not be identifiable in all ages or sexes, or at all seasons.
At least 11 subspecies of Common Yellowhthroat are recognized as occurring in North America, north of Mexico. Most are perhaps best distinguished in adult male plumage - so it is difficult (at best) to assign females and immature males, like this, to any subspecies when found outside the breeding season. Marin County, California, 31 August 2010. © Steve N. G. Howell.
Following the publication in 1859 of Darwin's worldview-changing opus, the concept of subspecies came to have two meanings: one was as an incipient species, the other as evidence of the adaptive response of a species to local climatic conditions (Mayr 1982). With the philosophical change from a morphological to a biological species concept from the 1880s to 1920s, subspecies came to be seen as measures of geographic variation. (The present AOU definition of a subspecies, quoted earlier, indicates that this is still the case.) Scientists looked anew at all of the old "species" to see if some were simply subspecies of a more widespread and variable species. Consequently, many "species" were lost in this period: in North America alone, some 315 avian taxa that had been described as species were "lumped" - subsumed as subspecies into species that exhibited geographic variation (Mayr 1982). The formal term for this is synonymizing: when one subspecies is considered indistinguishable from another, the subspecies that was described later is a synonym of the earlier described subspecies, and is lumped with it; the earlier name takes precedence.
As with many changes in fashion, the pendulum swung too far, and much of the indiscriminate lumping that characterized the 1900s to 1950s is now being slowly undone. For example, Murphy (1952) lumped seven former species of small shearwaters as subspecies of Manx Shearwater, which then became a species of worldwide distribution; no new data prompted this move, simply a change in philosophy. Nowadays, all seven of these taxa are split again as full species - and we have come full circle.
Concurrent with the loss of many "species" was the description of many new subspecies, as ornithologists sought further examples of geographical variation. The resultant fad of naming subspecies saw its peak in North America from the 1890s to 1950s. Ideally, to describe a subspecies one would compare a series of fresh-plumaged specimens, of the same age and sex, with another series of the same makeup, but often this was not done. Many subspecies were distinguished based on slight differences, and rarely with adequate samples of birds in comparable plumage. For example, in 1934 Ludlow Griscom described a subspecies of Calliope Hummingbird from the species' wintering grounds in southern Mexico - on the basis of a single molting male in worn plumage!
The 5th edition of the AOU checklist (published in 1957) was the last edition to list subspecies of North American birds, and their distributions; it doesn't tell you what the subspecies look like, however - for that you need to do a lot of digging in the literature and in specimen collections. Peter Pyle's two identification guides (1997, 2008) offer a starting point, but from his descriptions it is often difficult to infer what any given subspecies really looks like, or how easy it might be to distinguish in the field. For example, average color differences apparent in series of specimens are often expressed as "absolutes," as with Vermilion Flycatcher. For adult males, the Arizona subspecies flammeus is described as "head and breast ... red to orange-red, often with pale mottling" whereas the Texas subspecies mexicanus is "deep red, without orange or pale mottling" (Pyle 1997:242). This sounds like a fairly clear difference. Yet when Brian Sullivan, Michael O'Brien, Chris Wood, and I laid out about 20 male specimens of each subspecies at the American Museum of Natural History in New York, we found only a subtle average difference in coloration, and many specimens could not be placed confidently with one subspecies or the other based simply on plumage tones.
Implied but not stated explicitly in the 6th edition of the AOU checklist (published in 1983) is the notion that subspecies listed in the 5th edition had been evaluated critically (AOU 1983:xiii). The 1957 AOU checklist ostensibly marked the culmination of the subspecies era in North America. Subspecies description fell from fashion in the 1960s and 1970s, and the "glamour" in taxonomy is now with descriptions of cryptic species, and with revisions of genera or of families and other higher-level groups. Meanwhile, countless books and articles list subspecies as if they were inviolate and well-defined entities with mystical powers to elucidate geographic variation and distribution patterns. But are they?
A complete Literature Cited section will appear at the end of Part 4...