Articles - Crossbills: the species puzzle
Common Crossbill by Steve Young
David Callahan - Posted on 04 Feb 2010
It is widely believed that Britain has only one endemic bird species, the Scottish Crossbill. Separated from Common Crossbill almost 20 years ago (Knox 1990), its unique status among Britain’s birds has led to it becoming something of a figurehead species, as well as a cause célèbre among conservationists. So when it recently emerged that BirdLife International might be considering ‘demoting’ Scottish Crossbill from full species rank, there was understandable consternation.
Scottish Crossbill is treated as a species in its own right due to a number of discrete differences in measurements, bill depth and morphology, call and ecological niche, all of which prevent it interbreeding with other crossbill forms. Such ‘assortative mating’ – the tendency for organisms to reproduce with others similar to themselves – results in ‘reproductive isolation’. This is necessary for a species to maintain its genetic integrity, the self-containment which makes it recognisable to other individuals of the same species, and indeed to us.
This female crossbill from Kazakhstan is of the subspecies tianshanica. Large, complex regions like the Central Asian Tien Shan Mountains may eventually be shown to hold more than one crossbill taxon. Photo by Askar Isabekov.
However, research over the last few decades has shown that these subtle essential differences are present in many other crossbill populations. In fact, judging by the criteria listed above, the genus Loxia may be composed of at least 35 species, 30 or more of which are currently ‘lumped’ within the most widespread taxon, Common Crossbill L curvirostra. Of these, only 20 or so forms are currently recognised as subspecies, and they are largely indistinguishable from each other in the field.
This situation poses immense identification and philosophical problems to anyone with an interest in birds. How can we tell them apart? Are they really what we understand as a species? And what does this say about how we assign species rank in the future?
The ‘wing-barred’ crossbills have already been split into three species, and this article deals with the plain-winged crossbill complex. Crossbills are a very distinctive genus of finch, and the species we call Common, Scottish and Parrot Crossbills appear to have diverged at some point between 10,000 and 27,000 years ago. DNA analyses show that Loxia split from a common ancestor with the redpolls, siskins and goldfinches about five million years ago, though the oldest fossil known is from north-west Bulgaria and dated to about 2.2 million years ago.
Common Crossbill is found throughout the Northern Hemisphere, from North America (where it is known as Red Crossbill) east to Japan, and it reaches south to Guatemala, Morocco, Vietnam and The Philippines. Its unique crossed mandibles are adapted to feed on the seed cones of conifers, and are specialised to extract the seeds from between the scales of the cones.
Most northern populations tend to be irruptive, forming nomadic flocks which roam over large areas, seeking cones on which to feed and moving on when the trees have been fully exploited. These exodus events are driven by the often cyclical failure of cone crops in the birds’ home territories and/or very productive breeding seasons. Ringing recoveries show that the main known source of irruptions into Europe seems to be Russia west of the Ural Mountains. Birds don’t return until the pines have recovered, and so they may often breed in their temporary homes. This can result in birds of different geographic origin mixing with one another.
Irruptions of the Scandinavian and Russian populations of Common Crossbill are often cyclical, tied in with the pine cone crop. These birds arrived on the Heather Drilling Platform in the North Sea last July. Photo by Kevan Leek.
It is this partially resident, partially nomadic lifestyle that has probably created the distinctive populations which are causing headaches for ornithologists. In broad terms, there is little visible difference between them, with adult males variably brick-red in colour (less so, for example, in the Carpathians), adult females greenish overall with a yellower rump, and juveniles rather grey-brown and streaky in appearance. Though similar in terms of plumage, these different crossbill populations have developed their own subtle measurements, bill structures and cone specialisations; the differences have been flagged up mostly by their vocalisations.
Jeff Groth’s recordings in North America got the ball rolling in 1993, showing that the continent’s crossbills included nine different vocal types, tallying with differences in bill and palate structure (Groth 1993 and research.amnh.org/vz/ornithology/crossbills). He initially named them ‘Type 1’ to ‘Type 7’, (a further two were defined later) and correlated them with existing named subspecies. He suggested that, as some types had overlapping breeding ranges without any sign of interbreeding, they might all be separate species.
Shortly afterwards, Michel Clouet and Jean Joachim (1996) recorded Common Crossbills from the Alps and Pyrenées, both involving birds of the nominate subspecies, and found that they differed in their vocalisations. Magnus Robb (2000) then published the results of his own four-year analysis of crossbill calls from The Netherlands, concluding that there might be six call types within Western Europe. Frequencies of each type varied in the areas he sampled in different years, and he suggested this was perhaps tied to irruption events.
The call types were based on matching flight and ‘excitement’ calls, and the birds giving them were allocated evocative names like ‘Phantom’, ‘Wandering’, ‘Glip’ and ‘Bohemian’ Crossbill when a total of seven were described in Constantine et al (2006). This was particularly interesting because, unlike North America, mainland European birds west of Ukraine (including those in Britain) were previously all believed to belong to the same subspecies, nominate curvirostra.
Ron Summers of the RSPB was developing similar work in parallel by further correlating the two types of call. He was able to identify nine types, using a number and letter code for each call combination, and several of these matched Robb’s call types. His investigations also included Scottish and Parrot Crossbills, as well as three Mediterranean subspecies of Common Crossbill.
Although classified within the nominate subspecies, the ‘Red’ Crossbills of North America, like this male in Connecticut, form a separate taxonomic clade to Eurasia’s Common Crossbills, and may be split in future. Photo by Julian Hough.
As a number of both Robb’s and Summers’s call types from within the range of nominate curvirostra could not be matched to known subspecies, they were clearly dealing with hitherto unrecognised crossbill forms. Further research on the breeding grounds involving the recording and comparison of song types – a key element in reproductive isolation – might help establish whether the differences in flight and excitement calls are truly population-specific. This would also be a first step towards mapping the breeding ranges of the various putative forms.
Other work in the last decade has resulted in Craig Benkman uncovering much of the complexity of crossbill evolution and ecology in North America. He has shown that not only are the forms of crossbill limited by the shape and size of the pine cones on which they feed, but in many places they are also constrained by competition with squirrels which exploit the same resource. A ‘geographic selection mosaic’ drives the evolution of Lodgepole Pine cones, Red Crossbills and Pine Squirrels through an ‘evolutionary arms race’, as the different forms exploit and prevent exploitation of each other via selection. This is directed by differing ecological influences across a landscape, creating varying responses among the organisms as they co-evolve (Benkman et al 2003, Mezquida and Benkman 2005).
The total absence of Pine Squirrel in the South Hills of southern Idaho has resulted in the emergence of an endemic sedentary crossbill form with high levels of reproductive isolation from the other two nomadic crossbill forms that occasionally invade its range. This distinctiveness led to it recently being named as a new species, South Hills Crossbill (Benkman et al 2009), and it is being recommended for recognition as a separate species by the Committee on Classification and Nomenclature (North & Middle America) of the American Ornithologists’ Union.
Could the same situation occur in Europe? Scottish Crossbill has been found to be reproductively isolated from Common Crossbill in its native Caledonian pine forests, and more recently also from Parrot Crossbill, small numbers of which are known to have been established there since at least 1991, after sporadic breeding attempts which began in Norfolk in 1984. All three species can be found breeding in the same area, largely without hybridising, although this has been proved on a very small scale (see below; Summers et al 2007). Additionally, as indicated by subtle distinctions in bill shapes, they also have different specialisations. The larger-billed Scottish and Parrot Crossbills can feed on the closed Scots Pine cones when they breed in February and March, but Common Crossbill, which feeds on spruce cones as well, has to wait until the cones open in April and May, resulting in staggered breeding cycles, and thus in ‘temporal’ reproductive isolation.
Scottish Crossbill (adult male, top) is intermediate but overlapping in size between Common and Parrot Crossbills – this overlap may account for the fact that crossbills matching Scottish in biometrics have been trapped as far south as Derbyshire. Britain’s sole avian endemic is best identified using a combination of the downcurved culmen, overall bill size and structure and, above all, calls. Parrot Crossbill (first-winter male, above) is up to 15 per cent larger in body mass than Scottish Crossbill and can often, though not always, appear bulkier. These two birds were trapped and photographed in Deeside, Scotland, in October 2009. Photos by Lindsay Cargill.
On the Continent, Pim Edelaar has shown that at least three of Robb’s curvirostra vocal types in The Netherlands breed alongside each other (‘sympatrically’) without hybridising – strongly suggesting reproductive isolation – so it seems likely that more splits may be forthcoming. Three different L c curvirostra call types have been recorded in Britain, indicating that separation of the different forms in the field may be a problem for British birders as well, in future.
How different is different?
The generally accepted definition holds that a biological species is a group of interbreeding individuals which are reproductively isolated from other such groups. Under this definition, the assortative mating shown by many of the vocal types of crossbills makes them ‘good’ species. But is the term ‘species’ an adequate label for what is going on with crossbills? There is very little genetic difference between any population, including Parrot and Scottish, meaning that all plain-winged crossbills could be regarded as a single extremely variable species; alternatively, all the vocal types may be equally as valid as full species as are Parrot and Scottish Crossbills.
The term ‘pseudospecies’ was coined to describe morphologically identical populations which otherwise act as separate species (Knox 1992), but this was before many of the recent discoveries and it may now be an unnecessary distinction. However, the genetic similarity of these crossbills certainly contradicts the modern idea of what a species should be, as most recently discovered cryptic species have been revealed by DNA analysis. This genetic homogeneity implies that each type evolved very recently, frequently hybridises, or even that all plain-winged crossbills are essentially ‘morphs’ of the same species that can and sometimes do interbreed.
Assortative mating between vocal types disproves this latter ‘morphs’ idea, and field observations have so far produced records of no more than 5.1 per cent of any population interbreeding, and usually far less. This is not a high enough figure to homogenise the entire gene pool of the crossbill complex. The known reproductive isolation mechanisms, then, seem to be call type and ecological niche.
Any subtle physical and vocal differences may be due to rapid adaptation to local factors, most likely the fluctuating distribution and species composition of coniferous forest after the last glaciation, marked by colder temperatures and advancing ice sheets. This leaves us with the majority of forms having diverged only recently. In Britain, this would not be surprising as the islands only emerged from under the ice sheets about 10,000 years ago. European forms, however, should be more separable if they were forced into warmer ‘refugia’ – regions supporting isolated populations of once more widespread species – during the longer periods of glaciation. In fact, some of the Mediterranean forms do seem to be slightly more genetically divergent than their mainland counterparts. Logically, the overall similarity of forms must be because they are very recently separated, and perhaps still incipient, species.
Call type currently remains the best clue to separate populations, and this mostly correlates with specialisation on one kind of conifer, or at least one size of cone or the stage at which the seeds can be accessed, as evidenced by the crossbills in Scotland.
At present, in morphological terms there generally seem to be two main plain-winged crossbill groups in continental Europe: the Parrot Crossbill types, with large bills specialising on pine cones and remaining largely sedentary; and the nomadic Common Crossbill forms, which broadly favour spruce and fir species. These crude visual divisions are only sometimes supported by vocal similarities within each group, and the curvirostra complex defies simplification on most counts.
What does the future hold? In Europe, few vocal types have been definitively associated with a discrete population or geographical range. More international fieldwork must be performed to match call types with food sources and breeding ranges, helping to accurately identify the source of the irruptive forms, and identify if any are truly nomadic. It should be noted that the North American forms are a separate phlyogenetic clade to the Eurasian forms – that is, Red and Common Crossbill are evolutionarily self-contained and separate from each other, and represent a potential split in the future.
Birds from Sicily, like this male (top) photographed on Mount Etna in April 2007, are currently lumped within the North African subspecies L c poliogyna, but the Sicilian population may be a discrete taxon in its own right. Male plumage becomes (perhaps clinally) greyer towards the south of Common Crossbill’s range, but the Atlas Mountains, Morocco, population of poliogyna, like this bird (above) photographed in March 2009, may be distinctive enough to be a future split. Photos by René Pop (top) and Arnoud B van den Berg (The Sound Approach; above).
The bill morphologies of Parrot and ‘traditional’ Common Crossbills are clearly different, but the bill depths of detected call types can vary by as little as 2 per cent – far too little to exclude observer bias and error – and standardisation will be very difficult.
The birds’ vocalisations need further examination, too. Are they inherited or learned? Do they remain the same throughout a bird’s lifetime?
Captive studies have shown that ‘foreign’ calls can indeed be learned between call types, and one of a mixed pair has also been observed to ‘change sides’, but both instances are apparently unusual. If this proves more prevalent, then perhaps each form is too flexible to be termed a species, and really they should be called ‘ecotypes’, unique populations adapted to their local environments. The presence of definable ecotypes possibly provides evidence for incipient sympatric speciation, a ‘holy grail’ of evolutionary biologists, but rarely proved in the wild.
Here’s another point to consider. There are more vocal types than suitable conifer species in Europe at least, which could mean that some forms are unspecialised or, conversely, highly wedded to local tree variation; a few may also derive irruptively from the more conifer-diverse areas of northern Asia. The sampling of stable isotopes from captured birds could help decide between these options.
When different forms live in close proximity, interbreeding, in so far as it has been established, is rare, and it was recently shown in captive breeding experiments that females overwhelmingly prefer males of their own call type, even when males of other call types are physically identical. Similar studies on more call types will further define population limits.
There are potentially serious conservation and monitoring problems ahead. When field identification is so difficult, how can we be sure that populations are stable or changing when they can’t be properly censused? The recognition of species rank confirms an organism as discrete and identifiable, and therefore, in theory at least, conservable. If a species is not defined properly for conservation workers and legislators, will we be unable to prevent its extinction?
The newly named South Hills Crossbill evolved in the absence of squirrels, but it’s chilling to realise that two similar forms, perhaps – with hindsight – full species, which were endemic to Newfoundland and Cypress Hills, Canada, have already become extinct due to the artificial introduction of squirrels, which out-competed them for cones. Climate change will also rapidly alter the ranges of different conifer species, and we don’t know if the crossbill types will respond in time.
There is the perverse possibility that Scottish Crossbill and some of the other forms evolved very recently, adapting to newly introduced plantations of Douglas Fir, Sitka Spruce and other alien tree species. Such rapid evolutionary progress appears to have been noted recently in Darwin’s finches in the Galápagos Islands, where a putative new species may have evolved in as few as seven generations (Grant and Grant 2009). This poses other interesting conservation questions – our only endemic bird seems almost as at home in such monocultures as it is in ‘native’ forests.
Chasing the gypsy finch
As bird orders and families become genetically better resolved, taxonomists are starting to concentrate on species- and population-level phylogeography – the historical processes responsible for the current distribution of individually distinct populations. Recent research is now uncovering the ways in which behaviours such as breeding display and vocalisation, as well as ecological niche, act to isolate populations and provide mechanisms for selection and evolution. Studies of crossbills are in the vanguard of such research.
Common Crossbills in the Troödos Mountains, Cyprus, are a separate subspecies, L c guillemardi. This form has already been mooted as a ‘split’ in a German taxonomic journal, and this striking spring male underlines this possibility. Photo by Ruud Altenburg.
What we do know is that each vocal type prefers its own kind and doesn’t mix with neighbouring types, and that each bill size is inherited. The study of vocalisations is providing the key to identifying the different crossbill groups in both Europe and America. It is unfortunate for field birders, however, that at present this is best detected by recording and analysis with a sonagram. Species which are virtually or completely unidentifiable in the field are no novelty to mammalogists, herpetologists and botanists, and perhaps in this way ornithology is catching up with the rest of biology.
Crossbills are slowly adding to our understanding of the speciation process in birds, and work is ongoing (exciting new research from both Europe and North America is expected to be published early in 2010). Birders can’t expect a process as universal, subtle and ongoing as evolution to always provide clear-cut, obvious species to identify, and perhaps seeing wild birds in that context will lead to a wider appreciation of birding than merely listing.
One thing’s for sure: we’re still some way from a European field guide featuring 10 or more species of crossbill.
While any mistakes in this article are my own, my understanding of this subject would have been substantially less without the patient and considered help and advice of Dr Craig Benkman, Lindsay Cargill, Dr Pim Edelaar, Dr Alan Knox, Roger Riddington and Dr Ron Summers, and especially Magnus Robb.
For a full list of references, please click here.
Interested readers may like to visit the following websitesfor practical advice and comment on crossbill vocal-types: