Figure 21.1 The Passerida, the largest songbird radiation, gave rise to over 3,500 species, from tits to tanagers.
The starling’s story is one of structural colours – how they evolved and how they promoted a dramatic diversification of a group of African birds. But first we need to discuss the origins of the starling family (Sturnidae) and how they came to inhabit the African continent in the first place.
The Passerida
As we have already highlighted (see Figures 16.1 and 19.1), the core corvoids emerged from the proto-Papuan archipelago around 45 million years ago. But this was only the first of several early songbird radiations. The core corvoids were followed by two smaller dispersals – the ‘transitory oscines’ – and then by the Passerida, the largest of all songbird radiations. It was from the latter group that the starlings evolved, although they would require a further 20 million years to do so.
The ‘transitory oscines’ first gave rise to three small clades that remained within Australasia: the berrypeckers and longbills (Melanocharitidae), satinbirds (Cnemophilidae), and Australasian robins (Petroicidae). The berrypeckers and longbills consist of a variety of small songbirds with generally dull plumage that are restricted to the mountain forests of New Guinea. The females of two species, the Fan-tailed and Streaked Berrypeckers, appear unique among the oscines in that they exhibit a reversal of the usual pattern of sexual dimorphism, with females being larger and bulkier. The three species of satinbird are also restricted to the central montane areas of New Guinea. They all build dome-shaped nests, with the female tending their single chick without any contribution from the male.
The Australasian robins comprise 49 species that occupy a broad range of wooded habitats from subalpine to tropical rainforest. They are ‘perch-and-pounce’ insectivores that can cling sideways onto tree trunks and survey the ground below without moving. The best known is the Black Robin, a species that occurs exclusively on the Chatham Islands, a small archipelago lying 700 kilometres east of New Zealand. By 1976, only two females and seven males remained, the lowest known population of any bird on Earth at the time. Despite only one of the females (nicknamed ‘Old Blue’ because of her coloured ring) being fertile, the population has slowly increased with intensive conservation efforts and now stands at around 250 individuals.
Figure 21.1 The Passerida, the largest songbird radiation, gave rise to over 3,500 species, from tits to tanagers.
The second divergence spawned the picathartes and rockjumpers in Africa as well as the elusive Rail-babbler in Malaysia, Borneo and Sumatra.1 Indeed, it was my interest in understanding the biogeography of this unusual group that first exposed me to the field of avian evolution and led to the research for this book. But it was the subsequent dispersal event, the Passerida, that was the most significant, as it spread rapidly across the world and gave rise to over 3,500 extant species.2 They include not just the starlings, but also thrushes, warblers, white-eyes, finches, tanagers and many others (Figure 21.1)
The early history of the Passerida is conflicting and obscured by many gaps and uncertainties that result, in part, from the group’s extremely rapid divergence. While some families must be closer relatives than others, the available data do not allow their precise evolutionary relationships to be resolved – a situation termed ‘soft polytomy’ by phylogeneticists.3 Nevertheless, it is widely accepted that the Passidera emerged from Australasia, sometime during the Eocene,4 and reached the Old World by one of two possible routes. Either they used the islands on the Sunda Shelf as stepping stones to Asia, or they crossed via the now submerged plateaus (Kerguelen, Crozet and Broken Ridge) to reach Africa via the Indian Ocean.5 Whichever route they took, the common ancestor of starlings, oxpeckers (Buphagidae) and mockingbirds (Mimidae) appeared around 23 million years ago, amid the extensive forests of the northern hemisphere.4 There then followed a prolonged period of global cooling, a climatic change that resulted in the population’s fragmentation and subsequent radiation. Eventually, these early birds gave rise to the oxpeckers in Africa (22 million years ago), mockingbirds in North America (21 million years ago) and the starlings and mynas in Africa (16 million years ago).6
Surprisingly, members of the Sturnidae never dispersed to the Americas, and the continent’s current superabundance of Common Starlings resulted from a bizarre introduction programme. In the 1890s, Eugene Schieffelin, an eccentric drug manufacturer from the Bronx, released 100 starlings in New York’s Central Park. His idea was to bring to North America all the birds mentioned in Shakespeare’s plays, and although bullfinches, skylarks and song thrushes failed to survive, the starlings adapted quickly. Since then the birds have spread to all parts of the United States, most of Canada, and parts of northern Mexico, with an estimated population today of around 200 million. The starlings’ success relates to their exploitation of a large variety of habitats, nest sites and food sources, while their aggression and gregariousness make it difficult for native birds to compete. Sadly, the New World was not their only territorial conquest, for successful introductions also took place in the West Indies, New Zealand, Australia and South Africa at around the same time.
The present-day restriction of oxpeckers to Africa may not have always been the case, as environmental changes elsewhere could have considerably reduced their range. For oxpeckers have a highly specialised manner of feeding that involves the removal of ticks and other parasites from the skin of large grazing animals. As a result, the diverse megafaunal communities that once roamed across Eurasia and the Americas could have supported a widespread oxpecker-like ancestor. The rapid extinction of these herbivores at the end of the Pleistocene, however, would have led to a parallel extinction of any symbiotic birds. It has even been suggested that the association with large mammals is ancestral to the entire radiation, rather than merely an acquired behaviour confined to oxpeckers. Most of the starlings and their relatives, for example, regularly forage around the feet of large ungulates, and many will perch on animals while they feed. Although such associations are uncommon in the New World, several populations of Galápagos Mockingbird are known to exhibit oxpecker-like behaviour.7 In addition to fruit and marine arthropods, they will also feed on skin parasites, drink blood and pick at the wounds on marine iguanas, nesting seabirds and even sea-lions.8
During the early Miocene, the African climate became significantly drier, a change that resulted in a marked reduction in the continent’s forests. Since only a minority of extant African starlings are strictly arboreal, it is likely that such climatic changes and the subsequent emergence of grasslands kick-started the clade’s speciation. The population then underwent two further radiations, one that colonised Madagascar and a much larger one that re-invaded the Palaearctic and the Orient around 14 million years ago.9 Interestingly, the two species that are probably most familiar to western birders, the Common Starling and the Spotless Starling, belong to an isolated lineage at the base of the Eurasian radiation. It seems likely that both species evolved after their common ancestor became split following the formation of southern glacial refugia during the Pleistocene. Ongoing climate change, especially over the last 6 million years, encouraged further diversification of both the African and Eurasian starlings by creating extensive open habitats.10
Two additional traits, prying and flocking, deserve consideration, as they are germane to the evolution and speciation of the Eurasian clade. Prying or open-bill probing is unique to this lineage and is dependent on the evolution of powerful protractor muscles in the jaw. This anatomy enables the birds to insert their closed bills into a substrate and then open them forcefully to dislodge hidden prey. The thrust required is the exact opposite to that needed by other birds, which need the greatest force when closing the bill, for example when cracking seeds or holding prey. The anterior region of the skull also coevolved to enable the eyes to have a binocular view of the space between the parted beaks. Although both structural modifications are present throughout the clade, they are most prominent in the Common Starling and the White-cheeked Starling. Since both species lie some distance apart phylogenetically, the anatomy required for extreme prying must have evolved twice, which suggests a high evolutionary lability or plasticity for these traits.11
Most members of the Sturnidae move nomadically in flocks, especially when tracking their food sources, and roost in large numbers during the non-breeding season. These behavioural traits offer safety from predation, prevent heat loss at night, and facilitate the exchange of information. Importantly, the family’s flocking behaviour contributed to their successful colonisation of new regions and may explain why many mockingbirds and starlings inhabit remote islands and archipelagos, often with small populations. For example, the Rarotonga Starling occurs on just one volcanic island in the Cook Archipelago, while the Atoll Starling is restricted to a few tiny Pacific islets.
The speciation of African starlings, however, involved more than climate and habitat change. It was also dependent on the evolution of iridescent feathers.
Until I visited Tsavo East National Park in Kenya, my encounters with starlings had been limited to sightings of the ubiquitous and mischievous Common Starling. Despite having a widespread distribution and possessing a distinctive wheezing and whistling call, Common Starlings are nevertheless a frequently overlooked species. They appear uniformly black at a distance, and it is only when observed at close quarters and in certain light conditions that their glossy feathers, with subtle violet and green iridescence, can be fully appreciated. My awareness of their ephemeral beauty, however, did little to prepare me for the extravagant palettes adopted by their African cousins (Plate 30). It was while waiting at Tsavo’s entrance gate, with the famous dust-red elephants in attendance, that I first became beguiled by their remarkable plumages. The birds’ descriptive names – Superb, Golden-breasted, Greater Blue-eared and Violet-backed – failed to convey the range of spectral hues on offer as I watched them jostle for drinking water in the midday sun. But why are there so many species of starling in Africa, and why do they sport feathers with such wildly different colour combinations? Before answering these questions, we should say a few words about the underlying mechanisms of colour production.
All animal colours are the result of two basic mechanisms: pigment colours and structural colours. Pigments produce their effect by selectively absorbing and reflecting specific wavelengths of visible light. If no light is reflected, we see black. If it’s all reflected, we see white. Gradations in between give rise to many of the colours observed in the avian world, from the browns of thrushes to the reds of woodpeckers and redpolls. All avian pigments are derived from three classes of organic compound: melanins, porphyrins and carotenoids. For example, the sparrow’s earthy shades and the black plumage of corvids are the effects of different melanin molecules (from the Greek melanos, meaning ‘dark’), the same pigments that give human skin and hair its colour. The vivid greens and reds of turacos are produced by porphyrins, while the canary’s yellow plumage is the product of dietary carotenoids. Crucially, each of these substances has important biological functions other than giving feathers their colour. Melanins act as structural components and their expression in feathers makes them stronger and more resistant to wear – which explains why many white species have black wing-tips, for extra strength. Porphyrins are essential components of the oxygen-carrying protein haemoglobin and several detoxifying enzymes in the liver, while carotenoids have antioxidant activity and are important stimulants of the immune system.
Not surprisingly, the expression of these compounds is highly conserved, and the chance of evolving different colour combinations is small because it would require the development of novel and complex metabolic pathways. But such innovations must have occurred in the past. For example, many lineages independently acquired the ability to alter yellow dietary carotenoids to form orange and red pigments without significantly affecting the individual’s health. But blue pigments never evolved. This spectral omission suggests that blue may be extremely challenging or even impossible for birds to create, a concept referred to as the ‘blue rose’ hypothesis. (Horticulturists have long considered blue roses to be the unattainable ‘Holy Grail’ of the flowering world).12 Furthermore, anthocyanins, the naturally occurring blue pigments, are degraded during digestion and so are unavailable for use by birds. The creation of blue feathers, therefore, required a different solution: structural colours.
Structural colours are produced by optical effects – interference, refraction or diffraction – when light interacts with regularly spaced nanometre-scale structures with varying refractive indices. If the reflected light is randomly scattered, white light is perceived, whereas if the wavelengths are ordered, vivid metallic iridescence can result. To see this effect, just twist a compact disc in daylight and view the bright spectrum of colours produced by the light scatter from the regularly spaced grooves. It is the latter’s periodicity that allows the reflected rays to amplify each other and create the strong colours that readily change depending on the angle of view. Similarly, the bright blue iridescence of a male Morpho butterfly is produced by corrugated ridges present on the scales of its wings.
The structural colours of feathers, however, are generated by the nanoscale periodicity of their barbule constituents: melanosomes, keratin and air. Melanosomes are melanin-containing packages that self-assemble during the development of feathers to form stable, ordered structures. They are truly minute: the majority are only 200–600 nanometres in diameter, such that 200 could fit across a human hair (a nanometre is one-billionth of a metre).13 Melanosomes lie embedded in a thin layer of beta-keratin just beneath the barbule’s surface. Beta-keratins are filamentous proteins that readily cross-link to provide robust, lightweight bundles that give feathers, as well as beaks and claws, their strength. Many birds combine keratins with the melanin-containing melanosomes to increase the range of available colours, since the two proteins have different refractive indices and structural periodicities. Interestingly, keratins alone can produce blue structural colours. The male Eastern Bluebird, for example, generates its striking hues solely from air trapped within the barbules’ beta-keratin channels. In fact, blue colour-producing nanostructures have evolved independently many times in over 20 bird families, including kingfishers, jays, manakins and honeycreepers.14
The evolution of further structural complexity has enabled the generation of even greater spectral ranges. For example, the male Lawes’s Parotia, a member of the bird-of-paradise family, has evolved a unique cross-section structure of its barbules to create colours that change dramatically with feather orientation. The species’ basic nanostructure produces a bright orange-yellow reflection, but since each barbule is V-shaped in cross-section, its sloping surfaces also act as reflectors of blue light. As a result, small movements of the feathers during courtship displays can cause the colour to switch from yellow-orange to blue-green suddenly, so increasing the chance of catching a female’s eye.15 Other species combine the effects of pigments and nanostructures. The characteristic green plumage of parrots, for example, is the result of laying a yellow pigment over a blue reflective layer of melanin and keratin.
Should readers remain sceptical that the blue colouration of birds is structural in nature, try the following simple experiment. Take any blue-coloured feather and shine a light on it from above, and you will observe bright blue. However, if you shine the light from underneath, the feather will appear dirty-brown owing to its melanin and keratin content acting as pigments.
But how is it that African starlings have generated such a diverse palette of colours compared to other avian families, when all feathers possess melanosomes and keratin? It was a mystery that caught the attention of Matthew Shawkey, a biologist working at the University of Akron, Ohio, and an authority on the optics and evolution of animal colours. Shawkey’s approach was to determine the structure of the birds’ melanosomes and relate the findings to the African starlings’ phylogenetic tree.16 The results indicated that the common ancestor of all starlings possessed relatively simple melanosomes, solid rod-like structures, that can still be found in some extant species. They include all 11 red-winged starlings (genus Onychognathus), a clade characterised by dark iridescent sheens. However, the researchers determined that 6 million years ago, as the early starlings began to form new species, different types of melanosome emerged, ones with shapes that interacted in novel ways with light. Some species developed flattened, oval-shaped melanosomes that allow for a more tightly packed arrangement. Others evolved air-containing, tube-like structures that offered a greater number of interfaces for the scattering of light. The most complex structures to emerge, however, were platelet-shaped, being both flattened and hollow: forms that can form colour-producing single layers, multilayers or alternating platelet–keratin layers.17 Surprisingly, each starling species possesses only one of the four types of iridescence-producing melanosomes (Figure 21.2).
The American team concluded that the remarkable spectrum of colours possessed by African starlings results from the clade’s unique possession of all four types of melanosome. They also postulated that it would take only minor tweaks to the shape or spatial arrangement of any particular melanosome to produce a new range of metallic colours and, potentially, a new species. Such a conclusion makes sense, as the genetic changes required to produce such changes (e.g. alteration of the thickness of melanosome wall, the amount of air content, or the number of layers) are much more likely to evolve than those required to synthesise new pigments. Indeed, it is now known that the three complex melanosomal forms emerged on many occasions during the last 6 million years. And yet, although they can evolve from one to another, they have not been shown to revert to their ancestral state. This finding surprised Rafael Maia, one of the American investigators: ‘I thought that maybe you’d have a lot of changing back and forth, but actually, once these complex structures evolve, they stick.’18 It is likely that this unidirectional pathway of melanosome evolution, from simple to complex, may have contributed to the starlings’ diversification, since by thwarting a return to the ancestral state it forced phenotypic change. Furthermore, not only did novel colours emerge that were unattainable with pigments, but the colours also became significantly brighter, by up to twofold. According to Shawkey, ‘evolving these new melanosomes was like inventing the wheel for these birds – it allowed starlings to reach new colours at an incredibly fast rate.’19 Indeed, it transpires that African starlings have evolved new colours 10–40 times faster than their cousins with simple melanosomes.
Figure 21.2 Morphology of the four types of melanosome responsible for the iridescence in African starlings, and which underpinned the clade’s rapid speciation: (A) solid rods; (B) solid platelets; (C) hollow rods; (D) hollow platelets. Redrawn with permission from Maia et al. (2013).17
Sexual selection suddenly had another key ornamental innovation to play with, especially as plumage colour provides an ‘honest’ marker of a starling’s fitness. Iridescent colours may be costly to produce and maintain, and only the healthiest birds can produce the most brilliant and vibrant hues. Since African starlings rely on colour for social communication and courtship, any sudden change could have acted as a barrier to genetic exchange and encouraged the rapid evolution of the dozens of species we see today.