CHAPTER 18

The Bowerbird’s Story

EXTENDED PHENOTYPES

The next oscine branch to evolve after the scrubbirds and lyrebirds gave rise to the Australasian treecreepers (Climacteridae) and the bowerbirds (Ptilonorhynchidae), approximately 45 million years ago (Figure 17.1).1 While the Australasian treecreepers are very similar in appearance and behaviour to their northern namesakes (family Certhiidae), they are a much older lineage, and any likenesses are the result of convergent evolution due to the occupation of similar ecological niches. The Australasian family’s ancient origin is suggested by their early design of syrinx, which reveals a ‘striking departure from the basic muscle pattern that prevails throughout the rest of the oscines.’2 Also, the oldest Climacteridae fossils are known from the early Miocene, and were extracted from deposits at Riversleigh in northwestern Queensland.3 This important site, which has also produced the oldest example of lyrebird, is unusual because the fossils are located in soft freshwater limestone which has not been compressed.4 The resultant well-preserved fossils, together with recent phylogenetic studies, provide yet further evidence that the world’s songbirds originated in eastern Gondwana – present-day Australasia.

The Climacteridae is a small family of only seven species; six of them live in Australia, while one, the Papuan Treecreeper, inhabits the mountains of New Guinea. All are medium-sized, scansorial species, with greyish-brown to black plumage and streaked undersides, which forage for arthropods on the bark of eucalyptus and other trees. Unlike their northern counterparts, Australasian treecreepers do not use their tails for support while climbing but only employ their feet. They climb upwards, placing one foot ahead of the other, circling up the trunk and outwards onto the main branches, using their long, slightly curved beaks to seek out prey. Because they always keep one foot attached to the bark, they can climb upside down beneath branches, making them the only birds in Australia to access this niche. In contrast to other passerines, Australasian treecreepers possess a hind toe that lacks ligaments and an extensor system that allows the rotation of this toe to produce a firmer grip on the toughened bark.5 Australasian treecreepers are cooperative breeders, as are many of the ancient branches of oscines, and obtain help, often from the young males of previous broods, to feed the incubating female and defend the young.

Despite molecular studies placing bowerbirds on the same ancient branch as the Australasian treecreepers, their skeletal anatomy is very different. Indeed, Walter Bock, Professor of Evolutionary Biology at Columbia University, noted that bowerbirds possess ‘some of the most peculiar and unique cranial features in the entire order of perching birds.’6 In particular, their lacrymal bone is unusually large and unlike that of any other songbird, except lyrebirds. Given the marked anatomical, as well as behavioural, differences between treecreepers and bowerbirds, Tim Low has questioned in his book Where Song Began whether both families can belong to the same evolutionary branch. Could it be that the genetic evidence is misleading: the result of a phenomenon known as ‘long-branch attraction’? This recognised limitation of phylogenetics is the tendency of DNA sequences from lineages with long terminal branches to group together, regardless of their actual relationships. In part, the phenomenon relates to the limited number of possible states that rapidly evolving sites can change to (20 amino acids and four nucleotides). Since the branches of bowerbirds and treecreepers are long, sequences could contain many such changes, leading to the development of spurious similarities. In effect, these false resemblances override the real phylogenetic signal, allowing the sequences to group, or ‘attract’ to each other. However, most phylogeneticists, including Professor Christidis, co-author of Systematics and Taxonomy of Australian Birds, believe that both families do belong to the same branch: the second one after the scrubbirds and lyrebirds.7

The precise evolutionary relationship between treecreepers and bowerbirds has little bearing on our present story, however. What is important is that readers appreciate that phylograms based on gene sequences, including all those used to illustrate this book, are merely the most likely result from a number of possible options. In truth, despite the use of sophisticated statistical manipulations, phylogenetic trees are just scientific hypotheses, subject to falsification by the addition of further data or the use of alternative analytical techniques. This fact should come as no surprise, as genetic sequences constitute ‘noisy’ data: the result of millions of years of genetic recombination, horizontal gene transfer and hybridisation, not to mention the confounding problems of conserved sequences and long-branch attraction. Nevertheless, the analyses of known phylogenies, such as viral populations that have evolved in the laboratory, indicate that the widely used statistical approaches (e.g. bootstrapping) give, in general, a reliable measure of phylogenetic accuracy (values of 70 per cent or higher, supporting reliable groupings).8

Now let us return to the bowerbird’s story, one that explores the concept of the extended phenotype and its role in sexual selection.

The bower

Most male traits that catch the female eye have a clear benefit for the species. Many female insects, for example, require a ‘nuptial gift’ from their suitors. This is usually something edible, such as another bug, which can supply nutrients and so directly increase the chances of successful mating when food supplies are limited. The presence of butterflies dancing in the sunlight around a drying mud puddle is usually the result of males collecting salt crystals to present to potential mates. Birds are no different. In the United States, Blue Jays pass tasty morsels from beak to beak, while the majority of roadrunner matings involve the transmission of food. Birds also select partners by their nest-constructing abilities. The male Eurasian Wren, for example, offers his mate a choice of three or four nests that he has built. This approach makes sense, as a well located and optimally built nest will increase the offspring’s survival rate. For other species, including Marsh Wrens, European Robins and Red-winged Blackbirds, it is the male’s ability to defend a large territory that wins out, as bigger areas are more likely to provide sufficient resources for rearing young. Most seabirds, although they breed in colonies and do not hold territories, still need to show prospective mates that they have real estate of their own.

In contrast, some avian species favour traits that appear, at first sight, to be somewhat arbitrary. Why, for example, would a female select a mate by his song, mimic ability, plumage decoration or prowess on the dance floor? Although the genetic advantages of these seemingly ‘whimsical’ criteria are now partially understood (discussed in subsequent chapters), it is the biological benefit of the bowerbird’s bizarre displays, especially the decorated bower, which has been the most challenging to explain. In the words of the American ornithologist Thomas Gilliard, bowerbirds ‘raise difficult questions, questions that penetrate to the very foundation of our biological theories.’9 But before we attempt to explore such issues, we need to describe the bower itself, the ‘extended phenotype’ and central character of the bowerbird’s story.

Of all the structures built by animals, there can be none as strange and magical as the bowerbirds’ ‘love-shacks’. Indeed, the first Europeans to encounter these remarkable chambers were convinced that they were the handiwork of local women, playthings constructed for the amusement of local children. Bowers can be one of two types, each built by species that appear to be phylogenetically closer to each other than to the builders of the alternative configuration. The first form, known as a ‘maypole’ bower, consists of a central supporting vine or sapling, around which are added woven sticks that radiate outwards in all directions, in the form of a tepee or Christmas tree (Plate 24A). The Vogelkop Bowerbird and the Streaked Bowerbird, however, add massive roofed, hut-like structures up to 2 metres across, often with a double entrance and a covered porch.

The second form, or ‘avenue’ bower, consists of parallel vertical walls of grass stems and sticks, planted into a base of vegetation. Each wall curves gently outwards, with one end of the bower exiting onto a display area where most of the decorations are arranged (Plate 24B). The Satin Bowerbird’s avenue consists of just two rows of sticks, while that of the Yellow-breasted Bowerbird contains a double avenue, with two parallel paths on a raised platform.

Paradoxically, not all bowerbirds build bowers. Catbirds, as well as the Tooth-billed and Archbold’s Bowerbird, merely clear areas of forest floor for their displays, although Archbold’s Bowerbird covers his with a thick mat of ferns. The construction of bowers and their absence in several family members raise some interesting questions: how, for example, did the behavioural trait evolve, and what is the function of the bower?

To address the first question, Rab Kusmierski and colleagues from the University of Maryland constructed a molecular phylogeny, one that revealed all bowerbirds to have evolved from a single ancestral species.10 By studying individual relationships, the team concluded that the common ancestor of all extant species (currently 27) was probably a monogamous, dull-plumaged bird that lacked bower-building and court-clearing skills. Indeed, this ancestral state characterises the catbirds, the first bowerbird genus to evolve, around 20 million years ago.11 Catbirds are found in eastern Australia and the mountains of New Guinea and are mainly green, nondescript birds. Although they are known to attract mates by displaying food or colourful objects in their beaks, they do not build bowers. Furthermore, catbirds are monogamous, and once a mate is chosen they remain together for life since, unlike in other bowerbirds, both parents are required for raising a brood. It was only during the last 20 million years, following the divergence of catbirds, that the family’s characteristic court-clearing and elaborate decorating behaviours evolved. Such traits, coupled with the transition to polygyny (males mate with more than one female), provided the preadaptations that enabled the evolution of bower-building to occur. According to the leader of the research group, Gerald Borgia, the first bowers were probably simple structures consisting of a display court that surrounded an isolated sapling or vine. As the bower-building lineage evolved, a second major divergence occurred, one that produced the avenue- and maypole-building clades. Later, two members of the maypole clade, the Tooth-billed and Archbold’s Bowerbird, independently lost their bower-building abilities. Such dramatic shifts in behaviour must have evolved quite rapidly, given that both species have close phylogenetic relationships to the maypole-building Vogelkop Bowerbirds.

Learning is likely to play a significant role in bower construction. Young male Satin and MacGregor’s Bowerbirds, for example, are inept at building bowers and often adorn them with inappropriate objects, often of the wrong colour. Youngsters also spend extended periods of time watching old males at work and take up to 4–7 years to develop the necessary architectural skills, a fact that may account for bowerbirds having the longest life expectancy of any passerine. Similarly, females go around in groups to visit bowers, so that the younger females can learn the species’ style preferences from their older kin.

The wide variation in bower design between the different species may have evolved as a result of changes in female preference or choice, a view first proposed by Jared Diamond.12 While undertaking field work in West Papua (the Indonesian half of the island of New Guinea) in the 1980s, Diamond discovered an isolated population of Vogelkop Bowerbirds that, although only 100 kilometres away from other populations in the same mountain range, built remarkably different bowers with distinctive decorations. Instead of large, hut-like bowers, copiously decorated with colourful objects, the new population constructed maypole bowers without a roof, and decorated exclusively with drably coloured objects – stones, acorns and moss – even though bright objects were available. Importantly, these differences have evolved despite the males in this isolated group having access to the same range of environmental materials as other Vogelkop Bowerbirds. Diamond concluded that ongoing sexual selection was the most likely explanation for an individual male’s preference. In support of this idea, variations in female choice are believed to give rise to local, or fine-scale ‘cultural’ differences in the bower design of the Spotted Bowerbird.13

Albert Uy and Gerald Borgia from the University of Maryland decided to conduct an experiment to see if Diamond was right.14 They placed a collection of differently coloured small plastic tiles just beyond the males’ court and filmed the birds to see which objects were incorporated into their displays. Although both the new population and males from elsewhere readily made use of the novel items, their choices were very different. The traditional Vogelkop Bowerbirds used the red and blue tiles and ignored the dull ones, while the new population preferred the drab ones and disregarded the coloured tiles. Also, females preferred the display traits, including the bower shape, size and decoration, of males from their own population. Given these findings, the researchers concluded that the different displays are the result of rapidly evolving female preference or choice. Since the two populations are allopatric, with minimal genetic differentiation, it is likely that they are incipient species. In other words, if the trend continues, female preference will lead eventually to two different species of bowerbird. These findings are significant because they provided the first direct evidence for the ‘speciation by sexual selection’ hypothesis in birds: a model in which the reproductive isolation of a population results from the relatively few genetic changes associated with female preferences and male display traits. But what could account for a change in female preference? Uy wonders whether ‘visual conditions, such as the play of light and shadow on foliage, could select for signals that are most effective in those particular environments. Or are these changes arbitrary relative to habitat, analogous to fads we see in our own society?’15

Whatever the cause, the highlighted field study suggests that female choice over many millions of years may have contributed to bowerbird speciation. It is a mechanism that is likely to have acted in conjunction with other processes, especially vicariant evolution. For example, population fragmentation, due to Pleistocene glaciations and the uplift of New Guinea’s central mountains, is known to have underpinned the speciation of catbirds.11 Furthermore, New Guinea’s orogeny has contributed to the speciation of the genus Amblyornis, which includes a central cordilleran form (MacGregor’s Bowerbird), the Bird’s Head form (Vogelkop Bowerbird), and a North Coastal Range form (Golden-fronted Bowerbird).

Bowers are obviously important to their owners, since males spend hours constructing, decorating and maintaining them. In fact, no male bowerbird has ever been noted to have mated successfully without first having built one. Surprisingly, this unique behavioural trait was only made possible by the availability of fruit. For bowerbirds thrive on the region’s variety of abundant, high-energy fruit and, as a result, spend little time or effort foraging. The resultant release from the daily pressures of finding food opened up new evolutionary possibilities and allowed the development of bower-building and polygyny. Indeed, the ‘many-female’ breeding system is only possible if females have enough spare time to observe courtships displays, build nests, incubate eggs and provide for their chicks. Males, on the other hand, must be able to devote time and energy to tending their bowers and wooing females without the competing need to search for food. As a result, successful male Satin Bowerbirds may mate with up to 25 different females, with the record being held by one male who inseminated 53 individuals within 12 months. Interestingly, the non-bower-building Tooth-billed Bowerbird is the most committed leaf-eater, having evolved specialised serrated jaws, with cusps and notches, to facilitate its folivorous lifestyle. Without fruit, therefore, bower-building and the polygynous breeding system of bowerbirds could never have evolved.

But why are bowers and their associated decorations so important for female bowerbirds? In other words, what is the female assessing when she selects a mate? To try and answer these questions, biologists have undertaken numerous field observations, using remote cameras, and structurally altering bowers and their decorations, to determine the effect on male mating success.

Collectively, these experiments indicate that bowers may reduce the threat of sexual coercion, physical molestation and forced copulations, and allow females to observe court decorations from close range. Indeed, for females to profit from mate choice they need to be able to visit and watch the courtships of as many different males as possible and reject those deemed unsuitable. At the same time, a male has to carefully assess a female’s readiness to mate while performing his dynamic vocal and dancing displays. If the signals appear favourable, then the male approaches slowly behind her to copulate. Any female with second thoughts can easily escape while the male negotiates his way around the intervening bower wall or maypole. In support of the threat-reduction hypothesis, experimental tinkering with the structure of bowers has shown that female Spotted Bowerbirds prefer to assess males through an intact straw wall, only to choose the most vigorously displaying male. If the wall is experimentally removed, then courting males lessen the intensity of their displays – an observation that implies that threat reduction has influenced the evolution of the species’ physical and behavioural displays.16

Further evidence for a linkage between the bower and female sexual autonomy is provided by studies of the Satin Bowerbird, a species in which the male modulates his display intensity according to the female’s body language. Gail Patricelli, a graduate student working with Gerry Borgia, created a remote-controlled robotic female bowerbird, dubbed a ‘fembot’, to determine the finer points of the species’ sexual interactions. The highly engineered, feathered construction was able to rotate her head, fluff her wings, tilt her beak and assume a mating posture, to fully mimic the female’s range of reactions. After carefully placing the fembot inside a bower, Patricelli operated the controls from a hide and, once a male was in attendance, sent four different female courtship signals, including consent. By studying the male’s response, she was able to confirm that those individuals who adjust their display intensity to keep females more at ease and relaxed were the most sexually successful.17

Species that do not construct bowers but possess display courts have evolved alternative ways of reducing the risk of unwanted male attention. The female Tooth-billed Bowerbird selects a suitable male before arriving on a court, so there is little need for physical protection. In contrast, the male Archbold’s Bowerbird has modified his display court such that aggressive behaviour is prevented. Orchid vines are draped on numerous overhanging branches to produce a series of curtains that criss-cross the court, almost reaching the ground. Should a female inspect his handiwork, the male flies down and maintains a crouched position below the vine curtains. If interest is shown, he approaches in a non-threatening way by keeping his body low to the court floor. In effect, the low-hanging vegetation reduces any opportunity of taking the female by surprise.

The reduction of threat by male bowerbirds may help our understanding of the display adaptations of other species. The Blue Bird-of-Paradise, for example, hangs upside down when displaying, a strategy that may have enabled the species’ highly intense courtships to evolve, while at the same time reducing the threat to females. Since many species, especially waterfowl, gain reproductive success through aggressive sex (see The Waterfowl’s Story), why would male bowerbirds build structures that thwart such behaviour? A likely explanation, according to Borgia, is that bowers encourage more females to attend on a greater number of occasions, outcomes that more than compensate for any loss of forced copulations.18

Cognitive ability

Mating success, however, depends on much more than just the offering of a functional or ‘protective’ bower. Females also take note of the overall quality of the male’s building efforts. For instance, the mating success of Satin Bowerbirds depends on the degree of bower symmetry, as well as the density and type of sticks used – thick curved walls, made from fine tightly packed sticks, being the winners.19 The critical importance of a carefully sculptured structure is reflected in the extraordinary lengths males will go to repair any damaged or asymmetrical bower. Indeed, Satin Bowerbirds have evolved a unique mechanism of rebuilding known as ‘templating’, in which they use the standing wall as a template to measure the size and position of every stick to be placed in the new wall. This process is repeated along the whole length of the bower and ensures that the new wall is a mirror image of the original standing wall.20

The courtship of Satin Bowerbirds has received particular attention, with the activity of individual birds being monitored for many years using automatically triggered cameras. Typically, once a female Satin Bowerbird finds an ‘aesthetically’ pleasing bower, she steps into the avenue of parallel sticks and surveys the decorated court from a vantage point predetermined by the male. Once she is inside, the male becomes very excited and runs to his display area to pick up his most prized possession and shows it off to his potential mate. He then carries out a ritualised display of exaggerated movements and sounds, dubbed the ‘buzz/wing-flip’. This frenetic performance involves strutting and bowing, with outstretched and quivering wings, while simultaneously issuing a variety of mechanical-sounding vocalisations, such as buzzing, hissing and chattering. Finally, he makes a more intimate appeal, coming almost beak to beak, and softly mimics a variety of other Australian bird calls. If she is impressed, the female signals her willingness to mate by adopting a low, submissive crouching posture, and after copulation leaves to raise the next generation on her own, while the male readies himself for courting more prospective females. The relevance of this account is that it highlights two additional behavioural traits that influence female choice, or sexual selection: court decoration and mimicry.

The arrangement and quantity of displayed objects appear paramount, and this explains why males spend many hours improving their courts, even if it means stealing objects from rival males. Each species has its particular taste in colour, size and alignment of ornamentation. Satin Bowerbirds, for example, prefer bright blue items, Great Bowerbirds like green and white, Streaked Bowerbirds favour yellow and red, while Fawn-breasted Bowerbirds opt for green. ‘Novelty’ items, rather than objects of any particular colour, are preferred by Vogelkop Bowerbirds, while Regent Bowerbirds coat their bowers with pea-green saliva paint, sometimes using a leaf as a paintbrush. Objects may not be confined to natural debris and can include human rubbish, such as broken glass, bottle tops, snack-food wrappers, ballpoint pens, plastic straws and even children’s toys. In general, the more decorations on display, the more successful will be the male owner: a finding underscored when researchers demonstrated that experimentally removing objects reduces a male’s attractiveness to females.

Laura Kelly and John Endler from Deakin University have shown that male Great Bowerbirds use visual tricks to influence female judgement, by arranging objects so that they increase in size as the distance from the bower increases.21 The resulting size–distance gradient creates a forced perspective and results in false perceptions of the bower’s geometry when seen from inside the bower. If the displays are experimentally rearranged, by placing the larger objects closest to the bower and the smaller ones furthest away, the males set about restoring the original pattern. This happens very quickly – in all instances, the forced perspective is recreated within three days. The overall geometrical effect is that females may discern the court to be smaller than it is, while the male appears to be bigger, an illusion that seems necessary for successful mating.22

According to the biologist Seth Coleman, mimicry also plays a significant role in the final decision of mate choice by female Satin Bowerbirds. By analysing vocalisations from thousands of hours of recordings, Coleman showed that the best mimics, those with the broadest repertoires, were the most successful in the mating game. The top males accurately mimicked four to five species, including kookaburras, honeyeaters and cockatoos, while the inferior males could only imitate poorly one or two species.23 The best mimics also possessed the largest number of court decorations and built the better-quality bowers. Although mimicry is widespread within the bowerbird family, only Satin Bowerbirds use it during courtship, suggesting that it must have been a pre-existing trait that was later adopted for courtship displays by the Australian species.

Despite Thomas Gilliard’s view that understanding bowerbird courtship is a scientific challenge, progress has been made. Field experiments have shown that it is the female’s assessment of male cognitive ability that underpins their elaborate courtship routines and accounts for the importance of bower design, court displays and mimicry. Jason Keagy and his colleagues, for example, gave male Satin Bowerbirds a set of problems to solve as an indicator of their cognitive ability. Since males have a strong aversion to anything red on their courts, the team created situations in which added red objects were increasingly difficult to remove. One test involved placing a transparent container over the offending items so that the males needed to work out how to tip the container and remove the objects. In a more challenging test, the red objects were fixed to the ground so that they could not be eliminated. Only the ‘brightest’ birds realised the nature of the problem and concealed the objects with leaf litter. When the team looked at the mating success of the birds, they discovered that the males with the best performances were also the most sexually attractive and mated with the most females.24

Females, it seems, may be using a combination of display traits as a sort of sexually selected intelligence test of potential mates. If there are fitness advantages to better cognitive ability and these are inheritable, then females choosing the ‘cleverest’ males are likely to have offspring with the same benefits. The higher cognitive ability of the descendants may enable them to live longer by avoiding predation, find more food, establish better territories, and avoid parasitic infection by having better immune systems. According to Gerald Borgia and Jason Keagy, females can be seen as sophisticated decision-makers able to make complex fitness-enhancing mating decisions, while males can be perceived as using their cognitive processes to construct highly effective new displays.20

Extended phenotypes

I have not yet fully explained the rationale for the present chapter’s subtitle, ‘extended phenotypes’. Before the seminal work of the evolutionary biologist Richard Dawkins, the phenotypic influence of a gene was thought to be limited to an individual’s body. Even so, the outcome of a single genetic change can be wide-ranging and affect far more than a single facet of an organism’s phenotype.

As an example, let us return to one of the key adaptations of the Bar-headed Goose, namely its high-affinity haemoglobin (discussed in The Waterfowl’s Story). A mutation in one of the haemoglobin genes enabled the species to gain a significant advantage by being able to fly longer and higher for the same amount of expended energy. Those individuals that possessed the adaptation would have survived longer and been more likely to pass the mutated gene to subsequent generations. Over millions of years, the ‘improved’ section of DNA has come to characterise the species, rather than the alternative forms of haemoglobin gene present in other geese. The modified gene and its resultant protein have had a cascade of consequences for the species: a structurally different haemoglobin within red cells, increased oxygen binding by blood, enhanced oxygen delivery to heart and brain, and, ultimately, the ability to undertake the highest-altitude sustained migration of any species. This impressive range of consequences, from red cell to migratory behaviour, still falls within the classical view of a gene’s phenotypic effect.

However, in his influential 1982 book The Extended Phenotype, Dawkins stresses that such a cascade of events should not be limited to the physical characteristics of an individual, but should include ‘all effects of a gene upon the world.’25 Each stick in a beaver’s dam and each heavy stone carried by the male Black Wheatear during the breeding season are expressions of the individual’s genotype, and therefore belong to the organism’s phenotype.26 So it is with bowers: female bowerbirds can read a male’s genetic quality not only from his appearance and behaviour but also from the artefacts that he produces. To paraphrase Dawkins, although the bowerbird’s phenotype may be made of sticks and bottle tops rather than living cells (hence an extended phenotype), it is no less a true phenotype.27

Since bowers constitute an extended phenotype, then, by definition, there must be genes controlling their construction and decoration. Such an argument holds true even though there is a major learning component. But how can we be so sure, given that no genetic research has yet been undertaken? Firstly, males reared in isolation still construct bowers, albeit rudimentary ones, suggesting that such behaviour is innate. Secondly, a genetic predisposition is likely, given that most bowerbirds exhibit the trait, despite its absence in all other avian families.

Since bowerbirds have taken millions of years to evolve their unique behaviour, and each male takes many years to master the art of bower construction, it is maybe not surprising that scientists require a little longer to reveal their genetic secrets.